Increased Power Flow Guidebook

Electric Power Research Institute
3420 Hillview Avenue, Palo Alto, California 94304-1338PO Box 10412, Palo Alto, California 94303-0813 USA
800.313.3774650.855.2121askepri@epri.comwww.epri.com
Increased Power Flow Guidebook
Increasing Power Flow in Transmission and Substation Circuits
Effective December 6, 2006, this report has been made publicly available in accordance
with Section 734.3(b)(3) and published in accordance with Section 734.7 of the U.S. Export
Administration Regulations. As a result of this publication, this report is subject to only
copyright protection and does not require any license agreement from EPRI. This notice
supersedes the export control restrictions and any proprietary licensed material notices
embedded in the document prior to publication.

EPRI Project Manager
R. Adapa
ELECTRIC POWER RESEARCH INSTITUTE
3420 Hillview Avenue, Palo Alto, California 94304-1395 PO Box 10412, Palo Alto, California 94303-0813 USA
800.313.3774 650.855.2121 askepri@epri.com www.epri.com
Increased Power Flow Guidebook:
Increasing Power Flow in Transmission and
Substation Circuits
1010627
Final Report, November 2005

DISCLAIMER OF WARRANTIES AND LIMITATION OF LIABILITIES
THIS DOCUMENT WAS PREPARED BY THE ORGANIZATION(S) NAMED BELOW AS AN
ACCOUNT OF WORK SPONSORED OR COSPONSORED BY THE ELECTRIC POWER RESEARCH
INSTITUTE, INC. (EPRI). NEITHER EPRI, ANY MEMBER OF EPRI, ANY COSPONSOR, THE
ORGANIZATION(S) BELOW, NOR ANY PERSON ACTING ON BEHALF OF ANY OF THEM:
(A) MAKES ANY WARRANTY OR REPRESENTATION WHATSOEVER, EXPRESS OR IMPLIED, (I)
WITH RESPECT TO THE USE OF ANY INFORMATION, APPARATUS, METHOD, PROCESS, OR
SIMILAR ITEM DISCLOSED IN THIS DOCUMENT, INCLUDING MERCHANTABILITY AND FITNESS
FOR A PARTICULAR PURPOSE, OR (II) THAT SUCH USE DOES NOT INFRINGE ON OR
INTERFERE WITH PRIVATELY OWNED RIGHTS, INCLUDING ANY PARTY'S INTELLECTUAL
PROPERTY, OR (III) THAT THIS DOCUMENT IS SUITABLE TO ANY PARTICULAR USER'S
CIRCUMSTANCE; OR
(B) ASSUMES RESPONSIBILITY FOR ANY DAMAGES OR OTHER LIABILITY WHATSOEVER
(INCLUDING ANY CONSEQUENTIAL DAMAGES, EVEN IF EPRI OR ANY EPRI REPRESENTATIVE
HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES) RESULTING FROM YOUR
SELECTION OR USE OF THIS DOCUMENT OR ANY INFORMATION, APPARATUS, METHOD,
PROCESS, OR SIMILAR ITEM DISCLOSED IN THIS DOCUMENT.
ORGANIZATION(S) THAT PREPARED THIS DOCUMENT
EPRI
Power Delivery Consultants, Inc.

NOTE
For further information about EPRI, call the EPRI Customer Assistance Center at (800) 313-3774 or
email askepri@epri.com
Electric Power Research Institute and EPRI are registered service marks of the Electric Power
Research Institute, Inc.
Copyright 2005 Electric Power Research Institute, Inc. All rights reserved.

iii
CITATIONS
This report was prepared by
EPRI
115 East New Lenox Road
Lenox, MA 01240
Principal Investigator
B. Clairmont
Power Delivery Consultants Inc.
28 Lundy Lane, Suite 102
Ballston Lake, NY 12019
Principal Investigators
D. A. Douglass
E. C. Bascom, III
T. C. Raymond
J. Stewart
59 St. Stephens Lane N
Scotia, NY 12302
This report describes research sponsored by the Electric Power Research Institute (EPRI).
The report is a corporate document that should be cited in the literature in the following manner:
Increased Power Flow Guidebook: Increasing Power Flow on Transmission and Substation
Circuits. EPRI, Palo Alto, CA: 2005. 1010627.

v
PRODUCT DESCRIPTION

The Increased Power Flow (IPF) Guidebook is a state-of-the-art and best practices guidebook
on increasing power flow capacities of existing overhead transmission lines, underground cables,
power transformers, and substation equipment without compromising safety and reliability. The
Guidebook discusses power system concerns and limiting conditions to increasing capacity,
reviews available technology options and methods, illustrates alternatives with case studies, and
analyzes costs and benefits of different approaches.
Results & Findings
The IPF Guidebook clearly identifies those cases where increasing power flow might be an
alternative to upgrading the grid with major investment. The document reviews both established
technologies and new developments in technologies with the potential to increase power flow
and addresses how to apply them for lines, cables, and substations. Because the guide compares
the economic benefits of each available technology, it will assist utilities in making informed
decisions in terms of what options for IPF are available and which options are most economical
for application at their utility sites. By implementing one or more of the IPF technologies,
utilities can obtain increased asset utilization with minimal cost. For example, if a utility decides
to implement one of the IPF technologies, such as Dynamic Thermal Circuit Rating (DTCR)
technology, that implementation will allow increased power flows on the order of 15-20% over
the existing static ratings and, thus, increase utility revenue.
Challenges & Objective(s)
Motivations for increasing power flow limits on existing transmission facilities (rather than
constructing new facilities) are economic, environmental, and practical. Due to limited incentives
for new construction and time delays that may result from public opposition to new power
facilities, utilities around the world are being forced to find new ways of relieving modest
constraints or increasing power flow through existing transmission corridors with minimal
investment. This Guidebook will be an excellent reference document for transmission and
substation engineers since it provides all possible IPF options in one place for ease of use.
Applications, Values & Use
Training materials will be developed with the IPF Guidebook, including hands-on workshops at
EPRI's full-scale laboratories. In addition, it is anticipated that, in the coming years, technical
reports will be produced annually on new and updated aspects of IPF as well as new material on
costing, economics, power storage, voltage upgrading, and case studies. This information will be
incorporated into the Guidebook during subsequent work.
vi
EPRI Perspective
Due to limited incentives for new construction, utilities around the world are undergoing a major
transformation that is redefining the use of existing power equipment in the electric transmission
network. Under these circumstances, utilities are forced to find new ways of increasing power
flow through the existing transmission corridors with minimal investments. EPRIs Increased
Transmission Capacity Program directly responds to the needs of owners and operators of the
transmission grid to get the most out of existing equipment in todays competitive electricity
business while increasing the availability and reliability of transmission and substation
equipment. This Program helps customers accomplish these goals strategically, without
jeopardizing reliability and driving up costs. A number of projects have been undertaken by
EPRI in this Program under the Project Set 38AIncrease Power Flow Capability in the
Transmission Systems. One major effort under Project Set 38A is the IPF Guidebook, the only
available utility compendium of best practices for increasing power flow in transmission
circuits. Other major EPRI developments include DTCR (Dynamic Thermal Circuit Ratings)
software to calculate dynamic ratings of transmission circuits and Video Sagometer to measure
sag of transmission lines.
Approach
The Guidebook was developed by industry experts and draws on a combination of technology,
documented case studies, and associated engineering and safety guidelines.
Keywords
Overhead transmission
Substations
Transmission capacity
Underground transmission
vii
ABSTRACT
The Increased Power Flow (IPF) Guidebook documents the state-of-science for increasing
power flow capacities of existing overhead transmission lines, cables, and substation equipment.
The Guidebook provides an overview of the electrical, mechanical, thermal, and system concerns
that are important to increased power flow, presents all possible IPF options, uses case studies to
illustrate the options, and compares their potential economic benefits. The Guidebook also
provides an overview of dynamic thermal rating methods and summarizes other developments in
hardware and software that are instrumental for IPF.
The IPF Guidebook provides utilities with the only available compendium of best practices for
increasing power flow. The Guidebook will assist utilities in making informed decisions in terms
of what options for IPF are available and which options are most economical for application at
their utility sites.

Contents
Chapter 1 Increased Power Flow Fundamentals and
Principles
1.1 INTRODUCTION 1-1
1.2 POWER SYSTEM ISSUES 1-2
1.3 LIMITING CONDITIONS 1-3
Circuit Power Flow Limits 1-3
Surge Impedance Loading of Lines 1-4
Voltage Drop Limitations 1-5
Thermal Limits 1-6
Environmental Limits 1-7
Examples Overhead Lines 1-8
1.4 CHAPTER PREVIEW 1-9
Overhead Lines (Chapter 2) 1-9
Underground Cables (Chapter 3) 1-10
Power Transformers (Chapter 4) 1-10
Substation Terminal Equipment
(Chapter 5) 1-10
Dynamic Rating and Monitoring
(Chapter 6) 1-10
REFERENCES 1-11
Chapter 2 Overhead Transmission Lines
Surge Impedance Loading 2-2
Voltage Drop 2-2
Thermal Limits 2-2
Environmental Limits 2-2
2.2 UPRATING CONSTRAINTS 2-3
Introduction 2-3
Sag-tension Calculations 2-3
Limiting High Temperature Sag 2-5
Uprating Constraints Related to Wind-Induced
Conductor Motion 2-8
Electrical Clearance 2-10
Loss of Conductor Strength 2-12
Constraints on Structural Loads 2-13
Environmental Effects 2-15
2.3 LINE THERMAL RATINGS 2-15
Introduction 2-15
Maximum Conductor Temperature 2-16
Weather Conditions for Rating Calculation 2-16
How Line Design Temperature Affects Line Ratings 2-17
Heat Balance Methods 2-17
Thermal Ratings
Dependence on Weather Conditions 2-21
Transient Thermal Ratings 2-22
2.4 EFFECTS OF HIGH-TEMPERATURE OPERATIONS 2-23
Introduction 2-23
Annealing of Aluminum and Copper 2-23
Sag Tension Models for ACSR Conductors 2-27
Axial Compressive Stresses 2-28
BuiltIn Stresses 2-29
Sag Tension Calculations 2-29
Sag and Tension of Inclined Spans 2-31
Calculation of Conductor High-Temperature Sag
and Tension 2-32
Results of High-Temperature Sag Tension
Calculations 2-34
Effects of Wind Speed on Thermal Ratings 2-40
Thermal Elongation 2-41
Creep Elongation 2-42
Connectors at High Temperature 2-46
Conductor Hardware 2-49
2.5 UPRATING WITHOUT RECONDUCTORING 2-51
Introduction 2-51
Deterministic Methods 2-51
Probabilistic Methods 2-56
Development of a Measure of Safety as a Basis
for Line Rating 2-60
Comparison of Probabilistic Rating Methods 2-62
Device for Mitigating Line Sag - SLiM 2-62
2.6 RECONDUCTORING WITHOUT STRUCTURAL
MODIFICATIONS 2-65
Introduction 2-65
TW Aluminum Wires ACSR/TW or AAC/TW 2-66
ACSS and ACSS/TW 2-67
High-Temperature Aluminum Alloy Conductors 2-70
Special Invar Steel Core 2-70
Gapped Construction 2-71
ACCR Conductor 2-74
Conductors with Exotic Cores 2-74
Comparing ACSS and High-Temperature Alloy
Conductors 2-74
2.7 DYNAMIC MONITORING AND LINE RATING 2-75
Introduction 2-75
Dynamic Ratings Versus Static Ratings 2-75
Advantages of Dynamic Rating 2-76
Disadvantages of Dynamic Rating 2-76
Real-time Monitors 2-77
Dynamic Rating Calculations 2-79
Field Test Results 2-82
Summary 2-84
2.8 CASE STUDIES 2-84
Introduction 2-84
Selecting a Line Uprating Method 2-84
Preliminary Selection of Uprating Methods 2-85
Uprating Test CasesPreliminary Uprating Study 2-87
Economic Comparison of Line Uprating Alternatives 2-95
ix
Contents Increased Power Flow Guidebook
x
Detailed Comparison of Uprating Alternatives
An Example 2-99
Conclusions 2-103
2.9 REFERENCES 2-104
Chapter 3 Underground Cables
3.2 CABLE SYSTEM TYPES 3-2
High-Pressure Pipe-Type (Fluid- and Gas-Filled) 3-2
Extruded Dielectric 3-5
Self-Contained Liquid-Filled (SCLF) 3-8
Other Cable Types 3-10
3.3 POWER FLOW LIMITS AND SYSTEM
CONSIDERATIONS 3-11
Thermal, Stability, and Surge Impedance
Loading Limits 3-11
Load Flow Considerations 3-14
Uprating Hybrid (Underground and Overhead)
Circuits 3-14
3.4 UNDERGROUND CABLE RATINGS 3-15
Introduction 3-15
Concept of Ampacity 3-15
Losses 3-16
Equivalent Thermal Circuit and Thermal
Resistances 3-19
Calculating Ampacity 3-23
Effect of Various Parameters on Ampacity 3-24
Emergency Ratings 3-25
Inferring Conductor Temperatures from Measured
Temperatures 3-26
3.5 UPRATING AND UPGRADING CONSTRAINTS 3-26
Direct Buried Cable Systems 3-26
Fluid-Filled Cable Systems 3-26
Duct Bank Installations 3-27
Trenchless Installations 3-27
Other Installation Locations 3-27
Hot Spot Identification 3-28
Accessories 3-28
Hydraulic Circuit 3-28
3.6 INCREASING THE AMPACITY OF UNDERGROUND
CABLES 3-28
Route Thermal Survey 3-28
Review Circuit Plan and Profile 3-36
Evaluate Daily, Seasonal, or Other Periodic Load
Patterns 3-36
Temperature Monitoring 3-38
Ampacity Audit 3-40
Remediation of Hot Spots 3-40
Active Uprating 3-40
Shield/Sheath Bonding Scheme 3-43
3.7 RECONDUCTORING (UPGRADING) 3-43
Introduction 3-43
Larger Conductor Sizes 3-44
Cupric Oxide Strand Coating 3-44
Voltage Upgrading 3-45
Superconducting Cables 3-45
3.8 DYNAMIC RATINGS OF UNDERGROUND CABLE
SYSTEMS 3-46
Background 3-46
EPRI Dynamic Ratings on Cables 3-46
Benefit of Dynamic Ratings 3-49
Required Monitoring 3-51
Quasi-Dynamic (Real-Time) Ratings 3-51
3.9 CASE STUDIES FOR UNDERGROUND CABLE
CIRCUITS 3-51
CenterPoint Energy 3-51
United Illuminating Company 3-53
3.10 SUMMARY OF UPRATING AND UPGRADING
APPROACHES AND ECONOMIC FACTORS 3-55
REFERENCES 3-56
Appendix 3.1 Pipe-Type Ampacity Example 3-58
Appendix 3.2 Extruded Ampacity Example 3-65
Chapter 4 Power Transformers
4.2 TRANSFORMER DESIGN 4-2
General Construction 4-2
Types of Cooling 4-5
Losses 4-5
Factory Testing 4-6
4.3 RISKS OF INCREASED LOADING 4-9
Short-Term Risks 4-9
Long-Term Risks 4-11
Additional Risks 4-18
4.4 THERMAL MODELING 4-21
Mechanisms of Heat Transfer 4-21
Top Oil Model (IEEE C57.91-1995, Clause 7) 4-24
Bottom Oil Model (IEEE C57.91-1995, Annex G) 4-26
IEC Model (IEC 354-1991) 4-30
Proposed IEC Model 4-32
4.5 THERMAL RATINGS 4-33
Ambient Air Temperature 4-34
Load 4-34
Rating Type and Duration 4-35
Rating Procedure 4-35
Condition-Based Loading 4-36
Maintenance Considerations 4-37
4.6 WINDING TEMPERATURE MEASUREMENT 4-38
4.7 MODEST INCREASES IN CAPACITY FROM
EXISTING TRANSFORMERS 4-39
4.8 EXAMPLES 4-39
REFERENCES 4-43
xi
Contents Increased Power Flow Guidebook
Chapter 5 Substation Terminal Equipment
5.2 SUMMARYEQUIPMENT TYPES AND IPF
OPPORTUNITIES 5-2
Equipment Rating Parameters 5-2
Thermal Rating Parameter Comparison 5-4
5.3 THERMAL MODELS FOR TERMINAL EQUIPMENT 5-4
Bus Conductors 5-4
Switch (Air Disconnect) 5-6
Air-core Reactor 5-8
Oil Circuit Breaker 5-9
SF
6
Circuit Breaker 5-10
Bushings (Oil-immersed Equipment Only) 5-10
Current Transformers 5-11
Line Traps 5-11
Other Types of Terminal Equipment 5-12
5.4 UPRATING OF SUBSTATION TERMINAL
EQUIPMENT 5-12
Monitoring and Communications 5-13
Maintenance and Inspection Procedures 5-13
Reliability and Consequences of Failure 5-13
5.5 THERMAL PARAMETERS FOR TERMINAL
EQUIPMENT 5-14
Manufacturer Test Report Data 5-14
5.6 CONCLUSIONS AND SUMMARY 5-14
REFERENCES 5-15
Chapter 6 Dynamic Thermal Ratings Monitors and
Calculation Methods
6.2 ISSUES RELATED TO DYNAMIC THERMAL RATING
METHODS 6-2
Where Should Dynamic Thermal Circuit Rating
Calculations Be Performed? 6-2
CostsCapital and Otherwise 6-3
Why Dynamic Ratings Go With Increased Utilization 6-4
6.3 POWER EQUIPMENT CONDITION ASSESSMENT
AND REAL-TIME MONITORS 6-4
6.4 DYNAMIC THERMAL RATING MODELS FOR POWER
EQUIPMENT 6-5
Accounting for Heat Storage (Pre-load Monitoring) 6-5
Overhead Lines 6-6
Power Transformers 6-7
Underground Cables 6-10
Substation Terminal Equipment 6-11
6.5 EPRI'S DTCR TECHNOLOGY 6-13
Power Circuit Modeling 6-13
DTCR Output 6-13
DTCR is a Calculation Engine for SCADA 6-14
Modeling Complex InterfacesCalifornia Path 15 6-14
Conclusions about the Dynamic Rating of Complex
Interfaces 6-16
6.6 OPERATING WITH DYNAMIC THERMAL
RATINGS 6-16
Traditional Rating Definitions 6-16
Traditional Operating Rules 6-17
Operating with Dynamic Ratings 6-17
6.7 FIELD STUDIES OF DYNAMIC RATINGS 6-19
Overhead Lines 6-19
Power Transformers 6-20
Underground Cables 6-20
Substation Terminal Equipment 6-20
Power Circuits 6-20
Communications and Monitoring 6-21
6.8 CONCLUSIONS 6-21
REFERENCES 6-22
Glossary G-1
Index I-1

1-1
CHAPTER 1 Increased Power Flow
Fundamentals and Principles
1.1 INTRODUCTION
The purpose of this guidebook is to provide technical information and explain concepts
that may aid power transmission company technical personnel in finding economic, tech-
nically sound ways to increase the power flow capacity of existing circuits without com-
promising safety or reliability.
The motivations for increasing the power flow limits on existing transmission facilities
(rather than constructing new facilities) are economic, environmental, and practical. The
methods discussed are generally modest in costranging from virtually free to about
30% of the cost of equipment replacement. The corresponding increase in equipment rat-
ing is similarly modest, usually between 5% and 30% (with the exception of overhead
lines where reconductoring may yield an increase of over 100%). The methods are practi-
cal since the environmental and/or visual impact is normally low, regulatory approval and
public hearings may not be needed, and extended power outages are often avoided.
Given the extended time delays that may result from public opposition to the construc-
tion of new power transmission facilities or even to any visible, physical modification of
existing facilities, the use of increased power flow (IPF) methods may offer the only prac-
tical solution to relieving modest constraints on power flow.
Determining the degree to which maximum power flow constraints can be eased on existing
power equipment (overhead lines, power transformers, etc.), power circuits (multiple power
equipment elements in series), and power system interfaces (multiple parallel power cir-
cuits connecting power system regions) can be quite complex. For example, consider the fol-
lowing:
For an overhead line, any increase in power flow capacity is dependent on its length, the
original design assumptions, present-day environmental concerns, the condition of its
existing structures, and the type of conductors originally selected. Depending on these
multiple factors and which of the IPF methods suggested in Chapter 2 is applied, the
resulting increase in the lines thermal rating could be as little as 5% or as much as 100%.
But overhead lines are only part of the transmission path (circuit). The lines are termi-
nated at substations by air disconnects, circuit breakers, line traps, etc. The power flow
through all of the circuit elements must be limited to avoid damaging the line or the termi-
nating equipment, and the maximum allowable power flow over this circuit may be limited
by any one of the circuit elements.
Finally, when seen as part of a power system interface, any increase in maximum allow-
able power flow through any component circuit or circuit element does not necessarily
yield a higher rating for the complex interface.
Chapter 1: Increased Power Flow Fundamentals and Principles Increased Power Flow Guidebook
1-2
In general, it may be stated that maximum power flow on
the transmission system is a function of the overall system
topology (transmission lines, transformers, generation,
series and shunt compensation, and load), and that many
non-thermal system considerations can also limit the
maximum power flow on a specific transmission circuit.
Therefore, transmission circuit ratings are often developed
on a system basis, rather than on an individual line basis.
The overall limit may be between operating areas irre-
spective of ownership or individual lines, and may change
during a day based on system conditions.
Chapter 1 provides an overview of the electrical,
mechanical, thermal, and system concerns that are
important to increased power flow. The chapter includes
three sections:
Section 1.2, Power System Issues, presents a simple
power flow example to illustrate several principles
about increasing power flow.
Section 1.3, Limiting Conditions, describes limits on
power flow imposed by circuit power flow, surge
impedance loading, voltage drop, thermal factors,
and environmental constraints.
Section 1.4, Chapter Previews, presents brief descrip-
tions of the chapters in this guidebook.
1.2 POWER SYSTEM ISSUES
The power transmission system, in any region, is a com-
plex combination of lines (including underground cable)
and substations. With the exception of relatively short
radial lines connecting generating stations to the sys-
tem, power flow reaching any load point in the system
flows over multiple parallel paths (circuits). In any
path (circuit), the power flow moves through multiple
series elements.
This can be illustrated by the following simple power
system (NERC 1995) shown in Figure 1.2-1. There are
three load areas (A, B, and C). Each load area has suffi-
cient generation to supply the local load. With the sys-
tem operating normally, there is no net power transfer
between load areas. Nonetheless, as a result of the avail-
able electrical paths connecting the load areas, the dia-
gram shows a loop flow of 200 MW. This loop flow
occurs even though there is no net power transfer to any
of the areas.
Consider the situation where power generated in load
area A is considerably less expensive than local genera-
tion in load area B. It would then be advantageous for
power customers in load area B to buy power from the
generators in load area A. In doing this, the power
transmission system operator sets a transfer limit of
2834 MW from A to B. Given this level of transfer, the
power flows would be as shown in Figure 1.2-2.
Notice in Figure 1.2-2 that, even though load area C is
not importing power, the lines connecting load area C to
the other areas are carrying almost half of the total
power transferred.
Now let us assume that the customers in load area B
would like to buy even more than 2834 MW from the
low-cost generator in load area A. Consequently, they
contest the limit of 2834 MW set by the system opera-
tor, noting that the emergency rating of the lines is 1000
MW. The power system operator explains that the limi-
tation on power import to load area B is not due to nor-
Figure 1.2-1 Base system operating
"normally" with local generation being
similar in cost and able to supply all local
load.
Figure 1.2-2 Base system operating
"normally" with local generation at A
being much cheaper than at B,
causing a net power transfer of 2834
MW.
1-3
Increased Power Flow Guidebook Chapter 1: Increased Power Flow Fundamentals and Principles
mal power flows but rather to the emergency power flow
through one of the lines (#2) between A and C, as
shown in Figure 1.2-3! With line #1 out of service, the
redistributed power flow through line #2 reaches the
emergency thermal limit of 1000 MW. Thus the imposed
power transfer limit of 2834 MW from A to B.
As shown in Figure 1.2-3, if the net power transfer from
A to B with all lines in service had exceeded the transfer
limit of 2834 MW then, under this single contingency
loss of line #1 between A and C, the power flow in line
#2 from A to C would have exceeded the lines emer-
gency thermal limit.
One can reach a number of conclusions regarding power
flow limits from this simple example:
Economic power transfers can be limited by circuits
that do not directly connect the low-cost generation
source and the customer.
A 5% increase (50 MW) in the emergency rating of
lines #1 and #2 connecting A and C from 1000 MW
to 1050 MW might allow a similar 5% increase
(141 MW) in the power transfer limit from 2834 to
2975 MW.
A 5% increase in the emergency rating of either line
#1 or #2 between A and B would not allow any
increase in power transfer from A to B.
The long-term value of projects to increase the
power flow in any particular circuit is dependent on
changes in the cost of generation and the power flow
limits and electrical impedance of interconnected
power circuits.
Note that these observations do not depend on the
reason for the power flow limit in any of the circuits.
They would be equally valid whether the limitation
on power flow is due to equipment temperature lim-
its, limits on voltage drop, or electrical phase shift
stability issues.
1.3 LIMITING CONDITIONS
1.3.1 Circuit Power Flow Limits
Power circuits consist of series and parallel combina-
tions of electrical equipment (each subjected to mechan-
ical, electrical, and thermal stresses) whose collective
purpose is to transmit power safely and reliably under
widely varying operational situations. Each element of
such circuits is typically specified to have certain power
flow limits that allow their safe, reliable operation for an
extended period of time (e.g., 40 years).
Increased power flow inevitably means increased electri-
cal current flow or increased circuit voltage since power
is the product of these quantities. In general, for substa-
tion equipment and underground cables, increasing the
operating voltage is difficult or impossible, whereas
increasing the maximum electrical current is both possi-
ble and economic. Overhead lines are often capable of
either higher voltage or higher current levels if certain
modifications are undertaken.
Power transmission circuits are typically bimodal in
terms of power flow. Under normal operation, it is not
unusual for power transformers and lines to operate at
much less than half of their power flow capacity, only
approaching their operational limits under relatively
rare emergency events.
There are basically three methods of increasing power
flow: load control, improved modeling and monitoring,
and physical modification of existing circuits.
Improved models may allow operation of equipment
with reduced safety factors without reducing safety and
reliability. Examples are the bottom oil model in
Annex G of the IEEE loading guide and the improved
models for high-temperature sag of ACSR conductor.
Similarly, monitoring of environmental factors (air tem-
perature, wind speed, humidity, etc.) may allow the use
of less conservative assumptions, again without reduc-
ing safety and reliability.
Figure 1.2-3 Base system operating in
response to a single contingency
outage of line #1 between A and C
while there is a power transfer of 2834
MW from A to B.
1-4
With monitors communicating data in real-time, it may
be possible to run equipment at higher power levels most
of the time by avoiding the use of worst case assump-
tions. This approach is called dynamic thermal ratings. It
is unlikely that such real-time monitoring would allow
any increase in non-thermal operating limits.
Overhead transmission lines are the primary means of
power transfer over long distances. They have thermal
ratings just as power transformers, substation terminal
equipment, and underground cables but, for long lines,
power flow limits may also be necessary to avoid exces-
sive voltage drop or system stability problems. In addi-
tion, since the public has access to the area under lines,
there may also be limits on voltage and current related
to environmental effects. This section concerns the rela-
tionship between the various types of power flow limits
for overhead lines.
1.3.2 Surge Impedance Loading of Lines
Sometimes a power transmission line possesses a defi-
nite power flow limit based on the design parameters for
the specific line, but at other times, the line as a compo-
nent of the overall transmission system determines the
limit. System limits can result from factors such as volt-
age drop, possibility of voltage collapse, and system sta-
bility, both steady state and transient.
System limits are functions of transmission line reac-
tances in relation to the overall power system. Series
reactance, shunt admittance, and their combination,
surge impedance, are relevant to system transfer limits.
System planners have long recognized this relationship,
particularly where there are prospects of changing the
line surge impedance, either by adding equipment (e.g.,
series capacitors) or by modifying the line itself (e.g.,
reconductoring, voltage uprating, etc.).
Transmission line series inductive reactance is deter-
mined by conductor size, phase spacing, number of con-
ductors, relative phasing (double-circuit lines), and line
configuration. In transmission lines the series reactance
is significantly larger than the series resistance, and is
the dominant factor in a first-order explanation of sys-
tem behavior. For this reason, simple reconductoring of
a transmission line results in only minor changes in sys-
tem power flows.
Power flow on a transmission line, neglecting resistance
of the line, is given by Equation 1.3-1, which can be
derived from a simple circuit consisting of sending and
receiving end voltage sources connected by a series reac-
tance.
1.3-1
Where:
P =Real power transfer on the transmission line.
V
1
= Magnitude of sending end bus voltage.
V
2
= Magnitude of receiving end bus voltage.
X =Line series inductive reactance between V
1
and
V
2
.
= Phase angle difference between V
1
and V
2
.
Increasing voltage magnitude for the same line voltage
and same phase difference between ends increases the
power flow. By increasing the voltages V
1
and V
2
together, the power transmitted increases by the square
of the voltage for the same phase angle. Power flow
increases for the same end voltage magnitudes are
accommodated by an increase in the phase angle differ-
ence between the voltages at the two line ends.
Equation 1.3-1 imposes a fundamental limit on the
amount of power that can be carried by a transmission
line corresponding to a phase difference between line
ends of 90. Further increases i n angl e resul t in
decreases in power flow. This is an unstable situation
that can be realized in practice in two ways. If the
steady-state power flow were to slowly increase to the
point that the angle reached 90, an attempt to further
increase power flow would actually decrease the power
flow. An increase in the power angle when is in the
range from 90 to 180 results in a decrease in sin() and
a consequent decrease in power flow. The condition try-
ing to increase the flow on the line actually results in a
decreased flow, and system instability.
Secondly, a system disturbancefor example, tripping
of a linecauses a redistribution of power flow among
the remaining lines, and consequent changes in the bus
voltage angles. It is insufficient that the new angle differ-
ences on all the lines are less than 90, because the angle
differences must remain lower than 90 during all the
transient system swinging from the time of the distur-
bance until the system settles in its new operating state.
If a line were to experience its angle difference momen-
tarily passing 90, it would try to accommodate the
power requirement by opening up the angle beyond 90,
decreasing the power flow. This is an unstable situation,
and would cause the line to pass through the electrical
point where its relay protection would sense a fault
(even though none exists on the line), and result in a line
trip and probable system separation.
Surge impedance loading (SIL), defined in Equation
1.3-2) provides a useful rule of thumb measure of trans-
1 2
sin( ) V V
P
X

=
1-5
mission line loading limitation as a result of the effects
of series reactance.
1.3-2
Where V is the line voltage, and Z
S
is the surge impedance
of the transmission line given by:
1.3-3
Surge impedance Z
S
is a resistance in ohms. L and C in
Equation 1.3-3 are positive sequence inductance and
capacitance in henries per mile and farads per mile,
respectively. Surge impedance loading is that loading on
a three-phase power transmission line that it would have
if it were loaded by a Y-connected set of resistances of
Z
S
ohms per phase. This is the same physical condition
as a radio frequency transmission line impedance
matched to its termination (72 ohm coaxial cable termi-
nated in 72 ohms in television cable). In electromagnetic
theory, it corresponds to a pure TEM wave. The reactive
power (vars) generated in the line capacitance is exactly
canceled by the vars absorbed in the line inductance in a
power transmission line at surge impedance loading
(neglecting line resistance and real power losses). Surge
impedance loading thus is a loading value based on
physical principles related to the line design itself.
Surge impedance loading is a handy tool for estimating
the relative loading capabilities of lines of different volt-
ages, constructions, and lengths from a system stand-
point (St. Clair 1953). SIL is oversimplified for use in
specifying actual line ratings on an operating system.
However, it is a useful guide both for assessing actual
loading limits and for understanding the different fac-
tors that limit line loading. Figure 1.3-1 gives a curve of
line loadability in per unit of SIL as a function of line
length for heavy loading conditions. Slightly different
versions of Figure 1.3-1 have been published, but they
are all very similar (Dunlop et al. 1979, Gutman 1988).
The fundamental observation from Figure 1.3-1 is that
transmission line loadability decreases as length of the
line increases. Three different regions come into play in
derivation of Figure 1.3-1. Short lines tend to be ther-
mally limited, irrespective of system conditions. As line
length increases, voltage drop considerations frequently
come into play. At longer line lengths, stability factors
may dominate. Short lines are often loaded at 2 or 2.5
times SIL and thus need reactive power (var) support to
maintain the voltage. Long lines may be limited to 1.0
times SIL or less.
An important observation from Equation 1.3-2 is that
surge impedance loading is a function of the square of
line voltage. This has been a driving force in increased
transmission voltages over the years, especially for
longer lines.
For an overhead transmission line, typical surge imped-
ance is on the order of 300 ohms, while for a cable it
may be 50 ohms or less. At 345 kV, SIL of an overhead
line is on the order of 400 MW. Short lines may be able
to carry 800 MW or more, while long lines of exactly the
same construction may be limited to less than 400 MW
by system considerations. Because of limitations on heat
dissipation, underground transmission cables always
operate very far below SIL. A consequence is that
underground transmission cables are a net source of
vars to the system, a condition that must be considered
in system design.
1.3.3 Voltage Drop Limitations
Voltage control on the power system is of concern as
system loadings increase. The system voltage distribu-
tion is affected by the series inductance and shunt
capacitance of the transmission lines, and is related to
the flow of reactive power in the system. Depending on
the relative real and reactive power flow on a given
transmission line, the voltage may increase or decrease
from one end to the other. It is not desirable for voltage
to vary more than 5%, or at most 10%, from one end to
2
S
V
SIL
Z
=
S
L
Z
C
=
Figure 1.3-1 Line loadability in terms of surge
impedance loading (Dunlop et al. 1979).
1-6
the other. In some cases, a voltage drop limit is placed
on power flow corresponding to the maximum allowable
decrease in voltage magnitude. The longer the line or
cable, generally the lower the power flow required to
reach a voltage drop limit. Voltage control is a system
problem, and is not generally solved by modifications to
any one transmission circuit.
Methods to improve voltage control on transmission
circuits may take a variety of forms:
1. In some cases, bundled conductors have been in over-
head lines used for short lower voltage lines to reduce
series reactance, where the use of bundled conductors
is required neither for thermal or corona reasons.
2. Supply of vars at various points on the system can be
used to control voltage. The supply can be fixed,
switched, or adjustable. In former years synchronous
condensers were used to supply vars in a continu-
ously adjustable basis. Capacitor banks are com-
monly used, and may be switched on or off
depending on the local voltage. Static var compensa-
tors (SVCs) are also used to control voltage on the
bulk power system.
3. Shunt reactors may be used for long EHV lines where
the var supply from the line capacitance is greater
than the system can absorb.
Because voltage drop is primarily a function of line
reactance rather than resistance, simple reconductoring
does very little to decrease the voltage drop per unit
length. Reconductoring an existing 230-kV line by
replacing the original 636 kcmil Hawk ACSR with a 954
kcmil Rail ACSR only increases the voltage drop limit
by 5%. For an overhead line, adding a second conductor
per phase to form two conductor bundles results in a
more significant reduction in series reactance, and a
greater improvement in voltage drop power limit.
Shunt reactors may be applied for reasons other than
voltage controlfor example, to control transient over-
voltages during line switching. Series capacitors may be
used to partially compensate for the line series reac-
tance, but this is usually reserved for the longest lines in
relation to system stability. Whenever capacitors are
installed in series with the transmission line inductance,
the possibility of a series resonant condition exists. Sub-
synchronous resonance has been the cause of genera-
tor/turbine shaft failure and is a serious consideration
for a series capacitor installation.
Other problems present themselves with series capaci-
torsfor example, provision for passage of fault current
without causing failure of the capacitors. The overall
effect of the concern with system voltage is that a partic-
ular transmission line may be limited in its power-han-
dl i ng capaci ty by system vol tages and var fl ows
irrespective of the thermal capacity of the line conduc-
tors. In some cases, it is possible to increase the line
flows by addition of capacitors or similar measures.
Flexible AC Transmission (FACTS) is a scheme where
thyristor-controlled devices are arranged to provide real-
time control of transmission line flows in excess of those
that would normally be allowed by system voltage and
stability considerations.
While voltage drop has long been known as a transmis-
sion limitation, attention has also been focused in more
recent years on voltage collapse, which is a system insta-
bility that can occur under heavy loading conditions.
Figure 1.3-2 shows a voltage collapse condition follow-
ing a system disturbance, where the 115-kV voltage
drops to 50% of the nominal operating voltage (0.5 p.u.).
Voltage collapse can occur for several reasons on a
heavily loaded system where there is insufficient var sup-
port. An example is the geomagnetic storm of March 13,
1989, with its resulting voltage collapse and blackout.
The March 1989 storm increased attention to system
problems that result from solar activity. Utilities in areas
subject to geomagnetic disturbances monitor solar activ-
ity (Lesher et al. 1994), and can re-dispatch generation
to reduce loading on affected lines during times of high
geomagnetic activity. However, geomagnetic distur-
bances are not the only cause of voltage collapse.
1.3.4 Thermal Limits
Thermal limits are discussed in considerable detail later
in this guidebook (see, for example, Section 2.3). In
brief, the current-carrying capacity (thermal rating) of
an overhead transmission circuit is determined by the
assumed worst-case weather conditions, assumed
Figure 1.3-2 Voltage collapse condition following a
system disturbance.
1-7
conductor parameters, and the maximum allowable
conductor temperature. Some of the specific thermal
rating parameters are:
Conductor construction: outside diameter, conductor
strand diameter, core strand diameter, number of
conductor strands, and number of core strands.
Conductor AC resistance, which itself is a function of
the conductor temperature.
Conductor surface condition: solar absorptivity and
emissivity.
Line location: latitude, longitude, conductor inclina-
tion, conductor azimuth, and elevation above sea
level.
Weather: incident solar flux, air temperature, wind
speed, and wind direction.
The temperatures experienced by terminal equipment
must also be limited. In certain circuits, the thermal rat-
ing of substation equipment, in series with an overhead
line, may determine the circuit rating. Disconnect
switches, wave traps, current transformers, and other
substation equipment all have current ratings that can
be lower than those of the line. An example of terminal
equipment limitations on older lines is 600 A disconnect
switches. At EHV, bundled conductors are employed to
reduce the conductor surface electric field and conse-
quent corona phenomena of radio, television, and audi-
bl e noi se. Bundl ed conduct ors were ori gi nal l y
introduced to lower line reactance and increase the line
loadability, and their use for noise reduction was recog-
nized later. Especially at the higher transmission volt-
ages of 500 and 765 kV, the thermal current-handling
capacity of a bundled conductor may be far in excess of
the ratings of the circuit breakers. In such cases, the
thermal limit of the circuit is entirely dominated by the
terminal equipment. A survey of utility 345-kV circuit
thermal limits in New York State gave the following lim-
itations:
41% of the circuits were limited by the line or cable.
18% of the circuits were limited by current transform-
ers.
4% of the circuits were limited by wave traps.
4% of the circuits were limited by the bus-work.
3% of the circuits were limited by disconnect
switches.
4% of the circuits were limited by the circuit breakers.
Lines and substation equipment may have different
thermal ratings for normal and for emergency system
conditions. Emergency ratings typically apply for a lim-
ited period of time, not exceeding 24 hours and as short
as 5 minutes. Emergency ratings are typically calculated
for higher temperatures, and allow for some equipment
deterioration in order to avoid load interruptions under
unusual operating conditions.
Broader voltage tolerances may also be appropriate
under contingency conditions compared to normal
operation. Lower voltage may be acceptable for a short
time. Likewise, conductor resistive power losses are
inconsequential during emergencies.
1.3.5 Environmental Limits
The electric field produced by overhead power transmis-
sion lines is influenced by the following factors:
Line voltage.
Height of conductors above ground.
Configuration of conductors (line geometry, con-
ductor spacing, relative phasing of multi-circuit lines,
and use of bundled conductors).
Lateral distance from the center line of the transmis-
sion line.
Height above ground at the point of field measure-
ment.
Proximity of conducting objects (trees, fences, build-
ings) and local terrain.
The electric field near ground level produced by an over-
head transmission line induces voltages and currents in
nearby conducting objects. These objects are typically
the size of people, animals, motor vehicles, sheds, and
similar-sized bodies. Electric field coupling is capacitive
coupling, and can be represented by a current source in
parallel with a high source impedance (Norton equiva-
lent circuit).
The allowable electric field is limited by the maximum
allowed induced current and voltage. For example, the
National Electrical Safety Code specifies a maximum of
5 mA short-circuit current induced into the largest vehi-
cle that could be stopped under the line, based on
human susceptibility to loss of muscular control (let-
go). Thus, if an existing line induces 4.9 mA on a large
tractor-trailer, it would not be possible to increase the
voltage without taking other measures to limit the elec-
tric field.
Electric field levels are limited by law in some jurisdic-
tions. Some regulations are specified at the edge of the
right-of-way for public exposure. Other regulations are
maximum levels on the right-of-way based on induction
to an assumed size object. These regulations may restrict
1-8
voltage increases on presently existing transmission lines
without taking electric field reduction measures.
Magnetic field is affected by the same variables as elec-
tric field, except line current replaces line voltage, and
nearby objects generally have minimal impact on the
magnetic field. Magnetic field coupling is generally of
significance for objects that parallel the transmission
line for a long distance. Such objects include pipelines,
telephone and railway signal circuits, and metal fences.
Because magnetic field is a function of line current, and
current increases during fault conditions, it may be nec-
essary to evaluate magnetic field effects under both nor-
mal operation and faults. Magnetic field coupling is
inductive coupling, and generally produces low voltages
with low source impedances.
Increasing current on a transmission line increases the
magnetic field, and thus increases magnetically induced
voltages and currents. This may be significant in cases
such as when a transmission line parallels a railroad
right-of-way. This is the inductive coordination problem
that has been around since the dawn of the power indus-
try with respect to telephone and railroad signal facili-
ties. Increasing current flow on existing lines may
require coordination with parallel infrastructure. In
some jurisdictions, maximum magnetic field levels are
specified by regulation. If an existing transmission line is
operating near the magnetic field limit set by regulation,
the ability to increase line current may be impaired,
unless measures are taken to reduce the magnetic field
levels.
Electric fields can be shielded by conducting objects.
Vegetation is sufficiently conductive to reduce electric
field levels. Grounded wires can be strung under the
phase conductors at road crossings to reduce electric
field levels. Grounding measures can be taken for fixed
objects to eliminate induced voltages. On the other
hand, magnetic field shielding is significantly more diffi-
cult than electric field shielding. Shielding a transmis-
sion line by magnetic materials is impractical. Flux
canceling loops have been developed, but incur power
loss and complexity in actively driven loops. Shielding is
less practical as a mitigation measure for magnetic fields
than it is for electric fields.
1.3.6 Examples Overhead Lines
Figure 1.3-3, adapted from (Gutman 1988), repeats the
generalized SIL curve of Figure 1.3-1 with the addition
of a curve for thermal limitation. Superimposed on the
SIL curve are curves for:
Thermal limit for a single 1414 kcmil conductor per
phase.
Voltage drop limitation of 5%.
Steady-state stability margin of 35%.
The thermal and voltage drop limitation curves cross at
a line length of approximately 110 miles. The voltage
drop and stability limitation curves cross at a line length
of approximately 190 miles. Based on the thermal, volt-
age, and stability curves, three regions are identified in
Figure 1.3-3.
Figure 1.3-3 Three line loading limits: thermal limit, voltage drop, and steady-state stability.
1-9
Less than approximately 110 miles line length, the
line is thermally limited.
Between approximately 110 and 190 miles line length,
the line is limited by voltage drop.
Beyond approximately 190 miles, the line is stability
limited.
Figure 1.3-3 thus illustrates the three regimes of line-
loading limits. Figure 1.3-3 further illustrates that, for a
specific transmission line example, the data points for
the line fall near, but not on, the generalized SIL curve.
The fact that the values are similar, but not identical,
illustrates the point that the SIL curve is a handy refer-
ence for sanity checking and rule of thumb analysis, but
is not to be considered exact for any specific line.
Sample surge impedance and thermal loading values for
transmission lines of different voltages are given in
Table 1.3-1.
For comparison with the 345-kV example in Figure 1.3-
3, the 230-kV example in Table 1.3-1 has a thermal rat-
ing of 440 MW and surge impedance loading of 145
MW. Stability and voltage control limits for this line
depend on the system to which it is connected. As an
example of voltage drop, assume the 230-kV line is 100
miles long. Further assume that the sending end bus has
a voltage of 1.0 per unit, and the receiving end bus has a
voltage of 0.95 per unit, a 5% difference. Also assume
the 230-kV line is at 1.0 power factor at the receiving
end, neither taking nor supplying vars to the bus. In this
case, the 230-kV line flow would be 220 MW, about 1.5
times SIL, but half the thermal rating. This result is in
line with the 345-kV example given in Figure 1.3-3.
Because SIL is primarily related to transmission line
series reactance rather than resistance, simple reconduc-
toring would produce only a minor effect on SIL limits
such as voltage drop. In this 230-kV example of 5% volt-
age drop, reconductoring from Cardinal to Falcon
ACSR would increase the loading from 220 MW to 230
MW, a minor difference. Adding a second Cardinal con-
ductor per phase to make two conductor bundles would
increase the loading to 310 MW. Adding a second con-
ductor per phase has a greater impact on surge imped-
ance, and thus on SIL and line loading. Full use of the
230-kV lines thermal rating would require system
changes to provide var support at the receiving end of
the line.
The thermal limit is determined by line current and line
voltage. Equation 1.3-2 shows that surge impedance
loading is proportional to the square of the line voltage.
Doubling line voltage doubles the thermal rating of the
line, but multiplies SIL by a factor of 4. This has been
the driving force during the history of the electric power
industry for increasing voltage levels, and sometimes a
motivation for voltage upgrades of existing lines.
1.4 CHAPTER PREVIEW
1.4.1 Overhead Lines (Chapter 2)
Overhead transmission lines are the predominant
method of transporting power in any but the most
urbanized power systems such as the New York City
area. Of all the types of power equipment, overhead
lines offer the largest opportunities for increased power
flow at modest cost. Limits are placed on power flow
through overhead lines in order to limit electrical phase
shift, avoid excessive voltage drop, and limit the temper-
ature of the current-carrying conductors. The emphasis
in this book is on the latter of these limits.
Chapter 2 discusses the reasons for limiting the temper-
ature of overhead lines and the consequences of exceed-
ing such limits. The chapter also covers the techniques
for modifying the clearance of existing lines, reconduc-
toring them without rebuilding structures, and real-time
monitoring of weather and line sag-tensions.
A number of interesting case studies are included at the
end of the chapter.
Table 1.3-1 Power Flow Limits on Lines and Cables
System XL XC
Surge
Impedance SIL
Thermal
Rating
kV (/mi) (/km) (M-mi) (M-km) () (MW) (MW)
Transmission Line Characteristics
230 0.75 0.47 0.18 0.29 367 145 440
345 0.60 0.37 0.15 0.24 300 400 1500
500 0.58 0.36 0.14 0.26 285 880 3000
765 0.56 0.35 0.14 0.26 280 2090 8000
Transmission Cable Characteristics
345 .25 .16 .0060 .0097 39 3050 2100
1-10
1.4.2 Underground Cables (Chapter 3)
Chapter 3 provides an overview on underground cable
systems and a very brief background on each of the
major transmission cable types. As with overhead lines,
the discussion on underground cable considers aspects
external to a specific cable circuit that may limit power
flow regardless of the cable circuits rating. The chapter
also includes an overview on cable system ampacity,
including worked examples.
The major barriers to increased underground cable rat-
ing are inherent to each cable system type or installation
location. Methods for increasing the rating of under-
ground cablesuch as surveying the soil thermal resis-
tivity along the route and removing thermal bottlenecks
due to other cable circuits or external heat sourcesare
discussed in some detail.
Given the relatively long thermal time constant of
underground cables, dynamic rating methods are very
attractive ways of increasing the rating. The chapter dis-
cusses monitoring methods and the necessary real-time
data required for dynamic rating calculations with
underground cable.
Case studies are included for actual cable uprating
projects, and the chapter provides a summary compari-
son of uprating methods.
1.4.3 Power Transformers (Chapter 4)
Power transformers represent a significant portion of
capital investment costs. Under existing conditions in
the industry, utility budgets are reduced and networks
are being forced to support greater power transfer over
existing transmission circuits than ever before. As such,
there is increased interest in safely utilizing all available
capacity of power transformers.
In general, transformer load capacity is limited by equip-
ment (winding and oil) temperatures. Industry standards
(IEEE C57.12.00 in the U.S.) specify a maximum average
winding rise that defines the rated load. In other words,
when operating at rated nameplate current, the average
winding rise shall not exceed the given value.
Chapter 4 describes the general construction of power
transformers, outlines short- and long-term risks related
to the loading of transformers, provides an overview of
heat transfer mechanisms and describes the four most
prevalent thermal models, and discusses factors behind
thermal ratings, including ambient air temperature,
load, and maintenance considerations.
1.4.4 Substation Terminal Equipment
(Chapter 5)
Substation terminal equipment consists of many differ-
ent types and designs of power equipment. Included in
this classification are line traps, oil circuit breakers, SF
6
circuit breakers, rigid tubular bus, line disconnects, cur-
rent transformers, bolted connectors, and insulator
bushings. The increase in circuit rating, resulting from
applying the various methods of increasing power flow
in overhead transmission lines, underground cable, and
power transformers is often limited by terminal equip-
ment. In some cases, a large increase in circuit rating
may be obtained for a very modest expenditure on ter-
minal equipment rather than a relatively large invest-
ment in lines, cables, or transformers.
Chapter 5 describes practical, rather simple methods of
increasing the power flow through less capital-intensive
equipment such as switches, bus, line traps, breakers,
and power transformer auxiliary equipment. The chap-
ter includes a summary of terminal equipment types,
specific thermal models for each type of equipment,
dynamic thermal rating of terminal equipment, and
methods of determining specific thermal parameters
from field test, laboratory test, and manufacturer heat-
run tests.
1.4.5 Dynamic Rating and Monitoring
(Chapter 6)
Since the mid-1980s, considerable attention has been
paid to increasing the power flow of overhead lines,
power transformers, underground cables, and substation
terminal equipment by means of monitoring weather
and the equipment thermal state and by developing
more accurate thermal models. The resulting dynamic
thermal rating techniques typically yield increases of 5
to 15% in capacity.
Chapter 6 provides an overview of dynamic thermal rat-
ing methods. The chapter aims to present a balanced
overall view of when dynamic rating methods are appro-
priate, how they are best implemented in a practical
operational application, and how such methods can be
applied to complex interconnections consisting of multi-
ple circuits and many circuit elements.
The chapter discusses concerns related to dynamic rat-
ings; outlines the need for inspections and/or real-time
monitors and the problems that may arise without them;
provides an overview on models for overhead lines, trans-
formers, underground cables, and substation terminal
equipment; describes the use of DTCR software; identi-
fies operating issues related to dynamic thermal ratings;
and describes field studies of dynamic ratings used for
overhead lines, transformers, underground cables, substa-
tion terminal equipment, and power circuits.
1-11
REFERENCES
Boteler, D. H. 1994. Geomagnetically Induced Cur-
rents: Present Knowledge and Future Research. IEEE
Transactions on Power Delivery. Volume 9. Number 1.
J anuary. pp. 50-58.
Dunlop, R. D., R. Gutman, and P. P. Marchenko. 1979.
Analytical Development of Loadability Characteristics
for EHV and UHV Transmission Lines. IEEE Transac-
tions on Power Apparatus and Systems. Volume 98. Num-
ber 1. March/April. pp. 606-617. correction May/J une.
page 699.
Federal Power Commission. 1964. National Power Sur-
vey. Part II-Advisory Reports. U. S. Government Print-
ing Office. Washington, D. C. October.
Gutman, R. 1988. Application of Line Loadability
Concepts to Operating Studies. IEEE Transactions on
Power Systems. Vol. 3. Number 4. November. pages
1426-1433.
Koessler, R. J. and J. W. Feltes. 1993. Voltage Collapse
Investigations with Time-Domain Simulation.
IEEE/NTUA Joint International Power Conference.
Athens Power Tech Proceedings. Athens, Greece. Sep-
tember 5-8.
Lesher, R. L., J. W. Porter, and R. T. Byerly. 1994. Sun-
burstA Network of GIC Monitoring Systems. IEEE
J anuary. pp. 128-137.
North American Electric Reliability Council (NERC).
1995. Transmission Transfer Capability.
St. Clair, H. P. 1953. Practical Concepts in Capability
and Performance of Transmission Lines. AIEE Trans-
actions on Power Apparatus and Systems. Volume 72. Part
III. December. pages 1152-1157.

2-1
CHAPTER 2 Overhead Transmission
Lines
2.1 INTRODUCTION
The degree to which the maximum power flow can be increased on an existing overhead
line depends on its length, the original design margins, environmental concerns, and
many other issues. Because power flow on the transmission system is a function of the
overall system topology (transmission lines, transformers, generation, series and shunt
compensation, and load), system considerations can also limit the maximum power flow
on a specific transmission line. Transmission line ratings are sometimes developed on a
system basis rather than on an individual line basis. The overall limit may be between
operating areas, irrespective of ownership or individual lines, and may change during a
day based on system conditions.
Sometimes a power transmission line possesses a definite power flow limit based on the
design parameters for the specific line; at other times the line as a component of the over-
all transmission system determines the limit. System limits can result from factors such as
voltage drop, possibility of voltage collapse, and system stability, both steady state and
transient.
Power system limits, on the power flow through individual overhead lines, are described
in more detail in Chapter 1, which discusses power system limits on increased power flow.
System limits are functions of transmission line reactances in relation to the overall
power system. Series reactance, shunt admittance, and their combination, as well as surge
impedance are relevant to system transfer limits. Transmission line series inductive reac-
tance is determined by conductor size, phase spacing, number of conductors, relative
phasing (double circuit lines), and line configuration. In transmission lines, the series
reactance is significantly larger than the series resistance, and is the dominant factor in a
first-order explanation of system behavior. For this reason, simple reconductoring of a
transmission line results in only minor changes in system power flows.
Critical factors related to power flow limits for overhead lines include:
Surge impedance loading
Voltage drop
Thermal limits
Environmental limits
Chapter 2: Overhead Transmission Lines Increased Power Flow Guidebook
2-2
2.1.1 Surge Impedance Loading
Surge impedance loading (SIL, defined in Equation 2.1-
1) provides a useful rule-of-thumb measure of transmis-
sion line loading limitation as a result of the effects of
series reactance.
2.1-1
For an overhead transmission line, typical surge imped-
ance is on the order of 300 ohms, while for a cable it
may be 50 ohms or less. At 345 kV, SIL of an overhead
line is on the order of 400 MW. Short lines may be able
to carry 800 MW or more, while long lines of exactly the
same construction may be limited to less than 400 MW
by system considerations.
2.1.2 Voltage Drop
Voltage control on the power system is of concern as
system loadings increase. The system voltage distribu-
tion is affected by the series inductance and shunt
capacitance of the transmission lines. It is not desirable
for voltage to vary more than 5%, or at most 10%, from
one end to the other. In some cases, a voltage drop limit
is placed on power flow corresponding to the maximum
allowable decrease in voltage magnitude. The longer the
line, generally the lower the power flow required to
reach a voltage drop limit. Voltage control is a system
problem, and is not generally solved by modifications to
any one transmission line.
Because voltage drop is primarily a function of line
reactance rather than resistance, simple reconductoring
does very little to decrease the voltage drop per unit
length. Reconductoring an existing 230-kV line by
replacing the original 636 kcmil (324 mm
2
) Hawk ACSR
with a 954 kcmil (487mm
2
) Rail ACSR only increases
the voltage drop limit by 5%. Adding a second conduc-
tor per phase, to form two conductor bundles, results in
a significant reduction in series reactance, and yields an
increase in the voltage drop power limit.
2.1.3 Thermal Limits
Thermal limits are discussed in considerable detail in
this chapter. In brief, the current carrying capacity
(thermal rating) of an overhead transmission circuit is
determined by the assumed worst case weather condi-
tions, assumed conductor parameters, and the maxi-
mum allowable conductor temperature. Some of the
specific thermal rating parameters are:
Conductor construction: outside diameter, conductor
strand diameter, core strand diameter, number of
conductor strands, and number of core strands.
Conductor AC resistance, which itself is a function of
the conductor temperature.
Conductor surface condition: solar absorptivity and
emissivity.
Line location: latitude, longitude, conductor inclina-
tion, conductor azimuth, and elevation above sea
level.
Weather: incident solar flux, air temperature, wind
speed, and wind direction.
2.1.4 Environmental Limits
The electric field produced by overhead power transmis-
sion lines is influenced by the following factors:
Line voltage
Height of conductors above ground
Configuration of conductors (line geometry, con-
ductor spacing, relative phasing of multi-circuit lines,
use of bundled conductors)
Lateral distance from the center line of the transmis-
sion line
Height above ground at the point of field measure-
ment
Proximity of conducting objects (trees, fences, build-
ings) and local terrain
The electric field near ground level produced by an over-
head transmission line induces voltages and currents in
nearby conducting objects (St. Clair 1953, Federal
Power Commission 1964, Dunlop et al. 1979, Koessler
and Feltes 1993, Boteler 1994, Lesher et al. 1994, EPRI
2005). These objects are typically the size of people, ani-
mals, motor vehicles, sheds, and similar-sized bodies.
Electric field coupling is capacitive coupling, and can be
represented by a current source in parallel with a high
source impedance (Norton equivalent circuit).
Electric field levels are limited by law in some jurisdic-
tions. Some regulations are specified at the edge of the
right-of-way (ROW) for public exposure. Other regula-
tions are maximum levels on the ROW based on induc-
tion to an assumed size object. These regulations may
restrict voltage increases on presently existing transmis-
sion lines without taking electric field reduction mea-
sures.
Magnetic field is affected by the same variables as elec-
tric field, except line current replaces line voltage, and
nearby objects generally have minimal impact on the
magnetic field. Magnetic field coupling is generally of
significance for objects that parallel the transmission
line for a long distance. Such objects include pipelines,
S
Z
V
SIL
2
=
2-3
Increased Power Flow Guidebook Chapter 2: Overhead Transmission Lines
telephone and railway signal circuits, and metal fences.
Because magnetic field is a function of line current, and
current increases during fault conditions, it may be nec-
essary to evaluate magnetic field effects under both nor-
mal operation and faults. Magnetic field coupling is
inductive coupling, and generally produces low voltages
with low source impedances (St. Clair 1953, Federal
Power Commission 1964, Dunlop et al. 1979, Koessler
and Feltes 1993, Boteler 1994, Lesher et al. 1994, EPRI
2005).
Increasing current on a transmission line increases the
magnetic field, and thus increases magnetically induced
voltages and currents. This may be significant in cases
such as when a transmission line parallels a railroad
ROW. This is the inductive coordination problem that
has been around since the dawn of the power industry
with respect to telephone and railroad signal facilities.
Increasing current flow on existing lines may require
coordination with parallel infrastructure. In some juris-
dictions maximum magnetic field levels are specified by
regulation. If an existing transmission line is operating
near the magnetic field limit set by regulation, the ability
to increase line current may be limited, unless measures
are taken to reduce the magnetic field levels.
Electric fields can be shielded by conducting objects.
Vegetation is sufficiently conductive to reduce electric
field levels. Grounded wires can be strung under the
phase conductors at road crossings to reduce electric
field levels. Grounding measures can be taken for fixed
objects to eliminate induced voltages. On the other
hand, magnetic field shielding is significantly more diffi-
cult than electric field shielding. Shielding a transmis-
sion line by magnetic materials is impractical. Flux
canceling loops have been developed, but incur power
loss and complexity in actively driven loops. Shielding is
less practical as a mitigation measure for magnetic fields
than it is for electric fields (St. Clair 1953, Federal Power
Commission 1964, Dunlop et al. 1979, Koessler and
Feltes 1993, Boteler 1994, Lesher et al. 1994, EPRI
1994, EPRI 2005).
Chapter 2 includes seven sections:
Section 2.2, Uprating Constraints, discusses con-
straints on electrical and mechanical safety, with
information on sag-tension calculations, limiting
high-temperature sag, constraints related to wind-
induced conductor motion, electrical clearance, loss
of conductor strength, constraints on structural
loads, and environmental effects.
Section 2.3, Line Thermal Ratings, explores the calcu-
lation of line thermal ratings, and describes common
heat balance methods.
Section 2.4, Effects of High-Temperature Operations,
describes annealing, calculation of sag and tension,
thermal and creep elongation, and connectors and
conductor hardware at high temperature.
Section 2.5, Uprating without Reconductoring, dis-
cusses deterministic and probabilistic methods of
uprating without reconductoring.
Section 2.6, Reconductoring without Structural Modi-
fications, reviews the various reconductoring choices
using new commercially available conductors.
Section 2.7, Dynamic Monitoring and Line Rating,
introduces the principles of dynamic rating methods.
Section 2.8, Case Studies, includes a number of uprat-
ing test cases and an economic comparison of line
uprating alternatives.
2.2 UPRATING CONSTRAINTS
2.2.1 Introduction
Increasing the thermal rating of an existing line requires
dealing with constraints on electrical and mechanical
safety. The uprated line must remain safe under all elec-
trical power flows up to its maximum without compro-
mising the mechanical safety under severe ice and wind
loads.
This section discusses issues related to constraints on
uprating, including determining what constitutes a con-
straint in various areas of design, operation, and the
environment.
2.2.2 Sag-tension Calculations
Normally, sag-tension calculations are performed
using numerical programs in order to determine the sag
and the tension of a conductor catenary as a function of
ice and wind loads, conductor temperature, and time.
Calculation examples from a program like SAG10 are
shown below to illustrate how tension limits are applied
and how maximum conductor tension and maximum
final high temperature sag are taken for the purposes of
strain structure design and tower placement. Details of
sag-tension calculation methods are not included, but
examples and key references are cited.
In the design, uprating, or simple maintenance of power
transmission lines, the concern of primary importance is
public safety. It is more important that a line be safe
than it carry power. Other than designing the support-
ing structures such that they remain standing under
even the most severe weather conditions, the safety of a
line is essentially determined by the position of its ener-
gized conductors relative to nearby people, buildings,
and vehicles. Maintaining minimum distances to nearby
2-4
objects and people is primarily a matter of limiting the
sag of the energized conductors under high mechanical
loads and high temperature conditions.
In addition to making lines safe, other important con-
straints are the level of electric and magnetic fields pro-
duced (e.g., electric fields increase as the conductor gets
closer to the ground), the maximum structure loads dur-
ing occasional high wind and ice loads, and the maxi-
mum temperature at which the energized conductors are
allowed to operate. Given standard worst-case
weather conditions, the thermal rating of an existing line
is determined by the maximum allowable conductor
temperature. Thus, uprating such lines without recon-
ductoring normally requires finding ways to maintain
electrical clearances while operating at a higher conduc-
tor temperature.
Figure 2.2-1 is a basic sag-clearance diagram, which
illustrates how minimum ground clearance must be
maintained under both heavy loading and high temper-
ature events over the life of both new and re-rated trans-
mission lines. The figure shows ground clearance and
line sags under normal conditions, high ice/wind load,
and high temperature conditions for a ruling (or equiv-
alent) span. Note that the sum of the minimum ground
clearance, the buffer, and the sag at maximum tempera-
ture is the minimum attachment height, which deter-
mines structure height and spacing. In a detailed line
design that has many different spans, this sort of sag-
clearance calculation must be developed for all spans
(Ehrenburg 1935, Winkleman 1959).
Definitions of the labels in Figure 2.2-1 are as follows:
Init is the initial installed unloaded (with no ice or
wind) sag of the conductor. It is typically at a con-
ductor temperature of 10C to 25C (50F to 80F).
This is also typically referred to as the line ruling
span stringing sag.
FinalSTC is the final sag of the conductor at 15
o
C
(60
o
F) after an ice/wind-loading event has occurred
for a short timetypically an hour. STC stands for
short-time creep.
FinalLTC is the final sag of the conductor at 15
o
C
(60
o
F) after an extended periodtypically 10 years
where the conductor simply persists at a conductor
temperature on the order of 15C (59F) without ice
or wind. LTC stands for long-time creep, which
occurs even if heavy ice and wind loads never occur.
Max Load is the sag of the conductor during the
specified maximum ice and wind loading at a reduced
temperaturetypically 18C to 0C (0F to 32F).
Note that the sag prior to this event is normally
assumed to be the Init sag and the sag after this event
is the FinalSTC sag.
TCmax is the sag of the conductor when its tem-
perature is the maximum for which the line is
designedtypically 50C to 150C. The final sag at
15
o
C (60
o
F), prior to this high temperature event, is
assumed to be the larger of the FinalSTC and the
FinalLTC sags.
Figure 2.2-1 shows typical behavior of transmission
conductors where the sag under maximum ice and wind
load conditions is less than that at the maximum tem-
perature. For small or weak conductors experiencing
heavy ice loads, this may not be true.
Note that the diagram illustrates the snapshot nature
of traditional sag-tension calculations. The actual con-
ductor sag position at any time in the life of the line
depends on the actual mechanical and electrical load
history of the line. If the high load event is more severe
or persists for a longer time than assumed in determin-
ing the Max Load condition, then the corresponding sag
at Max Load and the sag increase will be greater. The
use of buffers is required because of such uncertainties.
For transmission conductors made up primarily of alu-
minum strands under tension, sag never stops increasing
Figure 2.2-1 Sag diagram showing sags for various times
and loading conditions.
2-5
with both time and high loading events throughout the
life of the line (Aluminum Association 1974, Harvey and
Larson 1972, Harvey 1979). That is, the sag at a given
conductor temperature (e.g., 15.5C, or 60
o
F) increases
steadily over the years after construction. However, with
moderate unloaded and loaded conductor tensions (typ-
ically 15% and 50% of rated strength), the rate of change
in sag with each such event decreases over the life of the
line. Thus, if a heavy ice load event occurs 10 years after
installation, the permanent increase in sag is much
smaller than if it occurred in the first 6 months after
construction. Similarly, under everyday unloaded condi-
tions, the rate of change in sag will decrease with time,
over the life of the line.
Tension-Elongation Diagram (Normal)
The tension-elongation diagram shown in Figure
2.2-2 shows how the conductor tension changes corre-
sponding to the changes in sag position with load, time,
and temperature shown in the preceding sag diagram.
The initial unloaded (Init) sag corresponds to the ini-
tial unloaded (Init) tension. In the design of a new over-
head line, increasing this initial tension decreases
maximum temperature (Tcmax) sag and can allow the
use of fewer and shorter structures. However, increased
everyday tension levels also increase the maximum
(Max) tension loads (and thus the cost) on angle and
dead-end structures, and decrease the mechanical self-
damping of the conductor, which can lead to Aeolian
vibration-induced fatigue damage unless dampers are
applied.
With an older existing line that has reached its final sag,
increasing the conductor tension reinitiates creep
(though at a reduced rate). It also increases angle and
dead-end structure loads (though perhaps not higher
than they were upon initial installation) and is likely to
increase Aeolian vibration activity.
When reconductoring an existing line, an increase in the
maximum tension load may lead to the need for rein-
forcement or replacement of angle and dead-end struc-
tures and may be a critical factor in determining whether
reconductoring is an economic uprating solution.
The modulus (actually the spring constant) of the con-
ductor determines the increase in tension between
unloaded and loaded states. Figure 2.2-2 shows typical
behavior for a transmission conductor where the differ-
ence in tension between unloaded and loaded states may
result in a tension increase by a factor of two or more.
Specification of a realistic, nonlinear conductor modu-
lus (stress-strain behavior) under high tension loads is
important to the correct calculation of maximum ten-
sion. Use of a linear modulus will result in an overesti-
mate of the maximum tension.
As the temperature of the conductor increases, its length
and the resulting sag increase while the line tension
decreases. Errors in modeling the conductor modulus at
high temperatures have little or no effect on the calcu-
lated sag, but the thermal elongation behavior of con-
ductors at high temperatures is very important. As is
noted in Section 2.4, the thermal elongation of ACSR
can be particularly complex.
2.2.3 Limiting High Temperature Sag
The thermal elongation of stranded conductors is essen-
tially the same as that of its component strands. There-
fore, for an all aluminum or copper conductor, once the
sag at final everyday conditions is established, the sag
at high temperatures can be calculated and limited with
relatively small uncertainty.
High Temperature Sag with All Aluminum Conductors
For example, consider a line section of an all-aluminum,
37 strand (Arbutus) conductor having a ruling span of
600 ft (183 m) installed to meet the following con-
straints: maximum tension of 50%, 33% initial unloaded
at 15F and 25% final unloaded at 15F (-9.4 C). An
equally typical SAG10 program line design sag-tension
run is shown in Table 2.2-1.
Figure 2.2-2 Tension diagram showing conductor tension
for various times and loading conditions.
2-6
High Temperature Sag with ACSR
Because steel elongates thermally at half the rate of alu-
minum, the thermal elongation rate of ACSR conductor
is less than that of all aluminum conductor. Therefore,
older lines (which often have relatively small conductors
with high steel content) sag less than all aluminum con-
ductors for the same change in temperature.
The degree to which an ACSR conductors thermal
expansion is less than that of an all aluminum conduc-
tor (AAC) is dependent on the ratio of the steel to alu-
minum area. This ratio, expressed as a percentage, is
usually referred to as the ACSR Type number. Table
2.2-2 lists the composite thermal elongation of ACSR
conductors with different type numbers. Typical values
for the coefficient of thermal expansion () of an ACSR
are shown in Table 2.2-2.
Although we have listed composite thermal elongation
coefficients for ACSR, in reality the aluminum strands
elongate at twice the rate of the steel strands. The
reduced thermal elongation coefficient of the composite
is actually the result of both this difference in expansion
with temperature and the change in component tensions
that it produces.
Ignoring Aluminum Compression in ACSR at High
Temperature
Over the past 40 years, the Varney graphical method
(Aluminum Company of America 1961) has been the
basis of most sag-tension programs. The Alcoa SAG10
program is widely used. The sag-tension Table 2.2-3,
taken from the SAG10 program, shows the sag and ten-
sion (total, aluminum, and steel component tensions)
for initial and final conditions for 30/19, 795 kcmil (405
mm
2
) ACSR (Mallard) initially sagged so as not to
exceed a final unloaded tension of 25% of Mallards
Rated Breaking Strength at 60
o
F (15.5
o
C). NESC
Medium Loading conditions and conductor tempera-
tures up to 302
o
F (150
o
C) are included.
Notice that the knee point temperature, where the alu-
minum tension goes to zero, under final conditions,
occurs at only 90
o
F (32
o
C).
Figure 2.2-3 shows final sag versus conductor tempera-
ture for ACSR (Mallard) in four different ruling span
lengths. Note the change in slope of the curves below
50
o
C where the knee point is predicted to occur.
Many older lines that are typical candidates for uprating
were designed with high steel ACSR such as 30/19, 30/7,
and 26/7. The low thermal elongation beyond the knee
point temperature, illustrated in the preceding calcula-
tions, makes these older lines attractive candidates for
operation at higher temperatures. In such design situa-
tions, the difference in predicted sag at high temperature
can be very important.
Table 2.2-1 Sag-Tension Calculations for 37 AAC (Arbutus)
ALUMINUM COMPANY OF AMERICAN SAG AND TENSION DATA
Conductor Arbutus 795.0 kcmil 37 Strands AAC Area = 0.6234 sq in. Dia + 1.026 in.
Wt = 0.746 lb/F RTS= 13900 lb Span + 600.0 ft Creep is a Factor NESC Medium Load Zone
Design Points Final Initial
Temp
(F)
Ice
(in.)
Wind
(psf)
K
(lb/F)
Weight
(lb/F)
Sag
(ft)
Tension
(lb)
Sag
(ft)
Tension
(lb)
15. .25 4.00 .20 1.451 12.02 5446. 10.65 6140
32. .25 .00 .00 1.143 12.00 4294. 10.06 5118
0. .00 .00 .00 .746 8.77 3833. 6.63 5067.
15. .00 .00 .00 .746 9.67 3475.
a
a. Design condition.
7.27 4621.
30. .00 .00 .00 .746 10.58 3179. 7.98 4212.
60. .00 .00 .00 .746 12.34 2727. 9.54 3524.
90. .00 .00 .00 .746 13.99 2406. 11.19 3006.
120. .00 .00 .00 .746 15.54 2167. 12.82 2624.
167. .00 .00 .00 .746 17.78 1897. 15.24 2210.
212. .00 .00 .00 .746 19.73 1711. 17.37 1941.
Table 2.2-2 Coefficients of Thermal Expansion for Bare
Stranded Conductors
Conductor Type Number
(per degree C)
AAC 0
23.0 x 10
-6
36/1 ACSR 3
22.0 x 10
-6
18/1 ACSR 5
21.1 x 10
-6
45/7 ACSR 7
20.7 x 10
-6
54/7 ACSR 13
19.5 x 10
-6
26/7 ACSR 16
18.9 x 10
-6
30/7 or 30/19 ACSR 23
17.5 x 10
-6
2-7
Table 2.2-3 Sag-Tension Calculations for 30/19, 795 kcmil ACSR (Mallard)
Conductor Mallard 795.0 kcmil 30/19 ACSR Area = .7669 sq. in. Dia + 1.140 in.
Wt = 1.235 lb/F RTS = 38400 lb Span + 600.0 ft Creep is a Factor NESC Medium Load Zone
Temp
(F)
Ice
(in.)
Wind
(psf)
K
(lb/F)
Weight
(lb/F)
Sag
(ft)
Tension
(lb)
Sag
(ft)
Tension
(lb)
15. .25 4.00 .20 1.955 7.80
11283.
3423.A
7859.S
6.83
12880.
4986.A
7894.S
32. .25 .00 .00 1.667 7.68
9773.
2377.A
7395.S
6.36
11804.
4462.A
7342.S
0. .00 .00 .00 1.235 5.30
10495.
3193.A
7302.S
4.45
12499.
4972.A
7527.S
15. .00 .00 .00 1.235 5.79
9600.
a
2508.A
7092.S
4.69
11864.
4623.A
7242.S
30. .00 .00 .00 1.235 6.34
8775.
1860.A
6914.S
4.95
11241.
4277.A
6963.S
60. .00 .00 .00 1.235 7.56
7357.
693.A
6664.S
5.54
10039.
3605.A
6435.S
90. .00 .00 .00 1.235 8.65
6432.
0.A
6432.S
6.23
8921.
2966.A
5955.S
120. .00 .00 .00 1.235 9.26
6010.
0.A
6010.S
7.03
7910.
2373.A
5537.S
167. .00 .00 .00 1.235 10.27
5422.S
0.A
5422.S
8.45
6580.
1553.A
5027.S
212. .00 .00 .00 1.235 11.27
4939.
0.A
4939.S
9.94
5600.
894.A
4706.S
257. .00 .00 .00 1.235 12.30
4528.
0.A
4528.S
11.45
4864.
343.A
4522.S
302. .00 .00 .00 1.235 13.34
4178.
0.A
4178.S
12.80
4352.
0.A
4352.S
Figure 2.2-3 Sag for a strong 30/19 ACSR conductor (calculated ignoring aluminum strand
compression) as a function of conductor temperature and ruling span length.
2-8
Considering Aluminum Compression in ACSR at High
Temperature
Starting with the studies of Barrett at Ontario Hydro
(Barrett et al. 1982), the assumption of zero compressive
stress in ACSR beyond the knee point temperature has
come into question. The question centers not on the
correct calculation of the knee point temperature but on
whether the aluminum strands can support compressive
stresses above it.
The Canadian Electrical Associations STESS software
program incorporated Barretts research. Inclusion of
the compressive effects of the aluminum strands of high
steel content conductors such as 26/7 ACSR (Drake)
can add as much as 3 ft (0.91 m) to the sag at 150
o
C in a
1200 ft (366 m) span. The effect is less with smaller rul-
ing spans and with lower conductor temperatures.
Recent studies by Rawlins (Rawlins 1998) seem to con-
firm the existence of compressive effects as well as resid-
ual stresses (due to manufacturing) in aluminum strands
at high temperatures. The effect on sag at high tempera-
tures appears to be much smaller than those predicted
by Barrett. The widely used SAG10 program has incor-
porated Rawlinss studies as an optional calculation. In
a 1200 ft (366 m) span, Rawlinss method would add
about 1 ft (30 cm) to the sag of a high steel conductor
such as Drake at 150
o
C.
Figure 2.2-4 shows a comparison of sag as a function of
conductor temperature calculated with the following
assumptions:
The SAG10 computer program with an assumption
of zero compressive stress in the aluminum strands.
The SAG10 computer program with the default
assumption of 2500 psi (17.2MPa) residual stress and
allowance for aluminum compression.
The STESS computer program with the default
assumption of 10 MPa (1450 psi) for maximum com-
pressive stress and no residual stress.
Figure 2.2-5 is a similar plot that shows the somewhat
larger sag differences that occur in a 1200 ft (366 m) rul-
ing span.
At this point, there is no clear way to determine which
of these methods is correct. Indeed, there is no way to
be certain that the stress assumptions for any of the cal-
culations is correct for all ACSR conductors installed in
old and new lines. However, there is a distinct possibility
that the original line design sag-tension calculations,
assuming no compressive stress in the multiple alumi-
num layers, yielded sags above the kneepoint that were
too small. The uncertainty centers on how much the sag
should be increased to be certain that electrical clear-
ances will be maintained at an increased maximum con-
ductor temperature.
2.2.4 Uprating Constraints Related to Wind-
Induced Conductor Motion
Transmission lines must be designed not only to provide
adequate vertical clearance for electrical and safety con-
siderations, but also to allow for adequate horizontal
clearance to tall objects and buildings at the edge of the
ROW under high wind conditions. This conductor dis-
placement is termed conductor blowout and is normally
at its maximum midway between conductor support
points, as shown in Figure 2.2-6. Note that the horizon-
tal displacement at midspan (X
H
) is determined in part
by the conductor sag (D).
The maximum displacement of the outermost conduc-
tors from the center of the ROW under high wind condi-
tions can be one of the most important variables in
Figure 2.2-4 Sag at high temperatures calculated with and
without aluminum compression.
Figure 2.2-5 Final sags for Mallard ACSR in a 1200-ft
span.
2-9
determining ROW width or for a given ROW width,
determining the maximum structure spacing.
In addition to blowout in strong cross-winds, wind can
cause certain oscillatory conductor motions. See Table
2.2-4. By far the most common oscillatory wind-
induced motion is aeolian vibration since it occurs
under low-speed everyday wind conditions. Unless it is
controlled, aeolian vibration can accumulate millions of
cycles, which cause fatigue failure of copper, aluminum,
or steel strands. A less common, but more dramatic,
form of wind-induced conductor motion is ice gallop-
ing. It occurs for strong winds in combination with ice
on the conductors and can yield high amplitude, oscilla-
tory conductor motions that result in repeated flash-
overs between the phase conductors or between a phase
conductor and a shield wire.
The amplitude of wind-induced conductor aeolian
vibration is generally less than the conductor diameter.
It can be measured with special monitors, but the calcu-
lation methods are complex. Two simple methods of
control are widely used: (1) the tension of the line con-
ductors under everyday conditions is limited and;
(2) vibration dampers are clamped to the line conduc-
tors. In regions where aeolian vibration is a problem,
transmission line conductor tensions are typically lim-
ited to between 15% and 20% of RBS during the coldest
month of the year, and vibration dampers are routinely
installed in every span.
Ice galloping motions can be predicted in a crude way
through the use of ice galloping ellipses. By comparing
such ellipses to the spacing between the line conductors,
the likelihood of flashovers from galloping can be mini-
mized by providing sufficient phase spacing and by off-
setting any vertical phases. Also, since the major axis of
the galloping ellipse is proportional to the line conduc-
tor sag with ice and wind loading, the amplitude of ice
galloping motions can be reduced by minimizing the sag
of conductors in the typical span.
Wind-induced subconductor oscillation only occurs for
bundled phase conductors when wind speeds exceed a
certain critical velocity. If uncontrolled, it can result in
fatigue damage to spacers and suspension hardware.
Oscillations are controlled by keeping bundled conduc-
tors at a spacing-to-diameter ratio of about 20 or more
and by avoiding uniform spacer spacing.
With regard to increasing maximum allowable power
flow through existing lines, wind-induced conductor
motions are a primary constraint on increasing the line
operating voltage, on retensioning the existing conduc-
tors to allow operation at higher maximum conductor
temperatures, on reconductoring the line with high
temperature, low-sag conductor, and on bundling (add-
ing a second conductor per phase).
Figure 2.2-6 Illustration of midspan conductor blowout
due to wind.
Table 2.2-4 Cyclic, Wind-induced Conductor Motions
Aeolian Vibration Ice Galloping Subconductor Oscillation
Types of Overhead Lines Affected All All in regions with ice
Lines with bundled con-
ductors
Approx. Frequency Range, Hz 3 to 150 0.08 to 3 0.15 to 10
Approx. Range of Vibration Amplitudes (Peak-to-
peak, Expressed in conductor diameters)
0.01 to 1.0 5 to 300 0.5 to 80
Weather Conditions Favoring Conductor Motion
Wind Character:
Wind Velocity:
Conductor Surface:
Steady
1 to 7 m/sec (2 to 15 mph)
Bare or uniformly iced
Steady
7 to 18 m/sec (15 to 40 mph)
Asymmetrical, thin ice
deposits
Steady
4 to 18 m/sec (10 to 40 mph)
Bare, Dry
Damage
Characteristics:
Direct Causes of Damage:
3 months to 20 years for
damage to occur
Conductor fatigue due to
cyclic bending
1 to 48 hours per occurrence
Repeated Flashovers, High
dynamic loads on structures,
premature wear of hardware
1 month to 8+ years for dam-
age to occur
Conductor clashing, acceler-
ated wear of hardware
2-10
HTLS conductors such as ACSS are particularly advan-
tageous in reconductoring if they are prestressed. When
prestressed, ACSS has much higher self-damping than
standard ACSR. It can be installed with smaller initial
sag, which reduces ice galloping motions and may allow
operation of an existing line at higher voltage as well as
higher current levels.
In lines without vibration dampers, the addition of
dampers may allow the lines existing conductors to be
retensioned and operated to a higher maximum temper-
ature without the need for reconductoring.
When bundling new conductors with old, the use of a
vertical bundle can eliminate the problem of subconduc-
tor oscillation while keeping the bundle spacing to no
more than 9 to 12 inches.
Regardless of the uprating method, wind-induced
motions must be thoroughly considered as part of the
redesign.
2.2.5 Electrical Clearance
The National Electric Safety Code (National Electric
Safety Code 1997) specifies minimum spacings from
energized conductors to ground, to objects passing
under the line, to buildings nearby, and to other conduc-
tors (underbuild). These clearances must be main-
tained under The maximum conductor temperature for
which the line is designed to operate (NESC 232.A.2).
Failure to maintain such minimum distances is a public
safety issue of primary importance.
The National Electric Safety Code also specifies mini-
mum horizontal spacings from energized conductors to
other conductors and objects such as buildings at the
edge of ROW. When subjected to transverse wind, the
conductor catenaries blow out, and the horizontal
spacing of energized conductors to buildings, etc. is
reduced. This reduction in horizontal clearance can be
limited by using heavier conductors or shorter span
lengths; reducing energized conductor sag under blow-
out conditions; using insulators such as V strings, hori-
zontal V, and posts that do not move with wind; and
providing generous right-of-way widths. Reducing sag
under horizontal blowout conditions is limited by con-
cerns about vibration, but is complementary to uprating
methods such as retensioning.
Minimum electrical clearances must be maintained
under all line loading and environmental conditions.
Since the actual sag clearance of transmission lines is
seldom monitored, sufficient allowance for this clear-
ance must be included in the process of initial design or
in rerating of existing lines.
Applicable Code Clearances
In all cases, national codes may apply. In the United
States, the National Electric Safety Code (NESC) is
applicable. State codes may also apply. Minimum hori-
zontal and vertical distances from energized conductor
(electrical clearances) to ground, other conductors,
vehicles, and objects such as buildings are a function of
three things: the line-to-ground voltage, the use of
ground fault relaying, and type of object or vehicles.
The NESC Rules covers both vertical and horizontal
clearances. That is, the code sets minimum spacing for
energized conductors both above and next to people,
vehicles, and buildings. This chapter considers only ver-
tical clearances since our focus is on high temperature
operations (see Tables 2.2-5 and 2.2-6). Horizontal
clearances are typically specified for high winds where
the transmission line catenaries are horizontally dis-
placed by the wind. In such cases, the conductor tem-
perature is low due to high convection cooling.
Ground clearance minimums listed in the NESC code
are primarily due to the height of the object or person
that may pass beneath the span. For example, a person
with an overhead umbrella extended overhead at arms
length may physically reach 10 ft (3 m) above ground,
whereas a railroad car may be as much as 20 ft (6 m)
high. The NESC code calls for a minimum ground
clearance of 27 ft (8.2 m) for a low-voltage conductor
over a railroad and only 16.5 ft (5 m) over spaces or
ways accessible only to pedestrians. The difference in
minimum ground clearance is due primarily to the
height of the object under the line. In each case, the
clearance between the low-voltage conductor and the
top of the conflicting object is approximately the same.
Essentially, the minimum vertical ground clearance for
any supply conductor (0 to 750 V) is defined by the
NESC code as 16.5 ft (5 m) for lines going over places
such as roads, streets, driveways, parking lots, and farm-
land, or any other type of land which can be traversed
by vehicles. Conductors passing over waterways must
generally meet greater clearance requirements.
The Influence of Line Voltage on Clearance
For those lines having a line-to-ground voltage of 750 to
22 kV, the ground clearance for the 0 to 750 V supply
conductor is increased by 2 ft to 18.5 ft (0.6 m to 5.6 m).
For lines at higher voltages, the vertical clearance is
increased by 0.4 in. (1 cm) for every kV increase in line-
to-ground voltage above 22 kV. Note that the voltage
used in these calculations of added electrical clearance
are based on the maximum operating voltage, which is
typically 5% or 10% above nominal.
2-11
Reduced Clearance for EHV Lines with Limited Switching
Surge Levels
For lines exceeding 98 kV line to ground, the code
allows clearances to be calculated based on knowledge
of switching surge levels. If the switching surge level is
restrained to 2.2 p.u., the clearance at EHV voltages
may be decreased.
Power System Conditions when Clearances Apply
It is impossible to be certain that clearances will be
maintained under all foreseeable circumstances. For
example, in certain regions, tornadoes may occur, which
might cause conductors that are energized to fall to
earth. However, it is irresponsible to design lines or line
upgrades where clearance violations are likely to occur.
The minimum ground clearances specified by the NESC
code apply to energized conductors under the three con-
ditions specified in Rule 232A where the temperatures
specified are that of the conductor not the surrounding
air:
50C (122F) with no wind displacement.
At the maximum operating temperature for which the
line is designed to operate if greater than 50C
(122F) with no wind displacement.
0C (32F), no wind displacement, with radial thick-
ness of ice.
Even in these days of heavily utilized transmission
assets, it is unusual for lines to carry electrical loads that
cause the energized conductors to be more than 5C or
10C above air temperature. However, given the rela-
tively rare loss (outage) of a major generating station or
EHV transmission circuit, electrical loading on HV lines
can increase and cause much higher conductor tempera-
tures. Thus, all lines are designed to meet clearances at
the maximum operating temperature for which the line
is designed to operate (see above).
Heavy ice loads are also relatively rare events, but in any
modern HV or EHV line, the energized conductor sag at
0C (32F) with maximum ice is typically less than the
sag for high temperature, even when that maximum
operating temperature is only 50C (122F). Thus, the
assurance of adequate clearance involves the behavior of
transmission conductors at high temperatures, not
under heavy ice load.
Transmission line operators typically meet minimum
clearance requirements by limiting the current on the
energized conductors. The specification of any relation-
ship between the electrical current on the energized con-
ductors and the conductor temperature is left to the
discretion of the operator.
The NESC code describes the minimum clearances in
considerable detail as a function of voltage and poten-
tially conflicting activity. The code also prescribes the
conditions under which the clearance minimums must be
met. The code does not, however, specify: (1) how the
temperature of the conductor is to be calculated; (2) how
the physical position of the conductor above ground is to
be related to this maximum operating temperature; nor
(3) how adequate ground clearance can be confirmed
under rare occasions of high electrical loading. Conse-
quently, methods of ensuring adequate ground clearance
vary widely between transmission line operators.
Upgrading Buffers
On older transmission lines, the structure placement
along the ROW is fixed, and the final sag of the conduc-
tor is measurable. Thus any initial spacing buffer added
because of uncertainties in structure placement and ini-
tial sag can be reduced in uprating. There are, however,
certain irreducible uncertainties, and some clearance
buffer must be maintained.
The traditional method of determining clearances for
existing transmission lines involves standard survey
methods to determine the conductor attachment points
and the sag at span mid-point for one or more spans in
Table 2.2-5 Minimum Vertical Ground Clearances
According to NESC C2-1997, Rule 232C
L-L/L-G
Basic Clear-
ance @ 22 kV
Clearance Added for
Voltage Streets
kV ft m ft m
69/40 18.5 5.6 0.7 19.2 5.8
138/80 18.5 5.6 2.1 20.6 6.3
161/93 18.5 5.6 2.5 21.0 6.4
230/133 18.5 5.6 3.9 22.2 6.8
345/200 18.5 5.6 7.0 25.5 7.8
500/290 18.5 5.6 9.9 28.4 8.7
765/440 18.5 5.6 15.5 34.0 10.4
Table 2.2-6 Minimum Vertical Ground Clearances
According to NESC C2-1997, Rule 232D*
Nominal
Voltage
L-L/L-G
Reference
Height Listed in
Table 232-3
Alternate
Clearance Adder
Minimum
Clearance for
Streets
KV ft m ft m ft m
69/40 - 19.2 5.9
138/80 - 20.6 6.3
161/93 21.0 6.4
230/133 14 4.3 7.1 2.2 21.0 6.4
345/200 14 4.3 7.1 2.2 21.0 6.4
500/290 14 4.3 12.7 3.9 26.7 8.1
765/440 14 4.3 21.8 6.6 32.4 9.9
* In accord with Rule 232D4, the clearance calculated based
on Rule 232D2-3 cannot be less than the clearance calcu-
lated for 98 kV under Rule 232C, which is 21.0 ft.
2-12
each line section. These measurements are typically
taken with the line out of service so that the conductor
is at air temperature plus some solar rise. Vertical posi-
tion errors of up to a foot (0.3 m) are easily made in
determining the catenarys position and the attachment
heights. Additional errors may be expected in determin-
ing the ground clearance since the ground profile is only
checked at a few points along the line.
In more recent years, several photographic and laser-
based methods have been developed. These methods
help determine all attachment points at all structures,
and provide a complete description of catenary profiles
for all three-phase conductors with much better accu-
racy than is possible with conventional survey methods.
Such measurements are seldom done with the line out of
service, so the conductor temperature at the time of the
survey measurements must be calculated or measured.
The result of such detailed survey activity is very
impressive and can easily convince the novice that buff-
ers can be eliminated or reduced when upgrading exist-
ing lines. This is not the case.
Knowing the exact ground clearance with perfect cer-
tainty at the conclusion of a laser survey does not mean
that one can be certain of adequate clearance under
maximum electrical loading. There still remain several
uncertainties.
2.2.6 Loss of Conductor Strength
Construction codes also require that maximum conduc-
tor tension not exceed a certain percentage of the ener-
gized conductors breaking strength. A significant
reduction in the breaking strength can weaken the ener-
gized conductor and lead to a tensile failure during sub-
sequent high ice and wind loading events. To avoid this,
the conductor must not operate at a high enough tem-
perature for a long enough period of time so as to
reduce its breaking strength more than 10%, and it must
not be installed at such a high everyday unloaded ten-
sion that its strands fatigue due to wind vibration.
The American Society for Testing and Materials
(ASTM) or the International Engineering Consortium
(IEC) standards specify the minimum tensile strength of
aluminum and copper wires, which is the stress at which
the wire breaks. At temperatures above 75C, the tensile
strength decreases with time. Temperatures below 300C
do not affect the tensile strength of galvanized, alumi-
num-clad, or copper-clad steel wires. Thus, extended
exposure of conductors made up largely of aluminum or
copper wires to temperatures above 75
o
C can eventually
lead to tensile failures during high ice and/or wind load-
ing events.
Figure 2.2-7 shows the reduction in tensile strength with
time and temperature for a sample of 0.081 in. (0.2 cm)
diameter hard drawn copper wire, as described in (Hick-
ernell et al. 1949). There are 8760 hours in a year, so the
diagram clearly shows that sustained operation at 65
C
yields no measurable reduction of tensile strength, sus-
tained operation at 100
o
C yields a 10% reduction in 600
hours (25 days), and that only 40 hours at 125
o
C
reduces the wire tensile strength by 10%.
Figure 2.2-8 shows similar tensile strength reduction
data for 1350-H19 EC hard drawn aluminum wire. It
Figure 2.2-7 Annealing of 0.081 in. diameter hard drawn
copper wire.
Figure 2.2-8 Annealing of 1350-H19 hard drawn
aluminum wire (Aluminum Association 1989).
2-13
is taken from (Aluminum Association 1989). In general,
tensile strength reduction of aluminum wires at temper-
atures of less than 90
o
C is considered negligible. At
100
o
C, the tensile strength of the wire is reduced by 10%
after 5000 hours, and at 125
o
C the tensile strength is
reduced 10% after 250 hours.
When compared to copper, aluminum appears to anneal
somewhat more slowly, though the difference is proba-
bly not important in transmission line applications. The
source of the copper wire data also noted a significant
amount of variation in the annealing rates for wire
obtained from different manufacturers.
In applying these equations, the cumulative strength
reduction for multiple exposures at the same conductor
temperature may be found by simply adding up all the
hours and calculating the residual strength. However,
for multiple exposures at different conductor tempera-
tures, the calculation process is more complex. To deter-
mine the cumulative strength reduction for a series of
high temperature exposures at different temperatures
and times, all exposures must be expressed in equivalent
time at the highest temperature before adding.
Thus, if an all aluminum conductor consisting of 37-
0.1466 in. diameter strands is raised to 125
o
C for 100
hours and then at a later time for 50 hours, then the
strength reduction can be calculated for 150 hours at
125
o
C. If the same conductor is raised to 125
o
C for 100
hours and then at a later time is raised to 150
o
C for 50
hours, then the following calculation must be per-
formed:
For 100 hours @125
o
C,
2.2-1
At 150
o
C, RS = 91.0% after 7.2 hours, so the cumulative
loss of strength over the two high temperature exposures
is equal to the remaining strength after 50 + 7.2 hours.
It is 82.1%
2.2.7 Constraints on Structural Loads
Existing structures and foundations were designed for
certain maximum transverse wind loads and, for strain
structures, for maximum tension loads. Unless the exist-
ing structures are to be replaced, retensioning the exist-
ing conductor, or reconductoring the line with a new
conductor, must be done without greatly exceeding the
original design limits on structure loading. If many of
the line structures must be replaced, the design solution
is not the sort of minimal capital investment solution
that this guide emphasizes.
For tangent structures, the governing transverse loads
are primarily a function of the conductor diameter.
Thus the replacement conductor diameter must be
within about 10% of the existing conductor to avoid tan-
gent structure modification. For angle and dead-end
structures, the governing loads are primarily related to
maximum conductor tension. The replacement conduc-
tors maximum tension should not exceed the original
conductors initial maximum conductor tension unless
these structures are to be reinforced.
Before undertaking any uprating project, a review of the
existing structures and operating records of the line is
required. If structural failures at angle or dead-end
structures have occurred, any attempt at increasing
everyday installed tension is unlikely to succeed. Simi-
larly, if occasional high temperature operations have
yielded splice failures, increasing operating temperature
levels without replacing or inspecting the line is unwise.
On the other hand, if a review of structure and founda-
tion capacity indicates that the line was conservatively
designed, and that it has operated for many years with-
out any structural or foundation failures, it may be pos-
sible to replace the existing conductor with a new larger
conductor without structure modifications.
When a conductor span is ice covered and/or exposed to
high winds, the effective conductor weight per unit
length increases. During occasions of heavy ice and/or
wind load, the conductor tension increases dramatically,
along with the loads on angle and dead-end structures.
Both the conductor and its supports can fail unless
these high-tension conditions are considered in the line
design. The National Electric Safety Code (NESC) sug-
gests certain combinations of ice and wind correspond-
ing to heavy, medium, and light loading regions of the
United States.
The NESC Code (National Electric Safety Code 1997)
also suggests that increased conductor loads due to high
wind loads but no ice should be considered as noted in
the last column of Table 2.2-7.
Certain utilities in very heavy ice areas use glaze ice
thickness as much as 2 to 3 in. (5 to 7.6 cm) in order to
calculate iced conductor weight. This is especially true if
they have experienced extensive line failures due to ice
loads in excess of those recommended by the NESC.
Similarly, utilities in regions where hurricane winds
occur may use wind loads as high as 0.236 psi (1630 Pa).
Ice Loading
The formation of ice on overhead conductors may take
several physical forms such as glaze ice, rime ice, or wet
( )
% 0 . 91 100 100
1466 . 0
1 . 0
095 . 0 125 . 0
= =

RS
2-14
snow. The impact of lower density ice formation is usu-
ally considered in the design of line sections at high
altitudes.
The formation of ice on overhead conductors has the
following influence on line design:
Ice loads determine the maximum vertical conductor
loads that structures and foundations must with-
stand.
In combination with simultaneous wind loads, iced
conductor may also yield the highest transverse
design loads on structures.
In regions of heavy ice loads, the maximum sags and
the permanent increase in sag with time (difference
between initial and final sags) may be due to ice
loading.
In addition to the NESC loading region, ice loads for
use in designing lines may also derive from past experi-
ence, state regulations, and analysis of historical
weather data. Mean recurrence intervals for heavy ice
loadings are a function of local conditions along various
routings. Line design software can be used to investigate
the impact of a variety of assumptions concerning ice
loading. The calculation of glaze ice loads on conduc-
tors is normally done with an assumed ice density of
57 lb/ft
3
(913 kg/m
3
).
The ratio of iced weight to bare weight depends strongly
upon the conductor diameter. As shown in Table 2.2-8,
for three different conductors covered with 0.5 in. (1.27
cm) radial glaze ice, this ratio ranges from 4.8 for #1/0
AWG to 1.6 for 1590 kcmil (811 mm
2
) conductors.
Therefore, small diameter conductors may need to have
a higher elastic modulus and higher tensile strength
than large conductors in heavy ice and wind loading
areas to limit the sag.
Wind Loading
Wind loading on overhead conductors influences line
design in a number of ways:
The maximum span between structures may be deter-
mined by the need for horizontal clearance to the
edge of the ROW during moderate winds.
The maximum transverse loads for tangent and small
angle suspension structures are often determined by
infrequent high wind-speed loading.
Wind loading determines the permanent increase in
conductor sag in areas of light ice loads.
Wind pressure load on conductors, Pw, is commonly
specified in lb/ft
2
. Equation 2.2-2 gives the relationship
between Pw and wind velocity:
2.2-2
Where V
w
= the wind speed in miles per hour.
The wind load per unit length of conductor, Ww, is
equal to the wind pressure load, Pw, multiplied by the
conductor diameter (including radial ice of thickness t,
if any):
2.2-3
Combined Ice and Wind Loading
If the conductor weight is to include ice and wind load-
ing, the resultant magnitude of the loads must be deter-
Table 2.2-7 Definition of NESC Loading Areas
Loading Districts
Heavy Medium Light
Extreme
Wind
Loading
Radial thickness
of ice (in.)
12.5 6.5 0 0
Radial thickness of ice
(mm)
318 165 0 0
Horizontal wind
pressure (lb/ft
2
)
4 4 9 16 to 22
Horizontal wind
pressure (Pa)
190 190 430 16 to 22
Temperature (F) 0 +15 +30 +60
Temperature (C) -18 -10 -1 +15
Constant to be added
to the resultant (all
conductors) (lb/ft)
0.30 0.20 0.05 0.0
Constant to be added
to the resultant (all
conductors) (N/m)
4.40 2.50 0.70 0.0
Table 2.2-8 Ratio of Iced to Bare Conductor Weight
ACSR
Conductor D w
bare
w
ice
w
bare
+ w
ice
---------
w
bare
(in.) (lb/ft) (lb/ft)
#1/0AWG-6/1
Raven
0.398 0.1452 0.559 4.8
47-kcmil-26/7
Hawk
0.858 0.6570 0.845 2.3
1590-kcmil-54/19
Falcon
1.545 2.044 1.272 1.6
[ ]
[ ]
2
2 2
) / ( 0473 . 0 ) (
) ( 00256 . 0 ) / (
h km V Pascals P
mph V ft lb P
w w
w w
=
=
[ ]
[ ]
1000
) ( 2 ) (
) ( ) / (
12
) ( 2 ) (
) ( ) / (
mm t mm D
Pascals P m N W
in t in D
psf P ft lb W
c
w w
c
w w
+
=
+
=
2-15
mined. Equation 2.2-4 gives the weight of a conductor
under both ice and wind loading:
2.2-4
Where
w
b
= bare conductor weight per unit length.
w
i
= weight of ice per unit length.
w
w
= wind load per unit length.
w
w+i
= resultant of ice and wind loads.
2.2.8 Environmental Effects
The public considers overhead transmission lines as
very visible and, though most power engineers have dif-
ficulty in understanding why, unattractive. Thus, one of
the primary environmental effects of any transmission
line is their visual impact on their surroundings. A great
deal of effort has been expended on making lines more
visually acceptable with decidedly mixed results. Fortu-
nately, when uprating existing lines, most of the normal
opposition to new lines is avoided unless the appearance
of the uprated line is significantly changed.
Because lines are highly visible and perceived as unat-
tractive, they can have a negative impact on property
values. This is typically much less of an issue with the
modification of existing lines than with new lines.
Figure 2.2.9 shows a comparison of the relative impor-
tance of some of the major environmental issues involv-
ing overhead lines as determined by survey. It is
interesting that the top three factors are primarily a
matter of human perception or belief, while the three
least important issues are matters of physics.
Various attempts to reduce the visual impact of lines
and the corresponding impact on property values have
been made. There have been design competitions to find
more visually acceptable structures and research into
methods of compacting HV lines so they look more like
distribution lines.
Probably the most effective way to reduce public opposi-
tion to transmission lines concerns putting them away
from where people live and work. Clearly, this is not
always possible, but as shown in Table 2.2-9, it is quite
effective.
In the specific case of uprating, a variation on the classic
physicians ethic. First, do no harm makes sense. Spe-
cifically, uprating techniques that do not raise structure
peaks or make conductors more visible from a distance
are preferred.
This guide emphasizes line uprating methods where the
voltage of the line remains the same but current flow is
increased. Most of the techniques covered herein will
leave the original ground level electric field, electric
induction, corona discharge levels and audible noise lev-
els unchanged. However, the ground level magnetic field
and magnetic induction levels will increase with the
higher line currents. Both environmental effects are lin-
ear with current so that maximum original levels are
easily estimated by scaling with the increase in rating.
2.3 LINE THERMAL RATINGS
2.3.1 Introduction
The temperature and/or sag of overhead power trans-
mission lines can be measured, but seldom are. Rather,
in order to avoid excessive sag or loss of strength, a
maximum allowable conductor temperature is typi-
cally specified, and the conductor temperature is kept
below this maximum by placing limits on the level and
duration of power transferred over the line (MVA or
amperes). If such limits are based on worst-case weather
conditions, they are called static ratings, and if based on
actual weather conditions, they are called dynamic rat-
ings. The calculation of thermal ratings for overhead
lines is an essential part of the uprating process.
2 2
) ( ) (
w i b i w
W w w w + + =
+
Figure 2.2-9 Median survey results as to why people
oppose transmission lines.
Table 2.2-9 The Impact of Distance on Public Opposition
to Power Transmission Lines
Distance from Line
Feeling Less than 1 mile More than 1 mile
Like it 2.3% 3.2%
Dont care 32.6% 71.3%
Dislike it 65.1% 25.3%
2-16
The electrical power conductors of overhead transmis-
sion lines carry relatively large electrical currents, are
self-supporting, and energized at high voltage. They are
stranded from wires of aluminum or copper, which may
be reinforced with a steel core. As the current flowing
through a conductor i ncreases, i ts temperature
increases, and it elongates. This elongation increases the
sag of the conductor between support points, decreasing
the clearance to people, ground, other conductors,
buildings, and vehicles under the line. Beyond a certain
maximum allowable sag, the line may flashover,
resulting in either a power supply outage or injury to the
public. If the conductor temperature remains high for
an extended period of time, the strength of the conduc-
tor and tensioned connectors may decrease, resulting in
mechanical failure during the next occurrence of ice or
high wind loading.
In the design of a new transmission line, the thermal rat-
ing required for reliable system operation can be
attained either by selecting a large conductor at a mod-
erate maximum operating temperature or by using a
smaller conductor at a higher maximum temperature.
The higher sag resulting from higher operating tempera-
tures can easily be accommodated by using taller or
more closely spaced structures.
In uprating an existing line, it is unusual if the existing
conductor can be replaced by a significantly larger con-
ductor since this would increase both transverse wind
loads and maximum tension loads, and require costly
and time-consuming rebuilding of existing structures. It
is also unusual if the existing conductor can simply be
operated at a significantly higher temperature without
raising the existing attachment points or retensioning
the line, since this would lead to unacceptable violations
of minimum electrical clearances under maximum elec-
trical loading.
2.3.2 Maximum Conductor Temperature
Modern transmission conductors are typically stranded
from aluminum wires with a steel core added where
increased strength is required. The temperature limit on
all-aluminum or ACSR conductors is based on the max-
imum sag or maximum loss of strength in the alumi-
num. Temperature limits for normal ACSR conductors
in use today range from 50C to 150C (122F to
302F). The temperature limit is normally selected at
the time the line is designed. The higher this tempera-
ture, the higher the thermal capacity of the line, the
maximum conductor sag, and the higher (or closer) the
structures required to maintain ground clearance.
If aluminum or copper conductor temperatures remain
high (above 95 C, or 203 F) for an extended period of
time, the strength of the conductors and tensioned con-
nectors may decrease, which eventually results in
mechanical failure during ice or high wind occurrences.
Generally, rating durations are kept short if maximum
conductor temperatures are high (e.g., 4 hour maximum
at 115 C [239 F] and 15 minutes at 125 C [257 F]).
These high temperature effects on conductor, hardware,
and fittings are discussed in detail in Section 2.4.
2.3.3 Weather Conditions for Rating Calculation
Traditionally, power utilities use fixed worst-case
weather conditions in order to calculate (static) line rat-
ings. The impact of changes in these weather parameters
upon thermal line ratings depends on the specific rating
situation. Consider an overhead line with 795 kcmil (402
mm
2
) of aluminum, 26/7, Drake ACSR conductor,
whose static rating is based upon a maximum allowable
conductor temperature of 100
o
C with an air tempera-
ture of 40
o
C, full summer sun, and a wind blowing per-
pendicular to the conductor axis at 2 ft/sec (0.61 m/sec).
The static rating under these conditions is 1000 A.
Clearly, if the current in this conductor is 1000 A with
the assumed weather conditions, the conductor temper-
ature is 100
o
C. Table 2.3-1 shows how the conductor
temperature is affected by small changes in weather con-
ditions. For example, the conductor temperature drops
to 92
o
C if there is no solar heating. The table also shows
how the thermal rating (i.e., the current which yields a
temperature of 100
o
C) changes with small changes in
weather.
Note that with the conductor at a reasonably high tem-
perature and near worst-case heat transfer conditions,
the overhead line rating and conductor temperature are
very sensitive to wind direction, modestly sensitive to
Table 2.3-1 Variation in Conductor Temperature and
Rating with Weather Conditions
(for 795 kcmil [404 mm
2
], 26/7, Drake ACSR conductor
with a maximum allowable conductor temperature of
100C, an air temperature of 40
o
C, full summer sun, and a
wind blowing perpendicular to the conductor axis at 2
ft/sec)
Range in
Weather
Conditions
Line Rating @
100C
Conductor Temperature
at 1000 A
(amperes) (C) (F)
None 1000 100 212
Air temp
= 39C
1010 99 210
No sun 1070 92 198
3 ft/sec
(0.91m/sec)
1090 90 194
Parallel wind 750 133 271
2-17
changes in wind speed and solar heating, and less
affected by small changes in air temperature. Other
minor factors are gradual changes in emissivity and
absorptivity of the conductor with age, and seasonal
shifts in solar heating.
2.3.4 How Line Design Temperature Affects Line
Ratings
Line design temperature is the maximum allowable con-
ductor temperature for a particular line. As noted previ-
ously, for normal conventional ACSR, it varies from
50
o
C to 150
o
C. The impact of changes in the line design
temperature upon thermal line ratings depends on the
specific rating situation, but certain observations are
possible.
Until the early 1970s, the National Electric Safety Code
(National Electric Safety Code 1997) suggested that
minimum electrical clearances were to be met at conduc-
tor temperatures up to 120F (49
o
C). Line thermal
capacity was typically calculated by conductor manu-
facturers for a conductor temperature of 75
o
C, a tem-
perature sure to avoid possible annealing problems with
aluminum and copper.
In the 1970s, the NESC changed this position and stated
that the electrical clearances listed were to be met at
the maximum conductor temperature for which the
line was designed to operate, if greater than 50
o
C, with
no wind displacement (excerpted from Rule 232.A.2).
Thus the maximum allowable conductor temperature
(MACT) used in line rating calculations may vary from
50
o
C to 200
o
C according to available ground clearances,
and consistency, with concerns about loss of tensile
strength at temperatures above 90
o
C.
Consider an existing overhead line with 795 kcmil
(402 mm
2
) of aluminum, 26/7, Drake ACSR conduc-
tor, whose static rating is based upon an air temperature
of 40
o
C, full summer sun, and a wind blowing perpen-
dicular to the conductor axis at 2 ft/sec (0.61 m/sec).
The rating of this existing line depends on the line
design temperature as is shown in Table 2.3-2, where the
line design temperature with this ACSR conductor
ranges from 50C to 150
o
C.
For each line design temperature, the line rating is
shown. Also shown is the increase in rating that corre-
sponds to an increase of 10
o
C in the line design temper-
ature. For this ACSR conductor, the increase in sag for a
10
o
C increase in conductor temperature decreases with
increasing line design temperature.
It is clear from this table why simple physical line modi-
fications (such as raising support points or using float-
ing dead-ends) are an effective means of uprating older
lines with relatively low design temperatures. Even mod-
est increases in allowable sag result in relatively large
increases in rating for such lines. It is also clear why such
techniques are not usually helpful in uprating new lines
having higher line design temperatures.
2.3.5 Heat Balance Methods
Around the world, utilities perform overhead line rating
calculations in essentially the same way: by setting the
heat input from Ohmic losses and solar heating equal to
the heat loss due to convection and radiation (EPRI
1995). The specific formulas used to determine the heat
balance terms vary somewhat, but normally one of three
methods is used the IEEE method (IEEE 1993), the
CIGRE method (CIGRE 1992), or the EPRI
DYNAMP method (Black et al. 1983).
Given the same assumed wind speed and direction, the
same conductor temperature and the same conductor
electrical and physical parameters, the thermal rating
found with the three methods is similar if not identical.
To illustrate typical values of the heat balance terms, the
IEEE method is used in the following development for a
Drake ACSR conductor at 100
o
C.
Table 2.3-2 Variation in Line Rating with Design
Temperature
(for 795 kcmil [405 mm
2
], 26/7, Drake ACSR conductor
with an air temperature of 35
o
C, full summer sun, and a
wind blowing perpendicular to the conductor axis at 2
ft/sec [0.61 m/sec])
Line Design
Temperature Line Rating
Increase in Line Rating
for 10C Change in Line
Design Temperature
(
o
C) (amps) (amps) (%)
50 374 213 57
75 797 107 13
100 1039 78 7.5
125 1221 63 5.2
150 1370 54 3.9
2-18
Radiation
Radiation of heat from an overhead conductor is mod-
eled in terms of the absolute temperature of the conduc-
tor and its surroundings (taken as the air temperature):
2.3-1
As an example of radiation heat loss from a bare over-
head conductor, consider Drake ACSR at 100
o
C and an
air temperature of 40
o
C:
2.3-2
q
r
= 24.44 W/m
q
r
= 7.461 W/ft
Convection
Natural Convection
With zero wind speed, natural convection occurs, where
the rate of heat loss is:
2.3-3
Taking our example of Drake ACSR at 100
o
C, the natu-
ral convection heat loss is:
2.3-4
where: where:
D = 28.14 mm D = 1.108 in.
T
c
= 100C T
C
= 100
o
C
T
a
= 40C T
a
= 40
o
C
f
= 1.029 kg/m
3
f
= 0.0643 lb/ft
3

q
c
= 0.0205 (1.029)
0.5
q
c
= 0.283 (0.0643)
0.5

(28.14)
0.75
(1.108)
0.75

(10040)
1.25
(100-40)
1.25
= 42.4 W/m = 12.9 W/ft
Forced Convection
With the IEEE 738 and the CIGRE methods, forced
convection is calculated with two separate formulas, and
the larger of the two values for forced convection heat
loss is used.
2.3-5
2.3-6
The first two equations apply at low winds, but are too
low at high speeds. The last two equations apply at high
wind speeds, being too low at low wind speeds. At any
wind speed, the larger of the two calculated forced con-
vection heat loss rates is used.
The convective heat loss rate is multiplied by the wind
direction factor, K
angle
, where is the angle between
the wind direction and the conductor axis:
2.3-7
0.0178 /
0.138 /
4
+ 273
T
c

100
= D W m q
r
4
+ 273
T
a
100
4
+ 273
T
c

100
= D W ft q
r
4
+ 273
T
a
100
=
4
100
313
4
100
373
5 . 0 108 . 1 138 . 0
4
100
313
4
100
373
5 . 0 14 . 28 0178 . 0
r
q
r
q
0 05 /
83 /
0.5 1.25
0.75
= 0. 2 ( - W m q )
D T T
c a
c
f
0.5 1.25
0.75
= 0.2 ( - W ft q )
D T T
c a
c
f

25 . 1
) (
5 . 0
283 . 0
) 5 (
25 . 1
) (
75 . 0 5 . 0
0205 . 0
a
T
c
T
f c
q
s
a
T
c
T D
f c
q
=
=
100 + 40 100 40
= = 70
70
2 2
o
film film
C T C
T
+
= =
ft W
a
T
c
T
f
k
f
w
V
f
D
c
q
m W
a
T
c
T
f
k
f
w
V
f
D
c
q
/ ) (
52 . 0
371 . 0 01 . 1
1
/ ) (
52 . 0
0372 . 0 01 . 1
1

+ =

+ =
( )
( )
ft W
a
T
c
T
f
k
f
w
V
f
D
c
q
m W
a
T
c
T
f
k
f
w
V
f
D
c
q
/
6 . 0
1695 . 0
2
/
6 . 0
0119 . 0
2

=

=
cos cos 0.368 sin (2 )

angle
= 1.194 - ( ) + 0.194 (2 )
K
+
2-19
Referring once again to our example of Drake ACSR at
100
o
C we have:
2.3-8
2.3-9
Now select the larger of the two calculated convection
heat losses.
q
c
= 82.295 W/m q
c
= 25.052 W/ft
Since the wind is perpendicular to the axis of the con-
ductor, the wind direction multiplier, K
angle
, is 1.0, and
the forced convection heat loss is greater than the natu-
ral convection heat loss. Therefore, the forced convec-
tion heat loss will be used in the calculation of thermal
rating.
Notice that if the wind had been nearly parallel to the
line at 10
o
from the line direction, the wind direction
multiplier would be 0.517, and convection cooling
would have been similarly reduced.
Also, notice that the heat loss with no wind (natural
convection) is much less than that for forced convection
with even the low 2 ft/sec (0.61 m/sec) wind.
Solar Heating
Overhead conductors are typically 5
o
C to 10
o
C above
air temperature due to solar heating alone, even if the
current in the conductor is zero. The conductor heat
balance described in these notes applies when there is
only solar heat input as well as when the conductor car-
ries electrical current. The solar heat into the conductor
in direct sun is a function of the solar heat flux density,
the angle of the solar beam relative to the line direction,
and the conductor absorptivity (the fraction of incident
solar radiation absorbed by the conductor). The result-
ing temperature rise above air temperature is a function
of the conductor absorptivity and diameter as well as
the wind speed and direction.
According to the IEEE solar model, the maximum pos-
sible conductor solar temperature rise above air temper-
ature is on the order of 15
o
C (for still air). Field
measurements of actual lines indicate that the typical
solar rise is in the range of 5C to 10C.
With reference to Table 2.3-3, with a solar altitude of 70
degrees (noon in June in New York) and clear atmo-
sphere, the solar heat flux to a surface perpendicular to
the suns rays is approximately 1020 watts/ft
2
or 95
watts/m
2
. The maximum heat input to the conductor is
therefore:
2.3-10
Ohmic Losses
Conductor resistance per unit length and the electrical
current on the line determine the Ohmic losses. The
resistance of a stranded conductor is a function of the
conductivity of the component wires, the frequency, the
current density, the temperature of the wires, and the
stranded construction. Resistance values at 25C and
75
o
C are readily available from the manufacturer or the
Aluminum Association.
The IEEE standard suggests that electrical resistance
may be calculated solely as a function of conductor tem-
perature, ignoring dependence on current density. For
example, the values of conductor resistance at high tem-
perature, T
high
, and low temperature, T
low
, may be taken
from the tabulated values in the Aluminum Association
Handbook (Aluminum Association 1989). The conduc-
28.14 1.029 .6096
1.01 0.0372
1
5
2.04 10
.0295 (100-40) = 82.295 W/m
0.52
1.108 0.643 7200
1.01 0.371
1
0.0494
.00898 (100-40) = 25.052 W/ft
0.52
= + q
c
q
c

0.6
2
5
2
0.0119
28.14 1.029 .6096
2.04 10
.0295 (100-40) = 76.88 W/m
0.6
1.108 0.0643 7200
0.1695
0.0494
.00898 (100-40) = 23.464 W/ft
c
c
= q
q
Table 2.3-3 Total Heat Flux Received by a Surface at Sea

Level Normal to the Suns Rays
Solar Altitude,
H
c
Q
S
for a Clear
Atmosphere
Q
S
for an Industrial
Atmosphere
Degrees (w/m
2
) (w/ft
2
) (w/m
2
) (w/ft
2
)
60 1000 92.9 771 71.6
70 1020 95.0 809 75.2
80 1030 95.8 833 77.4
90 1040 96.4 849 78.9
W/ft 4 . 4 092 . 0 95 5 . 0
W/m 3 . 14 0281 . 0 1020 5 . 0
= =
= =
C
C
q
q
2-20
tor resistance at any other temperature, T
C
, is found by
linear interpolation according to Equation 2.3-11.
2.3-11
This method of resistance calculation allows the user to
calculate the high and low temperature resistance values
by whatever means is appropriate. See, for example, ref-
erences (Douglass and Rathbun 1985 and (Lewis and
Tuttle 1958).
In the example calculation, the resistance of the Drake
ACSR conductor is calculated for a conductor tempera-
ture of 100
o
C:
2.3-12
Steady-State Thermal Rating
Now that the radiation heat loss, the convective heat
loss, the solar heat input and the resistance of the con-
ductor have been determined, the steady-state thermal
rating can be calculated as follows:
2.3-13
For the example case:
q
c
= 82.295 W/m q
c
= 25.052 W/ft
q
r
= 24.44 W/m q
r
= 7.461 W/ft
q
s
= 14.0 W/m q
s
= 4.26 W/ft
(from equations) (from equations)
R(100) = 9.39010
-5
/m R(100) =2.86210
-5
/ft
Thermal Rating Dependence on Conductor Parameters
The rating of bare overhead conductors depends on the
various conductor parameters including (see Table
2.3-4):
Outside diameter
Emissivity and absorptivity
Electrical resistance per unit length
At the time of construction, the choice of conductor
type and size defines the resistance and outside diame-
ter. Normally, the emissivity and absorptivity of new
aluminum conductor are initially in the range of 0.2 to
0.3 but increase to values close to 1.0 as the conductor
ages. Figure 2.3-1 shows this increase in emissivity with
time for energized conductors.
The actual rate at which the conductor emissivity and
absorptivity increase with time is a function of the line
voltage and the density of particulates in the air. Two
observations, however, can be made. The emissivity and
absorptivity are correlated, so it is unlikely that one
parameter will be high and the other low. Also, new con-
ductors will have emissivity and absorptivity values in
the range of 0.2 to 0.3, and old conductors will have val-
ues in excess of 0.5.
As stated above, resistance and diameter are tightly cor-
related. Thus, aluminum stranded conductors of a given
diameter will have a corresponding resistance per unit
length. The exceptions to this are:
The component strands have a different conductivity
from that of standard aluminum (e.g., copper).
Conducting strands are trapezoidal rather than
round (e.g., TW conductor).
The steel core strands are not used or are replaced by
aluminum-clad steel wires (e.g., ACSR/AW).
)
T
low
R( + )
T
low
-
T
c
( *
T
low
-
T
high
)
T
low
R( - )
T
high
R(
= )
T
c
R(
( )
m
R R
R R
/ 10 390 . 9
75
50
5
10 283 . 7
5
10 688 . 8
5
10 283 . 7
25 100
25 72
) 25 ( ) 75 (
) 25 ( ) 100 (
5
=
+ =
( )
ft
R R
R R
/ 10 862 . 2
75
50
10 220 . 2 10 648 . 2
10 220 . 2
25 100
25 72
) 25 ( ) 75 (
) 25 ( ) 100 (
5
5 5
5
=

+ =

+ =
(100) (100)
c r s c r s
q q q q q q
I I
R R
+ +
= =
A A
I I
994 994
10 862 . 2
26 . 4 461 . 7 052 . 25
5 5
10 390 . 9
0 . 14 44 . 24 295 . 82
= =
+
=

+
=
Table 2.3-4 Illustration of the Effect of Diameter,
Resistance, Emissivity and Absorptivity on Thermal Rating
Conductor
Description
Outside
Diameter
(in.)
Resis-
tance
@ 25 C
(Ohms/mi)
Emissivity
and
Absorptivity
Thermal
Rating (A)
Drake 1.108 0.1170 0.5 & 0.5 996
Drake/TW 1.010 0.1170 0.5 & 0.5 976 (-2.0%)
Drake/AW 1.108 0.1129 0.5 & 0.5 1014 (+1.8%)
Arbutus AAC 1.026 0.1200 0.5 & 0.5 962 (-3.4%)
CU 500
kcmil
0.811 0.1196 0.5 & 0.5 909 (-8.7%)
Drake 1.108 0.1170 0.9 & 0.9 1046 (+5.0%)
Drake 1.108 0.1170 0.3 & 0.3 971 (-2.5%)
2-21
2.3.6 Thermal RatingsDependence on Weather
Conditions
It is clear from the preceding discussion that the thermal
rating of an overhead line depends on the weather con-
ditions along it, as well as on the type of conductor and
its maximum allowable operating temperature. Many
utilities around the world adjust their line ratings for
seasonal variation in air temperature, recognizing that
air temperature is lower and ratings can be higher in the
winter than in the summer. Of course, in areas where the
seasonal change is small (near the equator), or where
the fluctuations in any season are larger than the sea-
sonal average difference, this does not make sense. Other
utilities adjust thermal ratings for day and night by
including or ignoring solar heating, and others adjust
the wind speed, using a more conservative (lower) wind
speed for continuous ratings than for emergency ratings,
which tend to have a low probability of occurrence.
Many utilities have installed real-time monitoring sys-
tems, adjusting their line ratings for actual real-time
wind speed, wind direction, solar heating, and air tem-
perature. This technique is discussed in more detail in
Section 2.8.
In order to illustrate the effect of changing weather con-
ditions on ratings, consider Table 2.3-5.
Figure 2.3-1 Transmission line conductor emissivity as a function of time (House et al. 1963)
Table 2.3-5 Effect of Weather Conditions on Thermal
Ratings. (In all cases, the conductor is 26/7 795 kcmil
(0.61 mm
2
) ACSR (Drake) with emissivity = absorptivity =
0.5, Direct sun on June 10, clear air, at sea level, latitude =
40 with the conductor at 100C.)
Air
Temperature
(C)
Wind
Speed
(ft/sec)
Wind Direction
Relative to the
Line
(90 =
Perpendicular)
Time of
Day
Thermal
Rating
Amperes
40 2 90 2 PM 996
40 2 90 12 PM
986
(-0.8%)
40 2 90 6 PM
1045
(+4.9%
30 2 90 2 PM
1081
(+8.5%)
40 0 90 2 PM
838
(-15.6%)
40 3 90 2 PM
1183
(+18.7%)
40 6 10 2 PM
968
(-2.8%)
2-22
By reviewing this limited series of rating calculations, a
number of important aspects of line rating dependence
on weather can be drawn:
Rating variation due to solar heating changes
throughout the day is less than 5%.
Air temperature variation is important. A difference
of 10C in air temperature causes a line rating change
of nearly 10%.
Relatively small differences in wind speed, in the
range of 0 to 3 ft/sec (0.91 m/sec) can make a big dif-
ference in the line rating, generally 10% to 20%.
The wind direction relative to the line is as important
as the speed. A 6 ft/sec (1.8 m/sec) wind blowing near
parallel to the line (10) yields a slightly lower line
rating than a 2 ft/sec (0.61 m/sec) wind blowing per-
pendicular to the line.
2.3.7 Transient Thermal Ratings
The need for increased thermal capacity in overhead
lines is often driven by occasional, sharp increases in
load after certain system contingencies. For example, an
HV line might only reach high current levels after the
loss of an EHV line or a critical generating facility. Since
these occasions of high load occur infrequently and may
persist for short time periods, it is often useful to con-
sider transient thermal ratings for lines.
The temperature of an overhead power conductor is
constantly changing in response to changes in electrical
current and weather. In this method, however, weather
parameters (wind speed and direction, ambient temper-
ature, etc.) are assumed to remain constant, and any
change in electrical current is limited to a step change
from an initial current, I
i
, to a final current, I
f
, as illus-
trated in Figure 2.3-2.
Immediately prior to the current step change (t = 0),
the conductor is assumed to be in thermal equilibrium.
That is, the sum of heat generation by Ohmic losses and
solar heating equals the heat loss by convection and
radiation.
Immediately after the current step change (t = 0+), the
conductor temperature is unchanged (as are the conduc-
tor resistance and the heat loss rate due to convection
and radiation), but the rate of heat generation due to
Ohmic losses has increased. Therefore, at time t = 0+, the
temperature of the conductor begins to increase at a rate
given by the non-steady-state heat balance equation.
As time passes, the conductor temperature increases,
yielding higher heat losses due to convection and radia-
tion, and somewhat higher Ohmic heat generation due
to the increased conductor resistance. After a large num-
ber of thermal time constants, the conductor temper-
ature approaches its final steady-state temperature (T
f
).
The transient thermal rating is normally calculated by
repeating the preceding calculations of Tc(t) over a
range of I
f
values, then selecting the If value that causes
the conductor temperature to reach its maximum allow-
able value in the allotted time.
The transient thermal rating of an overhead line is
dependent on the duration of the elevated current, the
maximum temperature that the conductor is allowed to
attain during the rating period, and on the starting tem-
perature of the conductor. For example, with the Drake
ACSR that was used for rating calculations previously,
the transient ratings for various rating durations, maxi-
mum temperatures, and starting temperatures are as
shown in Table 2.3-6.
The advantage to using transient ratings is that the line
can be loaded above its continuous rating without vio-
lating the constraints on sag clearance or annealing, but
Figure 2.3-2 Temperature response of a bare overhead
conductor to a step-change in current.
2
c 2
c
s c r
p
dT
c
+ + = + R ( ) q q q
mC I T
c p
c r s
dt
1
dT
= R( ) + - - q q q
T I
dt
mC

Table 2.3-6 Transient Ratings Versus Rating Duration
Rating
duration
Maximum
temp Starting Temp Rating
(Minutes) (
o
C) (
o
C) (A)
continuous 100 N/A 1040
60 100 50 1045
30 100 50 1090 (+4.8%)
15 100 50 1230 (+18.3%)
15 100 75 1135 (+9.1%)
2-23
the drawback is that the load must be reduced to the
continuous rating or below within a short time (15 to 30
min). See, for example, (Black and Rehberg 1985) and
(Davidson 1969).
2.4 EFFECTS OF HIGH-TEMPERATURE
OPERATIONS
2.4.1 Introduction
For any given high-voltage conductor, there are usually
at least two current limits specifiedthe conductors
normal rating, and the emergency rating (see Section
2.3). The conductors normal rating specifies how much
current may flow in the circuit on a continuous basis,
whereas the emergency rating specifies how much cur-
rent can flow under emergency conditions for a specified
amount of timee.g., 30 minutes. A typical emergency
rating may be applicable in the case where there is an
unexpected outage on a parallel circuit requiring a
short-duration increase in the load flow.
The above normal current flow that is required during
emergency loading will, if it remains unchecked, result in
a thermal overload of the circuit, or significantly
reduced clearances leading to a flashover of the circuit.
Regardless of the case, the two issues that require atten-
tion are the loss of conductor strength and increased
conductor sag. Ideally, transmission line operators aim
to maximize the load on a particular circuit while mini-
mizing the annealing (softening) of the circuits conduc-
tor. Annealing causes a decrease in the conductors
strength and performance, necessitating the eventual
replacement of this component. Because of the difficulty
associated with taking a line out of service, and the large
expense associated with the replacement of a circuits
conductor, the operator clearly needs to balance the
need for increased load flow with the economic risk
associated with the premature replacement of the com-
ponent, and the loss of service life or the safety risk
associated with providing inadequate clearances.
This section reviews issues related to the effects of high-
temperature operation, including annealing, calculation
of sag and tension, thermal and creep elongation, and
connectors and conductor hardware at high temperature.
2.4.2 Annealing of Aluminum and Copper
The American Society for Testing and Materials
(ASTM) standards and the International Engineering
Consortium (IEC) standards specify the minimum ten-
sile strength of aluminum and copper wires, which is the
stress at which the wire breaks. For aluminum and cop-
per wires, the tensile strength of the materials decreases
with time if operated at temperatures above 75C, while
the tensile strength of galvanized, aluminum-clad, or
copper-clad steel wires remains constant at tempera-
tures below 300C. Thus, the extended exposure of con-
ductors made up largely of aluminum or copper wires to
temperatures above 75
o
C can eventually lead to tensile
failures during high ice and/or wind loading events.
Table 2.4-1 shows experimental results for a test con-
ducted by Troia (Troia 2000). The high-temperature
simulation used an ACSR conductor Raven with a 6/1
stranding ratio. In this study, four sets of conductors
and connectors, and three sample loops, were operated
at a temperature of 100C, and cycled for 125, 250, and
500 cycles, respectively. The fourth set of samples was
continued to 1000 current cycles at the 100C tempera-
ture rise. A second group of four sets of sample test
loops were operated at 175C, and the current cycle
counts were 125, 250, 500, and 1000. Upon completion,
a Rockwell H scale was used to measure the hardness of
each sample, and as expected, the average hardness was
determined to be directly proportional to the tempera-
ture and duration of heating. Conductors heated to
100C exhibited a decrease in hardness of up to 22%
after 1000 cycles, and conductors heated to 175C
showed a decrease in hardness of up to 92.5% after only
250 cycles. Thus, annealing of the conductor was exten-
sive at the higher operating temperature, and hardness
readings could not be measured for conductors having
been cycled 500 or 1000 cycles.
Clearly, the results of this and other studies indicate that
the prolonged operation of high-voltage conductors
including ACSR at very high temperatures reduces the
Table 2.4-1 Hardness Results
a
a. Rockwell H scale, 1/8 in. ball, 60 kgs.
Conductor Cycle Temp Average Hardness % Difference Temp Average Hardness % Difference
ACSR (6/1) Raven
0 - 33 - - 37.3 -
125 100 23.6 28.5 175 15.9 57.4
250 100 19.4 41.2 175 2.8 92.5
500 100 20.8 37.0 175 0
b
b. Material hardness too soft for accurate readings on Rockwell H scale.
-
1000 100 22.6 31.5 175 0
b
-
2-24
mechanical strength, integrity, and performance of the
overhead system. It is also clear that the damage to the
aluminum is of a cumulative nature and that the pro-
longed high temperature operation will significantly
reduce the expected service life of the delivery system.
Figure 2.4-1 shows the influence of prolonged high-
temperature operation on the tensile strength of alumi-
num conductor. The graph shown has been developed
based on test data of 1350-H19 EC hard drawn
aluminum wires. In general, tensile strength reduction of
aluminum wires at temperatures of less than 90
o
C is
considered negligible, and the effect of prolonged opera-
tion at this temperature will have very little effect on the
service life of the aluminum conductor. At 100
o
C, the
tensile strength of the wire is reduced by 10% after 5000
hours, equivalent to a little more than a half a year, and
at 125
o
C, the tensile strength of the aluminum conduc-
tor is reduced 10% after 250 hours, a little more than 10
days of continuous operation. The effects of the high-
temperature operation on the aluminum conductor are
irreversible, and the damages experienced by the
conductor are cumulative.
Aluminum anneals at a slower rate than copper wire
when exposed to identical conditions. Even though the
use of copper conductors has significantly decreased
over the years, a large number of low- to medium-volt-
age lines with copper conductors are still operated. The
issues associated with the prolonged operation of these
wires at very high temperatures are similar to the issues
encountered with aluminum conductors.
For example, Figure 2.4-2 shows the reduction in tensile
strength with time and temperature for a sample of 0.081
in. (0.2 cm) diameter hard drawn copper wire. Since
there are 8760 hours in a year, the logarithmic diagram
clearly shows that the sustained operation of the copper
wire at 65
o
C yields no measurable reduction in the ten-
sile strength, while the sustained operation of a copper
wire at 100
o
C yields a 10% reduction in the tensile
strength in 600 hours (25 days). More critically, the oper-
ation of the same wire at a temperature of 125
o
C for less
than 40 hours reduces the wire tensile strength by 10%.
The remaining strength of AAC, AAAC, ACAR, and
ACSR conductor wires can be estimated with the fol-
lowing predictor equations. The use of these equations is
acceptable even in the case in which several emergency-
rating episodes have occurred.
Figure 2.4-1 Annealing of 1350-H19 hard-drawn
aluminum wire.
Figure 2.4-2 Annealing of 0.081in. diameter hard-drawn
copper wire (Southwire).
2-25
Definition of Terms:
RS
1350
= Residual aluminum (1350 Alloy) strength as a
percentage of initial strength [%].
RS
6201
= Residual 6201 Alloy strength as a percentage
of initial strength [%].
RS
COM
= Residual strength of composite conductor as
a percentage of initial strength [%].
T = Temperature [C].
t = Elapsed time [hours].
d = Strand diameter [mm, in.].
A
1350
= Area of aluminum (1350 Alloy) strands [sq
mm, sq in.].
A
6201
= Area of 6201 alloy strands [sq mm, sq in.].
A
T
= Total area [sq mm, sq in.].
STR
1350
= Calculated initial strength of the aluminum
(1350 Alloy) strands [N, lbs].
STR
ST
= Calculated initial strength of the steel core
[N, lbs].
STR
T
= Calculated initial strength of the conductor
[N, lbs].
Predictor Equations: (Metric)
AAC:
2.4-1
If (-0.24T + 134) > 100, use 100 for this term.
AAAC:
2.4-2
ACAR:
2.4-3
ACSR:

2.4-4
ACSR:
2.4-5
Predictor Equations: (English)
AAC:
2.4-6
AAAC:
2.4-7
ACAR:
2.4-8
ACSR:
2.4-9
As previously shown, when applying these equations,
the cumulative strength reduction for multiple expo-
sures at the same conductor temperature is additive;
however, this is not true for multiple exposures at differ-
ent conductor temperatures. To determine the cumula-
tive strength reduction for a series of high-temperature
exposures at different temperatures and times, each of
the exposures must be expressed in equivalent time at
the highest temperature experienced by the conductor
before finding the equivalent time. The following exam-
ples illustrate a possible scenario that an ACSR and
AAC conductor might experience during one year of
service.
Exampl e 2. 4- 1: The conduct or i s 795 kcmi l
(405 mm
2
)ACSR Drake. During one year of opera-
tion, it is subjected to 7500 hours at 75C, 1200 hours at
100C, 50 hours at 125C, and 10 hours at 150C. What
is the remaining strength (RS) of the conductor?
Using Equation 2.4-5, we know the following equation:
ACSR:
and, if (134 0.24T) >100, use 100; if (0.241- 0.00254T)
> 0, use 0.
a) 7500 hours at 75C
7500 hours at 75C has 100% remaining, which equals 0
minutes at 150C.
( )
( )

+ =
d
RS
54 . 2
0.095 - T 0.001
1350
t 134 T 0.24 -
( )
( )
+ =
d
RS
54 . 2
0.118 - T 0.0012
6201
t 176 T 0.52 -
( ) ( )

=
T T
COM
A
A
A
A
RS
6201
6201
1350
1350
RS RS
( ) ( )

=
T
ST
T
COM
STR
STR
STR
STR
RS 109 RS
1350
1350
( )
( ) d
COM
RS
1 . 0
0.00254T - 0.241
t 0.24 - 134 =
( )
( ) d
COM
RS
1 . 0
0.095 - 0.001T
t 0.24 - 134

=
( )
( ) d
COM
RS
1 . 0
0.118 - 0.0012T
t 176 0.52T -

+ =
( ) ( )

=
T T
COM
A
A
A
A
RS
6201
6201
1350
1350
RS RS
( ) ( )

=
T
ST
T
COM
STR
STR
STR
STR
RS 109 RS
1350
1350
( )
0.1
(0.241 - 0.00254T)
134-0.24 t
d
COM
RS

=
( )
( )
0.1
0.241 - 0.00254T
0.1
-(0.241 0.00254T)
1.108
134-0.24 t
(134 - (0.24 x 75)) x (7500)
100%
d
COM
RS

=
=
=
2-26
b) 1200 hours at 100C
1200 hours at 100C has 99% remaining, which equals
20 minutes at 150C.
c) 50 hours at 125C
hours at 150C.
d) 10 hours at 150C
10 hours at 150C has 95% remaining.
To calculate the total loss of conductor strength, the
sum of the equivalent times (hours) at 150C is found.
In this example, the sum is:
The remaining strength for the in-service conductor is
95%.
Example 2.4-2: The conductor is 37 kcmil (19 mm
2
)
AAC Arbutus. During one year of operation, it is sub-
jected to 7500 hours at 75C, 1200 hours at 100C, 50
hours at 125C, and 10 hours at 150C. What is the
remaining strength (RS) of the conductor?
Using Equation 2.4-6,
AAC:
and, if (134 0.24T) >100, use 100.
a) 7500 hours at 75C
7500 hours at 75C has 98.3% remaining, which equals
1 hour at 150C.
b) 1200 hours at 100C
1200 hours at 100C has 87.4% remaining, which equals
125 hours at 150C.
c) 50 hours at 125C
hours at 150C.
d) 10 hours at 150C
10 hours at 150C has 93% remaining.
To calculate the total loss of conductor strength, the
sum of the equivalent times at 150C is found. In this
example, the sum is:
The remaining strength for the in-service conductor is
87%.
As expected, the example calculations predict that the
continuous high-temperature operation will cause the
tensile strength of the AAC conductor to deteriorate
faster than in the case of the ACSR conductor. In Fig-
ure 2.4-1, the graph showing the loss of strength for
0.1
1.108 -(0.241 - 0.00254T)
(134 - (0.24 x 100)) x (1200)
99.17%

=
=
0.1
1.108 -(0.241 - 0.00254T)
= (134 - (0.24 x 125)) x (50)
= 99.17%

0.1
1.108 -(0.241 - 0.00254T)
= (134 - (0.24 x 150)) x (10)
= 95.19%

0.1
0.1
1.108
-(0.241 - 0.00254T)
COM
-(0.241 - (0.00254 x 150))
0 + 0.3 + 2 + 10 = 12.3 hours at 150 C.
RS = (134 - 0.24T) t
= (134 - (0.24 x 150))
x (12.3)
= 95%
d

( )
( )
0.1
0.001T - 0.095
134-0.24 t
d
COM
RS

=
0.1
d
0.1
1.026
-(0.001 T - 0.095)
1350
-(0.001(75) - 0.095)
RS = (134 - 0.24 T) t
= (134 - (0.24 x 75)) x (7500)
= 98.3%

0.1
1.026 -((0.001 x 100) - 0.095)
= (134 - (0.24 x 100)) x (1200)
= 87.4%

0.1
1.026 -((0.001 x 100) - 0.095)
= (134 - (0.24 x 125)) x (50)
= 91.95%

0.1
1.026 -((0.001 x 100) - 0.095)
= (134 - (0.24 x 150)) x (10)
= 92.8%

0.1
0.1
1.026
-(0.001 T - 0.095)
1350
-(0.001(150) - 0.095)
1 + 125 + 15 + 10 = 151 hours at 150 C.
RS = (134 - 0.24 T) t
= (134 - (0.24 x 150)) x (151)
= 86.9%
d

2-27
individual strands of aluminum indicates that after 150
hours the remaining strength of the hard-drawn alumi-
num wire is 82%. When compared to the remaining
strength value calculated using the predictor equations,
the AAC conductor appears to be 5% stronger than pre-
dicted by using the data of the individual aluminum
strands. Much of this difference in the prediction of the
strength loss in the AAC conductor can be attributed to
differences in behavior between the stranded conductor
and an individual strand of aluminum and to differences
in the nominal and actual dimensions of the manufac-
tured product.
Similarly, using strand data, the prediction equations
show the ACSR conductor to be approximately 13%
stronger than the tensile strength predicted by using the
individual aluminum strand data itself. The difference in
the remaining tensile strength can be directly attributed
to the presence of the steel reinforcing strands which are
significantly stronger than a comparably sized alumi-
num strand and are also not affected by temperatures
below 300C.
2.4.3 Sag Tension Models for ACSR Conductors
The traditional sag and tension model used throughout
the industry is the model developed and promoted by
the Alcoa-Fujikura LLC (Alcoa) (Alcoa 2003). The
Alcoa sag and tension prediction model and methods
assume that the magnitude of the compression stresses
resulting from the manufacturing process are negligible
and therefore do not affect the behavior of the conduc-
tor regardless of the operating temperature. Therefore,
the model ignores the effects of aluminum compression.
The Alcoa sag and tension model focuses on the coeffi-
ci ent of thermal expansi on phenomenon, whi ch
accounts for the fact that the aluminum strands expand
at nearly twice the rate of the steel strands for the same
increase in temperature. Consequently, as the tempera-
ture of the conductor increases significantly, the alumi-
num strands expand more rapidly than the steel strands.
This shifts a continuously increasing percentage of the
catenary tension onto by the steel strands (i.e., the alu-
minum unloads its share of the original tension). At
some point (i.e., commonly called the knee point) the
tension in the aluminum strands approaches a value of
zero. At this time, the steel strands of the conductor
carry all of the catenary tension of the composite wire.
Based on this premise, and once the knee point temper-
ature has been exceeded, the sag of the ACSR conduc-
tor is proportional to the rate of thermal expansion of
the steel strands. It should be noted that the Alcoa sag
and tension prediction model was developed in the early
part of the last century when the stranding of the most
commonly used ACSR conductors was six aluminum
strands to one steel strand. Nevertheless, the sags and
tensions predicted by the Alcoa model generally corre-
late very well for most conductors operated at tempera-
tures not exceeding 75C to 100C.
Today, medium-to-large ACSR conductors may have as
many as four layers of aluminum strands surrounding a
multistranded steel core. Based on the construction
methods used to manufacture these multilayered con-
ductors, it is very likely that a significant number of alu-
minum strands are capable of supporting limited
compression forces (as a result of the confinement pro-
vided by underlying and overlaying layers) and that the
resulting conductor harbors noticeable built-in stresses.
The EPRI technical report Conductor and Associated
Hardware Impacts during High Temperature Operations
Issues and Problems, published in December 1997 by
Shan and Douglass (EPRI, TR-109044) concluded that
it is more likely that such multilayered conductors fol-
low a sag and tension model originally proposed by
Nigol (Nigol and Barrett 1980).
The Nigol sag and tension model, as shown in Figure
2.4-3, assumes that the aluminum strands do not change
from tension to compression until a fixed limiting value
is reached. In Nigols sag and tension model, as the alu-
minum changes from tensile to compression stresses, the
elastic modulus is assumed not to change. Nigol
hypothesizes the existence of compression stresses in the
aluminum below the birdcaging temperature but
assumes that the outer layer of aluminum strands
remains in tension. Therefore, Nigol assumes that the
tension in the outer layer strands denies the inner layer
to expand radially and buckle outward. Nigol concludes
that, since the inner layer is in compression and the
outer layer is in tension, and since the outer layer
presses radially inwards against the expanding inner
layer, the aluminum strands of the inner layer are
pressed against the steel cores and birdcaging is effec-
tively negated. Consequently, based on this hypothesis,
Nigol concluded that the elastic modulus remains
unchanged even though the net aluminum stresses
change from tension to compression.
Thus, Nigols sag-tension model assumes that, as the
conductor temperature increases towards the birdcaging
temperature, the tensions in the outer layers of alumi-
num strands and the corresponding inward directed
radial forces decrease. At the same time, Nigols model
assumes that there is an increase in the radial forces of
the outer layer and also an increased axial compression
in the strands of the inner layers. Nigol, therefore,
concludes that at the time when the birdcaging tempera-
ture is reached, the inner and outer radial forces are
2-28
balanced, and the compressive stresses that are created
reach their limiting value.
Barretts sag and tension prediction model (Barrett et al.
1982) departs from Nigols model at this point. Barrett
hypothesizes that at the birdcaging temperature, the
radial forces within each of the aluminum strand layers
balance. Based on Barretts hypothesis, the aluminum
strands of a layer move radially outward as the tempera-
ture rises, resulting in a different elastic modulus to be
used for the aluminum strands. Therefore Barrett
describes the use of two compression modulithe first
value to be used at temperatures below the knee point
and the second to be used at temperatures above the
birdcaging temperature.
Contrary to Nigol and Barrett, Rawlins (Rawlins 1998)
proposes that there are no compression stresses above
the knee point, but rather promotes the idea of large
built-in tensile stresses that are the result of the manu-
facturing process. In Rawlinss sag and tension predic-
tion model, the point at which the aluminum stresses
become tensile is also the point when there is a change
in the elastic modulus. Therefore, Rawlins proposes that
these stresses cause permanent elongations when com-
pared to the other sag and tension models at the same
level of tension. Figure 2.4-3 shows graphical represen-
tations of the four hypotheses explaining the behavior of
high temperature conductors.
2.4.4 Axial Compressive Stresses
Work performed previously by Nigol and Barrett (Bar-
rett 1982) clearly showed that compression stresses in
the aluminum strands are capable of greatly contribut-
ing to the sag of conductors at high operating tempera-
tures. A simplified illustration of the influence of
compression stresses in the aluminum strands at high
temperatures is illustrated in Figure 2.4-4.
The first sketch (Sketch 1) shows a length of the conduc-
tor at ambient temperature where the lengths of all of
the aluminum and steel strands are equal. The second
sketch (Sketch 2) shows an instance of the behavior of
the conductor once the conductor length is heated to a
temperature above ambient conditions. Because of the
differences in the coefficients of thermal expansion of
the aluminum and the steel strands (discounting for now
issues such as stranding, manufacturing, etc), the
lengths of the aluminum and steel strands will differ by
an amount proportional to the temperature difference
and the difference in the coefficients of thermal expan-
sion. Of course, this scenario assumes that each type of
material would be free to expand and not restricted in
any manner.
Similarly, the third sketch (Sketch 3) shows the behavior
of the same length of conductor heated to a temperature
significantly above ambient conditions, with the differ-
ence that the two materials are restricted from expand-
ing independently of each other. As a result of the
differences in the coefficients of thermal expansion, the
proportion of the stresses carried by each type of mate-
rial will change.
Finally, the fourth sketch (Sketch 4) shows the conduc-
tor in the case at which the compression in the alumi-
num strands is balanced by a tensile load in the steel
core (i.e., the phenomenon most commonly referred to
as aluminum compression). As the conductor reaches
the birdcaging temperature (i.e., the point at which the
aluminum strands expand outward to compensate for
the increase in length), the aluminum strands move out-
ward, yielding to the inherent compression stress
induced by the constrained thermal expansion. At this
Figure 2.4-3 Sag-tension models.
Figure 2.4-4 Manufacturing effects of ACSR conductor.
2-29
point (i.e., the point in the loading at which birdcaging
is first observed), the tensile stress in the steel core far
exceeds the value of the compression stress in the alumi-
num strands.
2.4.5 BuiltIn Stresses
When multistranded conductors are manufactured
using a combination of steel and aluminum wires, the
temperatures of the individual wire strands during the
manufacturing process may vary, especially that of the
steel core when compared to the aluminum strands.
Contrary to the manufacturing of these multimaterial
conductors, the same problem is not encountered in the
manufacturing of All Aluminum Conductors (AAC),
since they are constructed from the same homogeneous
material. For ACSR conductors, the variance in the
temperatures of the different types of stranding materi-
als contributes to the composite behavior of the conduc-
tor. Therefore, when the conductor is placed under load,
the variance in the temperature during manufacturing is
likely to impact the sag and tension characteristics of
the conductor. Unfortunately, quantitative values useful
in estimating these manufacturing effects are neither
provided by the manufacturer nor otherwise published.
For example, if one assumes that a cold steel core is
stranded with hot aluminum wires around the outside,
the resulting conductors steel core will be slightly
longer than the outer aluminum layers once the conduc-
tor is allowed to reach thermal equilibrium. As a result
of the temperature differences, the commonly named
knee point (i.e., the point at which the temperature of
the complete length and cross-section has increased to a
level at which the stresses in the aluminum strands
approach zero) will shift relative to the knee point of a
conductor constructed with steel and aluminum strands
at the same temperature.
2.4.6 Sag Tension Calculations
Bare overhead conductors are flexible and uniform in
weight. These characteristics allow conductors to
assume a catenary shape when suspended between sup-
port points. The shape of the catenary assumed by the
conductor changes continuously as a result of climatic
changes such as ambient temperature, operating condi-
tions such as the amount of current being transferred,
service life, and weather-related loads such as wind,
snow, and ice.
Because of the threat to human life and property, it is
critical to ensure adequate horizontal and vertical clear-
ances under all circumstances. Therefore the line
designer is challenged to consider all weather and elec-
trical loading cases to ensure that the breaking strength
of the conductor is not exceeded and that the suggested
minimum clearances are maintained.
In this subsection the mechanical and thermal proper-
ties acquired in the EPRI Conductor High Temperature
Project experiments are presented, and the calculated
sags and tensions are compared to values predicted
using traditional sag and tension prediction data, meth-
ods, and tools. For clarity, these comparisons have been
separated into two parts; the first addresses sags of level
spans, the second addresses sags for inclined spans.
To illustrate the use of the results and to demonstrate
the differences in the predicted sags and tensions, three
ACSR conductors of varying stranding ratios (low,
medium, and high aluminum- to-steel stranding ratios)
have been analyzed at different temperatures. The con-
ductor sag analyses and comparisons were made at
room temperature (23C), at a temperature of 120C,
and at a temperature of 150C. Also, sags computed at
level spans are compared to values at an inclined span of
15 and 30. The underlying principles behind sag ten-
sion calculations for both level and inclined spans are
outlined in the following subsections.
Sag and Tension of Level Spans
The shape of a catenary of a conductor is a function of
the conductors weight per unit length, w, the horizontal
component of tension, H, the span length, S, and the
conductor sag, D. Figure 2.4-5 shows an illustration
relating all of these parameters for a level conductor
span. The actual conductor length, L, constitutes the
stretched length of conductor (Trash et al. 1994).
Figure 2.4-5 Catenary sag model level span.
2-30
The exact catenary equation representing the height of
the conductor, y(x), at its lowest point along the span is
given by the following equation:
2.4-10
For level conductor spans, the low point is in the center
and the total sag, D, is found by substituting x = S/2 in
Equation 2.4-10. The resulting hyperbolic catenary rela-
tion for sag then becomes:
2.4-11
The horizontal component of tension, H, is located at
the point in the span where the conductor slope is hori-
zontal, or at the midpoint for level spans. The conductor
tension, T, is found at the ends of the spans at the point
of attachment and is calculated using the following
equation.
2.4-12
Rearranging the hyperbolic catenary equation for a level
span, along with the substitution of x = S/2 for level
spans corresponds to a conductor length of:
2.4-13
It should be noted that while the above equations
describe the behavior of ideal (with perfectly elastic stress
and strain characteristics) concentric-lay stranded con-
ductors, actual conductors such as ACSR conductors
exhibit nonlinear behavior when loaded from an initial
tension to some final value due to ice, wind, or tempera-
ture loading. Permanent elongation from creep and
heavy loading also affects the resulting sag. Also, high-
temperature operations result in thermal elongation of
the steel and aluminum strands, thus affecting sag.
Therefore, in order to calculate the correct sag, it is nec-
essary to separate the effects of conductor elongation
due to tensile loading as well as thermal loading. This
process requires an iterative procedure in which the
mathematical formulas describing the conductor elon-
gation caused by the temperature change are solved
simultaneously with the tension and conductor length
relationship. To calculate the change in length due to
temperature loading, the following equation is used:
2.4-14
where:
= coefficient of linear thermal elon-
gation for the AL/SW strands.
= final length of the conductor.
= reference length of the conductor.
= change in temperature.
The NESC (National Electric Safety Code 1993) recom-
mends limits on the tension of conductors based on a
percentage of their Rated Breaking Strength (RBS). For
example, the tension limits of an ACSR may be 60%
under maximum ice and wind loading, 35% upon instal-
lation at 60F, and 25% final unloaded after maximum
loading has occurred at 60F. Therefore, if the initial
tension in a span is known along with the initial conduc-
tor length, the total elongation resulting from the tensile
load is calculated as follows:
2.4-15
where:
= length of conductor under horizontal
reference tension.
tension.
= horizontal tension.
reference tension.
= modulus of elasticity of the conductor
(psi).
= cross-sectional area, in
2
.
It should be noted that the modulus E
c
in Equation
2.4-15 is the modulus of the steel and aluminum strands
determined by the stress and strain relationship of the
composite conductor. Since this relationship for a bime-
tallic construction of the conductor is quite complex,
the relationship of ACSR conductors is typically
described by a third or fourth order polynomial. In
addition, the thermal and permanent elongation drasti-
cally affects the mechanical behavior of ACSR conduc-
tors at high temperatures, and this further complicates
the analysis. Therefore, since ACSR conductors are
nonhomogenous, the stress and strain characteristics
are separated into their steel and aluminum components
as shown in Figure 2.4-6. Numerical methods are used
to calculate the resulting sags and tensions, with the aid
( ) [cosh( ) 1]
H wx
y x
w H
=
[cosh( ) 1]
2
H ws
D
w H
=
T H wD = +
2
sinh
2
H Sw
L
w H

=

Re
1 ( )
REF
T T AS r f
L L T T = +

AS
T
L
T
L REF
Re
( )
r f
T T
1
REF
REF
H H
c
H H
L L
E A

= +

REF
H
L
H
L
H
REF
H
c
E
A
2-31
of software programs such as Alcoa Fujikuras Sag10
package (Alcoa 2003).
2.4.7 Sag and Tension of Inclined Spans
For inclined spans, the length of the conductor between
supports is divided into two separate sections for con-
sideration. One is to the right of the lowest point of the
conductor, and one is to the left (see Figure 2.4-7).
The same equation used to represent the height of the
conductor, for a level span is valid for inclined spans and
is given by the following equation (Trash el al. 1994):
2.4-16
In each part of the span, the sag is dependent upon the
vertical distance between support points and can be
described by the following equations:
2.4-17
2.4-18
The maximum tension is
2.4-19
2.4-20
Figure 2.4-6 Decomposed ACSR Tern conductor at 120C.
Figure 2.4-7 Catenary sag modelinclined span.
( ) cosh( ) 1
H wx
y x
w H

=

2
1
4
R
h
D D
D

=

2
1
4
L
h
D D
D

= +

R R
T H wD = +
L L
T H wD = +
2-32
2.4.8 Calculation of Conductor
High-Temperature Sag and Tension
Simplified Calculations
Simplified calculations are used for demonstration using
a hand calculator. While most sag and tension calcula-
tions are typically performed with an analysis software,
the use of the simplified calculation clearly demon-
strates the process and provides the reader some insight
into how sags and tensions occur for overhead lines.
Example 2.4-3: What is the sag (D) and slack for a
1000-ft level span of 795 kcmil (405 mm
2
) ACSR
Drake conductor at ambient temperature (25C,
77F). The weight per unit length is 1.094 lbs/ft (1.6
kg/m), the horizontal tension component, H, is 25% of
the RBS.
H = 0.25 x 31,500 lbs = 7875 lbs (35,196 N)
Use Equation 2.4.11,
The sag for this level span is 17.37 ft (5.3 m).
The following equation is used to calculate the length of
the conductor:
The conductor slack is the length minus the span length,
0.80 ft (0.24 m).
Example 2.4-4: What happens to the conductor length if
the temperature increases from ambient to 50C (122F)?
What about if it increases to 150C (302F)? The coeffi-
cient of linear thermal expansion is 10.7 x 10
-6
/F.
At 50C, the length of the conductor is 1001.28 ft
(305.2 m). At 150C, the length of the conductor is
1003.21 ft (305.8 m).
Example 2.4-5: What is the sag of the conductor at the
elevated temperature levels?
The following equation is rearranged to estimate the
resulting sag, which only considers the change due to
thermal effects, and ignoring any changes that are due
to the changes in tension.
hence
Example 2.4-6: What is the tension of the conductor
when subjected to the two temperatures?
Rearrange the following equation to obtain the result-
ant tension of the conductor.
These values assume that the conductor has an infinite
modulus of elasticity. Actually, the elastic modulus of
the conductor is finite, and changes in the conductor
tension do affect the length of the conductor. Therefore,
these equations estimate sags that are greater than has
been observed in the field.
Estimating the actual change in sag due thermal effects
is complex, because one needs to look at the combined
thermal and elastic effects of conductors. The initial
loads of concentrically stranded ACSR conductors
result in elongation behavior that is different from that
caused by loading several years later. Software such as
Alcoas Sag10 program use numerical methods to esti-
mate the resulting sags.
2
[cosh( ) 1]
2 8
H ws wS
D
w H H
=
( )
( )
( )
2
2
1.094 1000
17.37 ft 5.29 m
8 8 7875
wS
D
H
= = =
( )
( )
( )
2
2
2
2 8
sinh
2 3
8 17.37
8
1000 1000.80 ft 305.05 m
3 3 1000
H Sw D
L S
w H S
D
L S
S

= +

= + = + =
( )
( )
( )
( )
( )
6
122
6
302
1
1000.80 1 10.7 10 122 77
1001.28 ft (305.2 m)
1000.80 1 10.7 10 302 77
1003.21ft (305.8 m)
REF
T T AS REF
L L T T
L
L
= +

= +

=
= +

=
( )
2
3
8
becomes
3 8
S L S
D
L S D
S
= + =
( )
( )( )
( )
( )
( )( )
( )
122
302
3 1000 1.28
21.9ft 6.68m
8
3 1000 3.21
34.7 ft 10.58m
8
D
D
= =
= =
2 2
becomes
8 8
wS wS
D H
H D
= =
( )
( )
( )
2
2
122
1.094 1000
6244 lbs (27,875 N)
8 8 21.9
wS
H
D
= = =
( )
( )
( )
2
2
302
1.094 1000
3941lbs (17,594N)
8 8 34.7
wS
H
D
= = =
2-33
The following example uses an AAC, and the effects of
the elevated temperature levels are estimated.
Example 2.4-7: What is the sag (D) and slack for a 1000
ft. (305 m) level span of AAC Arbutus conductor at
ambient temperature (25C, 77F). The weight per unit
length is 0.746 lbs/ft (1.09 kg/m), the horizontal tension
component, H, is 25% of the RBS.
The sag for this level span is 26.8 ft (8.18 m).
The following equation is used to calculate the length of
the conductor:
The conductor slack is the length minus the span length,
1.9 ft (0.58 m).
Example 2.4-8: What happens to the conductor length if
the temperature increases from ambient to 50C (122F)?
What about if it increases to 150C (302F)? The coeffi-
cient of linear thermal expansion is 12.8 x 10
-6
/F.
At 50C, the length of the conductor is 1002.48 ft
(305.56 m). At 150C, the length of the conductor is
1004.79 ft (306.26 m).
Example 2.4-9: What is the sag at the elevated tempera-
ture levels?
The following equation is rearranged to estimate the
resulting sag, which only considers the change due to
thermal effects, and ignoring any changes that are due
to the changes in tension.
hence
Example 2.4-10: What is the tension of the conductor
when subjected to the two temperatures?
Rearrange the following equation to obtain the result-
ant tension of the conductor.
These high readings reflect the thermal elongation
assuming the strands have an infinite modulus of elas-
ticity. Fortunately, the modulus is finite, and the changes
in the tension levels do affect the conductor length.
These equations are a useful way to get a rough idea of
how an ACSR conductor behaves when compared to an
AAC conductor. At 50C, the AAC elongates about
28% more than the ACSR, and at 150C, the AAC elon-
gates approximately 50% more than the ACSR.
Using Alcoa Fujikuras Sag10 Software
Software programs can be useful to obtain more accu-
rate results, since they use an iterative procedure to
obtain solutions to complex problems such as the sag
and tension calculation of a conductor. This iterative
process is commonly referred to as the binary chop
technique, and it separates the thermal as well as plastic
and elastic effects resulting from changes in the conduc-
tors tension.
Once the results of the iterative calculations have con-
verged, the results reflect the conductor length that
accounts for the elastic effects of an increased load as
well as the thermal elongation effects. Since H, and
H
REF
correspond to the first-, second-, third-, and
fourth-order composite stress and strain curves that
describe the initial and final conductor modulus, it is
typically convenient to rearrange the equations in terms
of strain elongation to permit the direct substitution of
values into the polynomial equations.
H = 0.25 x 13,900 lbs = 3475 lbs. (15,513 N)
( )
( )
( )
2
2
0.746 1000
26.8 ft 8.18 m
8 8 3475
wS
D
H
= = =
( )
( )
( )
2
2
8 26.8
8
1000 1001.9 ft 305.38 m
3 3 1000
D
L S
S
= + = + =
( ) 1
REF
T T AS REF
L L T T = +

( )
( )
6
122
1001.9 1 12.8 10 122 77
1002.48 ft (305.56 m)
L

= +

=
( )
( )
6
302
1001.9 1 12.8 10 302 77
1004.79 ft (306.26 m)
L

= +

=
( ) 3
8
S L S
D

=
( )
( ) ( )
( )
( )
( )( )
( )
122
302
3 1000 2.48
30.5 ft 9.3 m
8
3 1000 4.79
69.2 ft 21.1m
8
D
D
= =
= =
2 2
becomes
8 8
wS wS
D H
H D
= =
( )
( )
( )
2
2
122
0.746 1000
3057 lbs (13,647 N)
8 8 30.5
wS
H
D
= = =
( )
( )
( )
2
2
302
0.746 1000
1348 lbs (6,018 N)
8 8 69.2
wS
H
D
= = =
2-34
2.4.9 Results of High-Temperature Sag Tension
Calculations
Three different ACSR conductors were selected to dem-
onstrate the high-temperature behavior of conductors
and the calculation of the sag and tension. The ACSR
conductors have varying steel-to-aluminum ratios,
hence the effects of the construction of the conductor
on the sag and tension are demonstrated. The conduc-
tors selected for the examples are ACSR Drake,
ACSR Mallard, and ACSR Tern. All of these
ACSR conductors selected for the demonstration are
commonly used by electric utilities.
ACSR Mallard is constructed using 30 aluminum
strands and 19 steel core strands. The ratio of the alumi-
nums cross-sectional area to the conductors total cross-
sectional area for the ACSR Mallard conductor is
0.814, classifying this conductor as a wire with a high
steel-to-aluminum ratio. Similarly, ACSR Drake is con-
structed using 26 aluminum strands and 7 steel strands.
The ratio of the aluminums cross-sectional area to the
ACSR Drake conductors total cross-sectional area is
0.860, classifying this conductor as a wire with a
medium steel-to-aluminum ratio. Finally, ACSR Tern
is constructed using 45 aluminum strands and 7 steel
strands. The ratio of the aluminums cross sectional area
to the ACSR Tern conductors total cross-sectional area
is 0.935, classifying this conductor as a wire with a
low steel-to-aluminum ratio.
In support of the analysis and the illustration of the
high-temperature behavior of standard ACSR conduc-
tors, stress and strain data acquired in the EPRI Con-
ductor High Temperature Operation Project were used.
The data used for the prediction of high-temperature
sags included the stress and strain curves and coeffi-
cients of thermal elongation.
In order to illustrate the differences in the results, sag
analyses were performed for a level conductor span as
well as for inclined spans of 15 and 30. In each case,
sags were calculated for each type of conductor using
two different approaches at each of the selected conduc-
tor temperatures. The resulting sags are reported at the
initial and final condition at each temperature.
In addition, the span length was varied, and the sag
analyses were performed for 500 ft, 1000 ft, and 1500 ft
span lengths. Again, the results were reported at the ini-
tial and final condition at each temperature.
In the first approach, sags were calculated using Alcoa
Fujikuras Sag10 sag tension analysis program using the
provided stress and strain charts included in the soft-
ware. Values are reported for initial and final conditions,
respectively.
In the second approach, sags were calculated using
Alcoa Fujikuras Sag10 sag tension analysis program
and data obtained in the EPRI Conductor High Tem-
perature Operation Project. Values are reported for ini-
tial and final conditions, respectively. Tables 2.4-2
through 2.4-10 provide a summary and comparison of
derived sags at each operating temperature.
For ACSR Drake with the 500 ft (152 m) span, the
results of the comparison show that final sags calculated
using EPRI High Temperature Conductor data are less
than 1ft (0.3 m) greater than comparable values calcu-
lated using Alcoa Fujikuras Sag10 data. For the 15-
inclined span, the differences in the final sag were also
less than 1 ft (0.3 m). Similar results were obtained for
the 30 incline, which resulted in increased final sags
that were less than 1 ft (0.3 m).
using EPRI High Temperature Conductor data gener-
ally differ by about 2 ft (0.61 m) at ambient, but was
about 1 ft (0.3 m) higher for elevated temperatures when
compared to Alcoa Fujikuras Sag10 data. For the 15
inclined span, the difference in the final sags was less
than 1 ft (0.3 m). Similar results were obtained in the
case of the 30 incline, which resulted in increased final
sags that were minimal.
using EPRI High Temperature Conductor data was up
to 3 ft (0.9 m) at ambient, and 1 ft (0.3 m) at the elevated
temperatures when compared to Alcoa Fujikuras Sag10
data. For the 15 inclined span, the difference in the
final sags was less than a foot (0.3 m). Similar results
were obtained for the 30 incline, which resulted in
increased final sags that are less than 1 ft (0.3 m).
For ACSR Tern with the 500 ft (152 m) span, the results
of the comparison show that final sags calculated using
EPRI High Temperature Conductor data are generally
less by about half a foot (0.15 m) when compared to
Alcoa Fujikuras Sag10 data. For the 15 inclined span,
the difference in the final sags was negligible, and simi-
lar results were obtained for the 30 incline. Good corre-
lation of the data occurred for the 1000 ft (305 m) and
1500 ft (457 m) span, and the sag values were less than 1
ft (0.3 m) when compared to Alcoa data. In most cases,
the predicted sag was less than expected by the Alcoa
database curves.
2-35
For ACSR Mallard with the 500 ft (152 m) span, the
using EPRI High Temperature Conductor data are gen-
erally 1 ft (0.3 m) higher than comparable values calcu-
lated using Alcoa Fujikuras Sag10 data. For the 15
inclined span, the difference in the final sags was between
0 ft and 1.5 ft (0.46 m). For the 30 incline, the difference
in the final sags was between 0 and 1.6 ft (0.49 m).
using EPRI High Temperature Conductor data are 3 ft
(0.9 m) at ambient temperatures and about 2 ft (0.61 m)
at elevated temperatures, when compared to values cal-
culated using Alcoa Fujikuras Sag 10 data. For the 15
inclined span, the difference in the final sags was negligi-
ble at ambient temperature but 2.5 ft (0.76 m) at ele-
vated temperatures. For the 30 incline, the difference in
the final sags was negligible at ambient temperature but
up to 2.8 ft (0.85 m) at elevated temperatures.
using EPRI High Temperature Conductor data is an
additional 4 ft (1.2 m) at ambient temperature and is
about 2.5 ft (0.76 m) at elevated temperatures when
compared to values calculated using Alcoa Fujikuras
Sag10 data. For the 15 inclined span, the difference in
the final sags was negligible at ambient, and as high at
3 ft (0.9 m) and approached 4 ft (1.2 m) for elevated
temperature levels. Similar results were obtained for the
30 incline.
Table 2.4-2 ACSR Drake, Span Length = 500 ft
Alcoa Fujikura Sag 10 Software with Alcoa Data
Conductor
Temperature Level 15% 30%
Initial Sag (ft) Final Sag (ft) Initial Sag (ft) Final Sag (ft) Initial Sag (ft) Final Sag(ft)
23 Deg 3.26 4.35 3.26 3.37 3.26 3.76
120 Deg 8.61 8.61 8.61 8.30 8.61 9.26
150 Deg 9.21 9.61 9.21 9.52 9.21 10.62
Alcoa Fujikura Sag 10 Software with EPRI Data
Conductor
23 Deg 3.26 5.23 3.26 3.37 3.26 3.76
120 Deg 8.64 9.26 8.64 8.93 8.64 9.96
150 Deg 9.81 10.29 9.81 10.14 9.81 11.31
Comparison: Alcoa to EPRI
Conductor
Difference (ft) Difference (ft) Difference (ft) Difference (ft) Difference (ft) Difference (ft)
23 Deg 0 0.88 0.00 0 0 0.88
120 Deg 0.61 0.65 0.63 0.70 0.61 0.65
150 Deg 0.60 0.68 0.62 0.69 0.60 0.68
2-36
Conductor
23 Deg 13.05 16.27 13.05 13.05 13.05 15.05
120 Deg 22.48 25.32 22.48 23.22 22.48 25.89
150 Deg 25.62 26.94 25.62 26.45 25.62 29.49
Conductor
23 Deg 13.05 18.32 13.05 13.5 13.05 15.05
120 Deg 23.05 26.23 23.05 23.81 23.05 26.55
150 Deg 26.20 27.87 26.20 27.05 26.20 30.16
Conductor
23 Deg 0 2.05 0 0 0 0
120 Deg 0.57 0.91 0.57 0.59 0.57 0.66
150 Deg 0.58 0.93 0.58 0.60 0.58 0.67
Conductor
23 Deg 29.39 34.65 29.39 30.38 29.39 33.88
120 Deg 42.39 47.81 42.39 43.74 42.39 48.77
150 Deg 46.40 49.93 46.40 47.84 46.40 53.34
Conductor
23 Deg 29.39 37.61 29.39 30.38 29.39 33.88
120 Deg 43.04 48.83 43.04 44.41 43.04 49.51
150 Deg 47.10 50.97 47.10 48.56 47.10 54.14
Conductor
23 Deg 0 2.96 0 0 0 0
120 Deg 0.65 1.02 0.65 0.67 0.65 0.74
150 Deg 0.70 1.04 0.70 0.72 0.70 0.80
2-37
Table 2.4-5 ACSR Tern, Span Length = 500 ft
Conductor
23 Deg 3.81 5.67 3.81 3.94 3.81 4.39
120 Deg 10.90 12.89 10.90 11.27 10.90 12.56
150 Deg 13.11 13.73 13.11 13.53 13.11 15.09
Conductor
23 Deg 3.81 5.74 3.81 3.93 3.81 4.39
120 Deg 10.76 12.45 10.76 11.12 10.76 12.40
150 Deg 13.00 13.29 13.00 13.42 13.00 14.96
Conductor
23 Deg 0 0.07 0 -0.01 0 0
120 Deg -0.14 -0.44 -0.14 -0.15 -0.14 -0.16
150 Deg -0.11 -0.44 -0.11 -0.11 -0.11 -0.13
Conductor
23 Deg 15.24 19.84 15.24 15.76 15.24 17.58
120 Deg 27.30 34.46 27.30 28.17 27.30 31.41
150 Deg 30.84 35.78 30.84 31.80 30.84 35.45
Conductor
23 Deg 15.24 19.98 15.24 15.76 15.24 17.58
120 Deg 27.17 33.7 27.17 28.04 27.17 31.27
150 Deg 30.67 35.04 30.67 31.63 30.67 35.26
Conductor
23 Deg 0 0.14 0 0 0 0
120 Deg -0.13 -0.76 -0.13 -0.13 -0.13 -0.14
150 Deg -0.17 -0.74 -0.17 -0.17 -0.17 -0.19
2-38
Conductor
23 Deg 34.34 41.10 34.34 35.47 34.34 39.55
120 Deg 49.79 62.49 49.79 51.31 49.79 57.20
150 Deg 54.26 64.30 54.26 55.88 54.26 62.28
Conductor
23 Deg 34.34 41.27 34.34 35.47 34.34 39.55
120 Deg 49.69 61.51 49.69 51.21 49.69 57.08
150 Deg 54.13 63.26 54.13 55.73 54.13 62.12
Conductor
23 Deg 0 0.17 0 0 0 0
120 Deg -0.10 -0.98 -0.10 -0.10 -0.10 -0.12
150 Deg -0.13 -1.04 -0.13 -0.15 -0.13 -0.16
Table 2.4-8 ACSR Mallard, Span Length = 500 ft
Conductor
23 Deg 3.02 4.27 3.02 3.12 3.02 3.48
120 Deg 6.21 7.43 6.21 6.42 6.21 7.16
150 Deg 7.77 8.48 7.77 8.03 7.77 8.95
Conductor
23 Deg 3.02 5.45 3.02 3.12 3.02 3.48
120 Deg 7.63 8.49 7.63 7.89 7.63 8.80
150 Deg 8.67 9.57 8.67 8.96 8.67 10.00
Conductor
23 Deg 0 1.18 0 0 0 0
120 Deg 1.42 1.06 1.42 1.47 1.42 1.64
150 Deg 0.90 1.09 0.90 0.93 0.90 1.05
2-39
Table 2.4-9 ACSR Mallard, Span Length =1000 ft
Conductor
23 Deg 12.08 15.82 12.08 12.50 12.08 13.94
120 Deg 19.54 22.59 19.54 20.20 19.54 22.52
150 Deg 22.11 24.33 22.11 22.85 22.11 25.48
Conductor
23 Deg 12.08 18.58 12.08 12.50 12.08 13.94
120 Deg 21.99 24.46 21.99 22.72 21.99 25.34
150 Deg 24.51 26.20 24.51 25.31 24.51 28.22
Conductor
23 Deg 0 2.76 0 0 0 0
120 Deg 2.45 1.87 2.45 2.52 2.45 2.82
150 Deg 2.40 1.87 2.40 2.46 2.40 2.74
Table 2.4-10 ACSR Mallard, Span Length = 1500 ft
Conductor
23 Deg 27.21 33.36 27.21 28.13 27.21 31.37
120 Deg 38.20 43.22 38.20 39.45 38.20 43.98
150 Deg 41.62 45.51 41.62 42.95 41.62 47.89
Conductor
23 Deg 27.21 37.38 27.21 28.13 27.21 31.37
120 Deg 40.74 45.76 40.74 42.05 40.74 46.89
150 Deg 45.07 48.03 45.07 46.48 45.07 51.82
Conductor
23 Deg 0 4.02 0 0 0 0
120 Deg 2.54 2.54 2.54 2.60 2.54 2.91
150 Deg 3.45 2.52 3.45 3.53 3.45 3.93
2-40
2.4.10 Effects of Wind Speed on Thermal Ratings
High-temperature operations are greatly affected by
local wind conditions. This is discussed in Section 2.3
for its effect on clearances. Below, its effect on conduc-
tor strength is the focus.
The following will demonstrate the effect of wind on the
performance of the conductor. For the comparison,
Power Technologies Ratekit was used. Ratekit is a soft-
ware tool capable of evaluating the effects of wind speed
and operating temperature independently. In the discus-
sion, all conductor conditions are assumed steady state.
Also, the angle of the wind relative to the conductor is
90, the span is at sea level, and the time of year is sum-
mer. The coefficient of emissivity is 0.50, the coefficient
of solar absorptivity is 0.50, the conductor resistance at
25C is 0.1200 ohms/mile (0.075 ohms/km), and the con-
ductor resistance at 75C is 0.1442 ohms/mile (0.090
ohms/m). Atmospheric conditions are clear, and the ori-
entation of the conductor relative to the north is 0. The
ambient temperature is 25C , the ruling span is 1000 ft
(305 m), and the conductors used in the comparisons
are ACSR Drake and AAC Arbutus. The operating tem-
peratures are 120C and 150C , and the wind speeds
used are 0 ft/sec, 2 ft/sec (0.61 m/sec), and 4 ft/sec
(1.2 m/sec). Table 2.4-11 shows the remaining strength
and sags for the ACSR Drake conductor, and Table 2.4-
12 summarizes the results obtained for the AAC Arbu-
tus.
At 1180 A and 0 ft/sec wind, the operating temperature
of the ACSR Drake conductor is 150C. If subjected to
this high operating temperature for 100 hours, the
remaining strength of the ACSR Drake conductor
decreases by 1.4% to 98.64% of the rated tensile
strength. If the weather conditions are favorable and the
wind speed is 2 ft/sec (0.61 m/sec), the operating temper-
ature of the ACSR Drake conductor is 110C, and the
tensile strength of the conductor remains unaffected. If
the wind speed is 4 ft/sec (1.22 m/sec), the temperature
of the ACSR Drake conductor with 1180 A of current is
nearly 86C, and the strength of the conductor remains
unaffected. Similarly, the sag (final) of the ACSR con-
ductor at 1180 A, and 0 ft/sec wind is 39.6 ft (12.1 m),
while the sag of the ACSR Drake decreases by nearly
3.5 ft (1.1 m) if the wind speed increases from 0 ft/sec to
2 ft/sec. If the wind speed increases to 4 ft/sec (1.2
m/sec), the sag of the ACSR Drake conductor is 33.8
ft/sec (10.3 m/sec), nearly 6 ft (1.8 m) less than in the
case where the wind is 0 ft/sec. If the ACSR Drake car-
ries 1000 A and the wind speed is 0 ft/sec, the operating
temperature of the conductor is 120C, resulting in no
loss of tensile strength if operated for 100 hours. If the
wind speed increases to 2 ft/sec (0.61 m/sec), the operat-
ing temperature decreases 86C and the sag decreases by
approximately 2 ft (0.61 m). If the wind speed increases
to 4 ft (1.2 m), the temperature decreases to 70C and
the sag decreases an additional 2 ft (0.61 m).
The temperature of an AAC Arbutus conductor carry-
ing a current of 1130 A at a wind speed of 0 ft/sec is
150C. If the duration of this operation is 100 hours, the
tensile strength of the conductor decreases by nearly
30% to 70.12% of the rated tensile strength. However, if
the wind speed is 2 ft/sec (0.61 m/sec), the operating
temperature of the AAC Arbutus conductor is 108C,
and the operation at this condition for a period of 100
hours reduces the tensile strength of the conductor by
only 9% instead of 30% at a wind speed of 0 ft/sec. The
remaining strength at these conditions is 92.1% of the
rated tensile strength, an acceptable value. At a wind
speed of 4 ft/sec (1.2 m/sec), an AAC Arbutus conduc-
tor carrying a current of 1130 A will operate at a tem-
perature of 85C, resulting in no loss of the tensile
strength of the conductor. Also, a wind speed of 2 ft/sec
(0.61 m/sec) reduces the sag by nearly 4 ft (1.2 m) com-
pared to the sag calculated in the no wind scenario, and
the sag is reduced an additional 2 ft (0.61 m) if the wind
speed increases to 4 ft/sec (1.2 m/sec).
Table 2.4-11 ACSR Drake
Current
(A)
Wind
Speed
(ft/sec)
Tempera-
ture (C)
Remaining
Strength
(%)
a
a. The remaining strength is after 100-hours of high operat-
ing temperatures.
Initial
Sag (ft)
Final
Sag (ft)
1180 0 150 98.64 36.1 39.6
2 110 100 32.3 36.1
4 85.7 100 29.8 33.8
1000 0 120 100 33.3 36.9
2 85.8 100 29.8 33.8
4 68.5 100 28.0 32.1
Table 2.4-12 AAC Arbutus
Current
(A)
Wind
Speed
(ft/sec)
Tempera-
ture (C)
Remaining
Strength
(%)
a
a. The remaining strength is calculated when the conductor
is exposed to the high temperature for 100 hours.
Initial
Sag (ft)
Final
Sag (ft)
1130 0 150 70.12 47.4 49.0
2 108.5 92.11 43.5 45.2
4 84.6 100 41.2 43.0
960 0 120 85.89 44.6 46.3
2 84.9 100 41.2 43.0
4 68.1 100 39.5 41.3
2-41
AAC Arbutus carrying a current of 960 A at a wind
speed of 0 ft/sec operates at a temperature of 120C,
resulting in a loss of 14% of the tensile strength if oper-
ated continuously for a period of 100 hours. The
remaining tensile strength is reduced to 86% of the rated
tensile strength of the conductor. If the wind speed
increases to 2 ft/sec (0.61 m/sec), the sag of the conduc-
tor decreases by 3 ft (0.9 m), while an increase in the
wind speed will decrease the sag by nearly 5 ft (1.5 m)
when compared to the no wind case.
2.4.11 Thermal Elongation
Thermal elongation is a metallurgical phenomenon
where the material increases in length in proportion to
an increase in temperature. The rate of linear thermal
expansion for the composite ACSR conductor is less
than that of conductors that are made entirely of alumi-
num because the steel strands in the ACSR elongate at
approximately half the rate of aluminum. The theoretical
composite coefficient of linear thermal expansion (CTE)
of a nonhomogenous conductor, such as the ACSR
Drake may be found from the following equations:
2.4-21
2.4-22
where
E
AL
= modulus of elasticity of aluminum, psi.
E
ST
= modulus of elasticity of steel, psi.
E
AS
= modulus of elasticity of aluminum-steel
composite, psi.
A
AL
= area of aluminum strands, square units.
A
ST
= area of steel strands, square units.
A
TOTAL
= total cross sectional area, square units.
AL
= aluminum coefficient of linear thermal
expansion, per F.
ST
= steel coefficient of thermal elongation, per
F.
AS
= composite aluminum-steel coefficient of
thermal elongation, per F.
Example 2.4-10: Using elastic moduli of 10 and 30 mil-
lion psi (68,966 MPA and 206,897 Mpa) for aluminum
and steel, find the elastic modulus for ACSR Drake is:
and the coefficient of linear thermal expansion is:
A comparison of three different ACSR conductors of
varying aluminum-to-steel ratios is used to demonstrate
the effect of the steel reinforcing on the composite con-
ductor behavior and coefficient of linear expansion.
Table 2.4-13 summarizes calculated (ideal) and mea-
sured coefficients of linear thermal expansion. The mea-
sured coefficient of linear expansion of the ACSR Tern
conductor (low steel-to-aluminum ratio) is 10.9% larger
than the calculated value. The measured coefficient of
l i near expansi on of the ACSR Drake conductor
(medium steel-to-aluminum ratio) is approximately 6%
higher than the calculated value and the measured coef-
ficient of the ACSR Mallard (high steel-to-aluminum
ratio) is only marginally larger than the calculated value.
As the steel-to-aluminum ratio increases, the coefficient
of l i near expansi on of the composi te conductor
approaches a limiting value equal to the coefficient of
steel alone.
Note that the results of the EPRI High Temperature
Conductor Project identifies the conductors knee point,
and that the coefficient of linear expansion does not
typically address the effect of the knee point on the sag
and tension calculation. A typical set of test data is
shown in Figure 2.4-8. The knee point is shown at
approximately 120C. However, to accurately predict
sags and tensions at high temperature, the effects of the
knee point need to be considered and is discussed in
more detail.
The conditions under which calculations using a simple
composite modulus and coefficient of linear thermal
expansion fail may be illustrated by considering the ten-
sion distribution between steel and aluminum under
normal and high temperature conditions. The preceding
equations for composite modulus and CTE are derived
based on the assumption that the aluminum and steel
ST ST AL AL
AS AL ST
AS TOTAL AS TOTAL
E A E A
E A E A

= +

ST AL
AS AL
TOTAL TOTAL
A A
E E EST
A A

= +

( ) ( )
6 6
6 10
0.6247 0.1017
10 10 30 10
0.7264 0.7264
12.8 10 (8.83 x 10 Pa)
AS
E x x
x psi

= +

=
6
6
6
6
6
6
6
10 10 0.6247
12.8 10
0.7264 12.8 10
30 10 0.1017
6.4 10
0.7264 12.8 10
10.7 10 /
AS
x
a x
x
x
x
x
x F

=

+

=
Table 2.4-13 Coefficient of Thermal Elongation for ACSR
Conductors
ACSR
Conductor
(below 212F,
or 100C)
Calculated
Coefficient of
Thermal Elon-
gation (CTE)
Measured
Coefficient of
Thermal Elon-
gation (CTE)
Percent
Difference
Tern 45/7 11.7 x 10
-6
13.0 x 10
-6
10.9
Drake 26/7 10.7 x 10
-6
11.4 x 10
-6
6.09
Mallard 30/19 10.2 x 10
-6
10.1 x 10
-6
0.64
2-42
strands are of the same length. Based on this assump-
tion, the total conductor tension (H
AS
) is equal to the
sum of the component tensions, and the elongation of
the steel and aluminum is equal:
2.4-23
2.4-24
For example, an ACSR Drake conductor installed in a
600 ft (183 m) span at an initial tension of 9450 lbs
(42,188 N) carries 3124 lbs (13,946 N) of the tension in
the steel core with the remainder of the tension being
carried by the aluminum strands. Therefore, the tension
carried by the steel is roughly 33% of the total tension.
If the ACSR Drake conductor is heated to a tempera-
ture of 100
o
C, the sag of the conductor increases, and
the total tension decreases to 4780 lbs (21,339 N), but
the tension carried by the steel core increases from
3124 lbs (13,946 N) to 3305 lbs (14,754 N), or nearly
70% of the total tension.
As the operating temperature of the conductor increases
further, the tensi on i n the al umi num eventual ly
decreases to zero while the tension carried by the steel
increases further. At some temperature, commonly
referred to as the knee point temperature, the linear
expansion of the conductor continues, but the stresses in
the aluminum strands change from tension to compres-
sion (as shown in Figures 2.4-8 and 2.4-9). The knee
point temperature of an ACSR conductor is directly
proportional to the steel-to-aluminum ratio and is lower
for ACSR conductors with high ratios and higher for
conductors with low steel to aluminum ratios.
Conductors made by different manufacturers exhibit
different knee-point temperatures because of differences
in the manufacturing practices, processes, and machin-
ery. For example, Figure 2.4-8 indicates that the ACSR
Drake has a knee point at about 120C, while the data in
Figure 2.4-8 suggests that the ACSR Drake has a knee
point of only 70C. Disregarding experimental differ-
ences, built-in stresses, which are the result of manufac-
turing methods and tools, are the primary cause for this
difference.
In Figure 2.4-8, the coefficient of thermal expansion is
17.38 x 10
-6
/C prior to the knee point, and the corre-
sponding value in Figure 2.4-9 is 17.53 x 10
-6
/C, a dif-
ference of less than 1%. While the corresponding
coefficient beyond the knee point is also fairly consis-
tent, the temperature at which the knee point has been
observed differs by nearly 50C or nearly 70%. The coef-
ficient of thermal expansion beyond the knee point in
Figure 2.4-8 is 12.84 x 10
-6
/C, while the corresponding
value in Figure 2.4-9 is 13.74 x 10
-6
/C, a difference of
less than 7%.
2.4.12 Creep Elongation
Once a conductor has been installed to an initial ten-
sion, it can elongate further. Such elongation results
from three phenomenapermanent elongation due to
high-tension levels resulting from ice and wind loads,
creep elongation under everyday tension levels, and
creep elongation due to thermal overloads. The creep
resulting from thermal loading is discussed below.
High-Temperature Creep Elongation
The definition of normal creep is the accumulative
nonelastic elongation of a conductor under low tension,
over an extended period of time at modest operating
temperatures. Normal operating temperatures are gen-
erally agreed upon as temperatures below 75C. In any
AS AL ST
H H H = +
ST AL
ST ST AL AL
H H
A E A E
=

Figure 2.4-8 ACSR Drake (26/7) composite thermal
elongation, (30% of RBS Pre-Stressing, 10% of RBS
Applied Tension, Manufacturer 2).
Figure 2.4-9 ACSR Drake (26/7) composite thermal
elongation, (30% of RTS Pre-Stressing, 10% of RTS
Applied Tension, Manufacturer 1).
2-43
case, a conductor under tension experiences a nonelastic
elongation over a period of time. Time is usually mea-
sured in years. The magnitude and rate of creep are a
function of the conductor's composition, stranding, line
tension, and operating temperature.
Conductors exhibit creep under everyday tension levels
even if the tension level never exceeds normal operating
conditions. Creep can be determined by long-term labo-
ratory creep tests, and the results of the tests are used to
generate charts that document the relationship of creep
versus time. On the stress-strain graphs of conductors,
creep curves are often shown for 6-month, 1-year, and
10-year time periods. Figure 2.4-10 shows a typical
creep curve (experimental results) for an ACSR Drake
conductor at room temperature. Generally, the creep
elongation in aluminum conductors is quite predictable,
it is a function of time and temperature, and the rela-
tionship is exponential. Thus, knowing the initial condi-
tions of the conductor, the permanent elongation due to
creep at everyday tension can be found for any period of
time after the initial installation. Note that the creep
elongation of copper and steel strands is minimal, and
that the effects can essentially be ignored.
High-temperature creep occurs when a conductor is
operated for extended periods of time at operating tem-
peratures in excess of approximately 75C. Creep consti-
tutes an irreversible, nonelastic elongation occurring in
response to the molecular realignment in the conduc-
tors base material. Figure 2.4-10 shows the effects of
creep on the conductor strain at an operating tempera-
ture of 120C. Because aluminum exhibits a signifi-
cantly higher rate of creep elongation than steel, the sag
and tension behavior of all-aluminum type conductors
such as AAC, AAAC, and ACAR is much more suscep-
tible to high-temperature creep. Conversely, copper and
steel wire strands supported aluminum conductors (Cu,
ACSR, and AACSR) exhibit lower rates of creep, and
the sag and tension behavior of these conductors is less
affected by the high-temperature operation. Aluminum
Conductor Steel Supported (ACSS) conductor, a con-
ductor constructed using fully annealed aluminum
strands and steel strands for the strength member,
exhibits negligible creep rates since all of the tension is
carried by the steel strength member, which is essentially
not affected by creep.
Creep, and the associated rate of creep, are directly pro-
portional to the type of conductor, the steel-to-alumi-
num stranding ratio, the material characteristics of the
conductor, and the tension and operating temperature.
For example, using strands that are drawn from continu-
ous-cast rod instead of hot-rolled rod reduces the creep
elongation and the associated long-term sag and tension
behavior of the conductor. Other important parameters
required in the determination of the rate of creep and
the cumulative effect include the conductor stress or
strain, the operating temperature, the elapsed time, and
for ACAR and ACSR type conductors, the ratio of the
constitutive components.
Figures 2.4-10 and 2.4-11 show experimentally derived
creep elongation results for ACSR Drake at room tem-
perature and at 120C. In both experiments, the major-
ity of the creep elongation occurred during the initial
100 hours, a settling phase, so to speak. Note that at
ambient room temperature conditions (23C), the final
creep elongation after 1000-hours was 0.01%, while the
final creep elongation measured at the increased operat-
ing temperature was 0.03%. Thus, an ACSR conductor
operated at a high temperature will reach its final sag
Figure 2.4-10 ACSR Drake composite creep at 23C,
Manufacturer 1.
Figure 2.4-11 ACSR Drake composite creep at 120C,
Manufacturer 1.
2-44
approximately two to three times faster than a conduc-
tor operated at temperatures of 75C or below.
Creep Predictor Equations
Harvey (Harvey 1969) developed mathematical models
and associated equations that can be used to estimate
the magnitude of creep elongation. The models and
equations were empirically derived, and are able to pre-
dict the creep elongation of a conductor at room tem-
perature, as well as at an elevated temperature. The
required definitions, constants, and formulas are listed
(see Tables 4-14 and 4-15), along with an example.
Definition of Terms:
c
= Primary creep strain (units/unit)
= Strain - increase in length/original
(units/unit)
T
= Increase in conductor strain due to ele-
vated temperature operations (units/unit)
= Stress - tension/area (N/mm
2
, lbs/in
2
)
= Coefficient of thermal expansion
(units/unit/C)
t = Elapsed time (hours)
T = Conductor temperature (C)
T = Temperature change value (C)
A
EC
= Area of aluminum strands (sq. mm., sq.
in.)
A
ST
= Area of steel strands (sq. mm., sq. in.)
A
T
= Total conductor area (sq. mm., sq. in.)
%RS = Tension as a percentage of the rated
strength (%)
Predictor Equations:
All-Aluminum Conductors (Room Temperature, Metric)
2.4-25
2.4-26
2.4-27
All-Aluminum Conductors (Room Temperature,
English)
2.4-28
2.4-29
2.4-30
All-Aluminum Conductors (Elevated Temperature,
Metric)
2.4-31
2.4-32
2.4-33
All-Aluminum Conductors (Elevated Temperature,
English)
2.4-34
2.4-35
2.4-36
Steel Reinforced Conductors (Room Temperature,
English)
Aluminum strands drawn from hot-rolled rod:
2.4-37
Aluminum strands drawn from continuous cast rod:
2.4-38
Steel Reinforced Conductors (Elevated Temperature,
English)
Only for conductors with less than 7.5% steel by area:
Note: K
1
, K
3
, M
1
, and M
3
are for wire bar rolled rod and K
2
,
K
4
, M
2
, and M
4
are for continuous cast (rolled) rod.
Note: K
1
, K
3
, M
1
, and M
3
are for wire bar rolled rod and K
2
,
K
4
, M
2
, and M
4
are for continuous cast (rolled) rod.
Table 2.4-14 Formula Constants (Metric Units)
7 Strands 19 Strands 37 Strands 61 Strands
K
1 1.36 1.29 1.23 1.16
K
2 0.84 0.77 0.77 0.71
M
1 0.0148 0.0142 0.0136 0.0129
M
2 0.0090 0.0090 0.0084 0.0077
G 0.71 0.65 0.77 0.61
Table 2.4-15 Formula Constants (English Units)
7 Strands 19 Strands 37 Strands 61 Strands
K
3 0.0021 0.0020 0.0019 0.0018
K
4 0.0013 0.0012 0.0012 0.0011
M
3 0.000023 0.000022 0.000021 0.000020
M
4 0.000014 0.000014 0.000013 0.000012
G 0.0011 0.0010 0.0012 0.00094
1.3 0.16
c
AAC: = K t
1.3 0.16
c
AAAC: = G t
1.4 1.3 0.16
c EC T
ACAR: = (0.19 + 1.36 A / A ) (T t )
1.3 0.16
c
AAC: = K t
1.3 0.16
c
AAAC: = G t
1.4 1.3 0.16
c EC T
ACAR: = (0.0003 + 0.0021 A / A ) (T t )
1.4 1.3 0.16
c
AAC: = M T t
1.4 1.3 0.16
c
AAAC: = 0.0077 T t
1.4 1.3 0.16
c EC T
ACAR: = (0.0019 +0.012 A / A ) (T t )
1.4 1.3 0.16
c
AAC: = M T t
1.4 1.3 0.16
AAAC: c = 0.000012 T t
c
1.4 1.3 0.16
EC T
ACAR: =
(0.000003 +0.000019 A / A ) (T t )
1.3 0.16
c
ACSR: = 2.4 (%RS) t
1.3 0.16
c
ACSR: = 1.1 (%RS) t
2-45
2.4-39
Elevated creep in conductors with steel to aluminum
ratios of greater than 7.5% can be ignored.
Temperature Change Value
The equivalent temperature change value is a calculated
temperature that approximates the net increase in the
micro-strain due to elevated temperature creep above
and beyond general creep.
2.4-40
2.4-41
@ambient

is the strain in the conductor due to room tem-
perature creep only, and
@high
is the strain in the con-
ductor attributed to the elevated temperature creep.
Method for Using the Predictor Equations
The process of calculating the predicted elevated tem-
perature creep is straightforward and can be accom-
plished using sags and tensions derived from graphical
charts, long-hand calculations, or computer software
predictions. The following procedure must be followed:
1. Use the standard graphical charting, long-hand cal-
culations, or computer sag and tension methods to
predict the sags and tensions without elevated creep
for the given situation.
2. Compute the creep of the conductor at ambient tem-
perature.
3. Compute the creep of the conductor at the first ele-
vated temperature.
4. Compute the number of hours that would be
required to accumulate an equivalent amount of con-
ductor creep at the second elevated temperature.
5. Repeat steps 3 and 4 for each elevated temperature.
6. Calculate the value of the equivalent temperature
change by subtracting the predicted creep elongation
at the everyday temperature from the creep elonga-
tion at the elevated temperature.
7. Calculate the final sag of the conductor following ele-
vated temperature creep by adding this temperature
change value to the temperatures used in the stan-
dard sag and tension calculation.
Example 2.4-11: A 795 kcmil (405 mm
2
)ACSR Tern
conductor is subjected to 20 hours at 60F (16C), 7500
hours at 167F (75C ) and 1240 hours at 212F
(100C). What is the sag and tension if creep is a factor?
Use a ruling span equal to 1000 ft (305 m). Use Alcoas
Sag10 program to directly calculate the result and com-
pare it to the results of the Creep Predictor Equations.
Assume Alcoa Heavy Loading applies.
Using Equation 2.4-37, the ambient temperature creep is:
20 hours at 60F produces a creep of 0.64 micro-in./in.
or the equivalent of operating the conductor for 0 hours
at 212F. Using Equation 2.4-39, the elevated tempera-
ture creep is:
7500 hours at 167F produces a creep of 417.7 micro-
in./in. or the equivalent of operating the conductor for
1690 hours at 212F (100C).
To calculate the total elongation due to creep, the sum
of the equivalent times (hours) at 212F (100C ) is
determined. In this example, the sum is:
The temperature change value is calculated using Equa-
tion 2.4-40.
where
The final sag following elevated temperature creep is
determined by adding this temperature change value to
0.16
c
ACSR: = 2.4 (%RS) T t
T T
= T or T = /
T @high @ambient
= -
1.3 0.16
c
0.16
c
c
= 2.4 (%RS) t where %RS is the
Remaining Strength, and is the time (hours)
a) Conductor creep produced for 20 hours
at 60F (16C)
= 2.4 x 0.25 1.3 x 20
= 0.64 micro-in/in (1 in. = 2.
t
54 cm)
0.16
c
0.16
c
c
= 2.4 (%RS) T t
b) Conductor creep produced for 7500 hours
at 167F (75 C)
= 2.4 x 0.25 x 167 x (7500)
= 417.7 micro-in./in.
0.16
c
c
c) Conductor creep produced for 1240 hours
at 212F (100 C)
= 2.4 x 0.25 x 212 x (1240)
= 397.6 micro-in./in.
0.16
c
0.16
c
c
0 + 1690 + 1240 = 2930 hours at 212F (100C).
= 2.4 (%RS) T t
= 2.4 x.25 x 212 x 2930
= 456 micro-in./in.
T
T = /
T @high @ambient
6
= -
T (456 0.64) /(11.8 10 )
38.75 F(3.8 C)

=
=
2-46
the temperatures used in the standard sag and tension
software calculations. In this example, 60F (16C)
becomes 98. 75F (37C), 90F (32C) becomes
128.75F (54C), 120F (49C) becomes 158.75F
(70C), etc. Table 2.4-16 provides a summary compari-
son of the calculated sags including the effects of creep
at each of the reference temperatures. The differences in
the sags predicted by each method can be mostly attrib-
uted to the empirical nature of the predictor equations
and the creep data and numerical analysis methods
deployed by the Alcoa software. Regardless, the use of
the predictor equations has been illustrated to convey
the process required whenever dealing with the effects of
elevated temperature creep for conductors. While the
predicted sags vary depending on which analysis
method is used, the critical issue is the significant differ-
ence in the predicted sag that is directly attributed to the
high-temperature operation of the conductor.
As discussed, the results obtained by using the hand cal-
culations are in close agreement to the values obtained
using the software, and the percentage difference ranged
from only 2. 2 to 13%. Note that the correl ati on
improves as the temperature is increased, with a differ-
ence of only 2.2% when the operating temperature is
212F (100C).
2.4.13 Connectors at High Temperature
Splices and compression fittings, which join sections of
conductor and other components, must provide both
structural and electrical integrity. Therefore, the integ-
rity of these components is critical to the operation of a
power line, particularly as the temperature of operation
increases. Connector failures can become very costly to
utilities, and the cost of failure can be attributed to one
of four categories. First, there is the cost associated with
potential damages and injuries to the public. Second,
there is a cost for loss of revenue from rerouting power.
Third is the cost of repairs, which depends heavily on
the labor cost for using line crews and equipment, and
fourth, there is the cost associated with the loss of com-
petitive advantage in the marketplace.
Transmission line operators use two types of tension
connectors: limited-tension connectors and full-tension
connectors. Limited-tension connectors are primarily
designed to join conductors that are under little or no
mechanical tension. These connectors are typically used
to splice the ends of two conductors together in a low-
tension application, tap a second conductor from a con-
tinuous run conductor, or terminate the end of a con-
ductor in a low-tension application. Typical examples of
limited-tension connectors are bolted connectors, com-
pression connectors, formed-wire connectors, wedge-
type connectors, and implosive connectors. Because
these connectors exhibit a limited tensile strength, the
part of the connector directly in contact with the cur-
rent-carrying conductor is generally smaller in area than
that of its full-tension counterpart.
Full -tensi on connectors are desi gned to provide
mechanical strength up to, or in excess, of the rated ten-
sile strength of the conductor at all operating tempera-
tures and also to provide a low-resistance conductive
path for the current carried by the conductor. Thus,
splice connectors are used to join the ends of the con-
ductors in spans between supports, and dead-end (also
referred to as terminations) connectors are used to join
conductors to the attachment hardware at the end of
each tension segment, or special heavy angle, or termi-
nation structures.
Generally, full-tension connections can be achieved
using one- and two-piece compression connectors,
formed-wire splices, implosive connectors, and wedge-
type connectors, but the compression-type connector
appears to be the predominant method of creating a
full-tension connection on high-voltage conductors.
Although the term full tension is commonly used to
specify the tensile strength of connectors to match the
tensile strength of the conductor, the connectors are
typically designed to exhibit a tensile strength of at least
95% of the conductor's rated breaking strength.
The main consideration for connectors when evaluating
elevated conductor temperature operations is its impact
on the connectors short- and long-term performance
and the effect on the service life. Certainly, the high-
temperature operation of a circuit greatly increases a
connectors electrical, mechanical, and thermal stresses.
If operated unchecked, these stress increases are likely
to lead to the premature deterioration of the integrity of
the connector and the eventual failure of this critical
component. Since the prediction of such a failure is
Table 2.4-16 Comparison of the Predictor Equations and
Alcoas Software Results
Tempera-
ture
(F)
Predictor
Equations
(ft)
Alcoas Sag 10
Software (ft)
Percentage
Difference (%)
60 27.5 31.6 13.0
90 29.5 33.4 11.8
120 31.4 35.2 10.8
167 34.2 36.6 6.3
212 36.8 37.6 2.2
2-47
difficult, and the consequences are great, the successful
high-temperature operation of a circuit requires a care-
ful evaluation of all of the issues involved.
Connector Breakdown Process
A connector facilitates the transfer of current through
numerous contact points between the connector and the
conductor. High current densities and high operating
temperatures tend to encourage the buildup of resistive
compounds at these contact points, effectively reducing
the size of the contact area, or in some cases completely
restricting the current flow. Generally, the resulting
increase in the temperature of the connector will then
allow the formation of new contact points within the
component at locations where the buildup of resistive
compounds is small. The continuous cycle of closing and
reopening of contact points within the connector can be
thought of as an aging process in which the connector
will continue to perform well as long as there are loca-
tions where contact points can be easily established.
Once the connector has aged sufficiently so that all loca-
ti ons for easi ly establ i shi ng contact poi nts are
exhausted, the connector is forced to maintain contact
points at locations at which resistive compounds are
present in order to reach the parent metal . This
increases the overall resistance of the connector, its
operating temperature, and current density within the
remaining contact points. At this time, the higher cur-
rent densities and operating temperatures cause further
buildup of resistive compounds, which in turn further
increases the current density and operating tempera-
tures, leading to the eventual thermal failure of the con-
nector. The thermal failure will eventually lead to the
electrical failure of the connector, and eventually results
in the mechanical failure of the component. Figure
2.4-12 shows typical compression deadend connectors
as used on transmission lines, and Figure 2.4-13 shows
an infrared image of the connector indicating an imped-
ing failure as indicated by the significant temperature
differences at the surface of the component.
Elevated temperature operations of conductors will
increase the current density and operating temperature
of associated connectors. This increase in the associated
thermal, electrical, and mechanical stresses will acceler-
ate their aging process, effectively reducing service life.
The amount of accelerated aging connectors experience
is directly related to the magnitude and frequency of the
elevated current and operating temperature operation
experienced by the component. Unfortunately, the rela-
tionship between the aging of the connector and the
loads served is nonlinear, and little success has been
achieved in directly quantifying that relationship. EPRI
is currently conducting research to assess the perfor-
mance of transmission-line components when subjected
to increases in operating temperatures. Preliminary
results indicate that most conductor hardware appears
to operate more efficiently when compared to the con-
ductors as long as the components are installed cor-
rectly. Results show that failures occur mostly as a result
of improper installation, and that many of these issues
can be addressed effectively using an appropriate qual-
ity assurance strategy.
Most wel l -desi gned connectors (when properly
installed) are capable of operating at high current densi-
ties and high conductor temperatures while providing
acceptable performance and service life. The current
cycle test, an industry standard, is commonly used to
evaluate these connector designs. Current cycling the
connector results in the thermal expansion and contrac-
tion of the electrical contact interface and tends to
break down the contact points between the connector
and the conductor. Although this standard test identi-
fies procedures and qualification criteria for connectors
used under normal operating conditions, the application
of this test does have its limitations. Figure 2.4-14 shows
typical test results obtained from a standard heat
cycling test.
Figure 2.4-12 Visual image of problem deadend. Figure 2.4-13 IR image of problem deadend.
2-48
Connectors should be considered as approaching failure
if the operating temperature approaches or exceeds the
operating temperature of the conductor. Experiments
show that the operating temperature of an adequately
designed and correctly installed compression connector
should be approximately 10 to 30% lower than the oper-
ating temperature of the conductor. The reduced oper-
ating temperature can be directly attributed to the
significantly larger cross section of the compression
connector relative to the cross-sectional area of the con-
ductor, the increase in the surface area of the compres-
sion fitting relative to the surface area of a comparable
unit length of conductor, and the increase in the convec-
tion to occur at any measurable wind speed.
High-Temperature Effects on Connector Joint Compound
Most aluminum connectors, especially compression
type, employ a viscous compound in the interface
between the connector and underlying conductor. The
primary purpose of the joint compound is to provide a
barrier that prevents moisture and other contaminants
from leaching into the joint. The ingress of moisture
and other contaminants is not desirable since this will
lead to the internal corrosion of the connector, which is
likely to affect not only the tensile strength of the steel
strength member in the case of an ACSR conductor but
also deteriorate the connectors ability to effectively
transfer the current.
The repeated high-temperature operation (connector
temperatures above 200F (95C)) degrades the joint
interface by causing the viscosity of the joint compound
to decrease significantly, leading to the eventual outflow
of the joint compound or the carbonization of the joint
compound, leading to a reduction in the protection pro-
vided by the corrosion inhibitor. Thus, the later scenario
leaves a shrunken and hardened residue that is no longer
effective as a moisture barrier. The presence of moisture
and contaminants in the joint accelerates the connec-
tors aging process and effectively shortens the connec-
tors service life. Figure 2.4-15 shows an infrared image
of a compression splice on which the joint compound is
leaking out of the fill hole located at the center of the
component.
New and Existing Connectors
When designing overhead power lines for high-tempera-
ture operations or contemplating the refurbishment of
existing facilities, a great deal of consideration should be
given to the conductor temperatures at which the con-
nectors were tested. Prudence dictates that connectors
designed for high-temperature operation should be
Figure 2.4-14 Mechanical load data of a new Hawk conductor.
2-49
tested and qualified for temperatures well in excess of
those expected in service.
When evaluating existing connectors prior to the opera-
tion of a circuit at a higher operating temperature, it is
recommended to evaluate the original performance
specifications of the existing connectors. A review of the
standards against which the connectors were designed
and tested assists in evaluating whether these compo-
nents are acceptable for operation at significantly
increased operating temperatures. Operating electrical
connectors at temperatures above those for which they
were designed can be risky, and at the least, a standard
current-cycle test should be performed to evaluate the
performance of each connector.
Mitigation of Effects of Connector High-Temperature
Operations
A cost-effective and frequently used method used to
increase the operating life of connectors is by using rein-
forcing methods such as shunts. Typically, existing con-
nectors that are suspected of being inadequate for the
high-temperature operation of a particular line can be
shunted to reduce their electrical loading and to prolong
their service life. Shunts provide an alternate path for
the current flow, thereby reducing a connectors current
density and operating temperature. The reduction in the
connectors current density retards the aging process,
and extends the long-term performance and service life.
Similarly, shunting of marginal connectors can be used
to extend the service life and to improve the perfor-
mance of components.
Nondestructive Methods to Measure Connectors
The degradation of a connector can be observed by
increases in its resistance and temperature. The
increased resistance is measured with devices such as an
Ohmstik, developed by SensorLink Corp. The tool
takes advantage of the increased resistance by measur-
ing the ac voltage drop across the component that is
caused by the current. Typically, resistance measure-
ment devices such as the Ohmstick can operate at cur-
rents ranging from 2 to 1400 A and can measure
resistances from 5 to 2500 micro-ohms. Figure 2.4-16
shows an infrared image of a defective conductor con-
nection and a summary of the corresponding resistance
measurements.
2.4.14 Conductor Hardware
Conductor hardware, as defined in this report, refers to
noncurrent carrying devices attached directly to the
conductor. Conductor hardware includes components
such as suspension clamps (with or without armor
rods), dampers, repair sleeves and splices, spacers and
spacer dampers, shackles, pins, etc. Generally hardware
is manufactured from cast or forged aluminum, but
there are instances where metallic hardware has been
used, and the issues associated with the use of metallic
hardware are discussed.
Figure 2.4-15 Leakage of connector joint compound.
Figure 2.4-16 Resistance measurement of connector.
2-50
Metallic Conductor Hardware
Metallic conductor hardware for aluminum conductors
is fabricated primarily from aluminum alloys, while
hardware for copper conductors is mostly fabricated
from copper alloys. One of the reasons for the strict
selection and use of similar metals is the recognition of
the differences in the galvanic properties of these materi-
als. When metals of different galvanic properties are
permitted to react in the presence of moisture, the
resulting reaction quickly deteriorates the lower ranked
material, resulting in significant corrosion to an assem-
bly. Nevertheless, galvanized ferrous hardware and com-
ponents have been used extensively in the past because
of the inherent high strength-to-weight ratio and the
fact that the galvanized steel component is relatively
inert to both aluminum and copper in a mild climate
and environment.
High-Temperature Effects of Ferrous Conductor Hardware
Ferrous hardware that either partially or completely
surrounds a current-carrying conductor is subject to
hysteresis and eddy current losses caused by the mag-
netic flux associated with the current flow. These eddy
current losses manifest themselves in the form of signifi-
cant heat gains within the hardware, and hence signifi-
cantly increased operating temperatures. Thus, excessive
heating of the hardware or greatly increased operating
temperatures within the ferrous hardware at levels above
the acceptable level of the conductor is likely to cause
the annealing of any aluminum components. The exces-
sive heating and increased temperatures conveyed by the
ferrous hardware can anneal the conductor or other alu-
minum components, thereby reducing the tensile
strength and the expected service life. While the effect of
the heating is typically localized and adjacent to the fer-
rous component, the effect on the conductor can be det-
rimental and sufficiently destructive to mandate the
eventual replacement.
Heat gain due to hysteresis and eddy current losses in
ferrous hardware is a function of the magnitude of the
current in the conductor and the hardwares thermal
conductivity. Convection and radiation heat losses from
the ferrous hardware are primarily a function of the
hardwares surface area and surrounding ambient con-
ditions. Hence, the operating temperature of ferrous
hardware fluctuates in response to changes in the cir-
cuits load and the ambient conditions.
Conductor hardware is deployed in numerous configu-
rations to support and protect the conductor, and is
available in many different sizes and shapes. Smaller ver-
sions of ferrous hardware have a relatively low ratio of
mass to surface area and usually operate at tempera-
tures well below that of the conductor, regardless of cur-
rent. Conversely, larger versions of ferrous hardware
have a mass-to-surface-area ratio that can result in
hardware temperatures greater than the conductors
allowable annealing temperature at higher currents.
Hardware large enough to produce localized conductor
temperatures of concern are usually confined to suspen-
sion and strain clamps, but can occasionally be caused
by any ferrous device surrounding the conductor as long
as they have a large mass-to-surface-area ratio. Pub-
lished literature, which permits the calculation of the
local temperature increase in a conductors temperature
as a function of the load, is limited.
The mitigation of the effects of localized heating under
ferrous hardware usually involves either limiting the
current rating of a line, limiting the cumulative time a
conductor can operate at the elevated rating, or replac-
ing the hardware with nonferrous hardware. Experience
shows that the number of conductor and hardware con-
figurations are extensive, requiring each utility to
develop customized solutions to address this problem.
Nonferrous conductor hardware does not internally
generate any heat in response to a conductors current.
Generally, nonferrous hardware operates at tempera-
tures well below the conductors operating temperature
because of the increased surface area and the corre-
sponding increase in the convective cooling provided by
the wind, if any.
Nonmetallic Conductor Hardware
Nonmetallic conductor hardware is generally limited to
elastomeric compounds, which serve as compressive
bushings within a hardware assembly. Compression
bushings are typically used in spacers, spacer-dampers,
and armor grip suspension clamps to provide a resilient
interface between the conductor and the hardware.
Publications concerning the effects of high-temperature
operations on elastomeric hardware components are
limited. During and after high-temperature excursions,
the elastomeric components must retain their resilient
and semiconductive properties for long-term survival.
Loss of such properties can result in component deterio-
ration and/or component failure.
High-Temperature Tests on ACSS Conductor Hardware
Reynolds Aluminum performed high-temperature cur-
rent cycle tests of ACSS conductors and associated fit-
tings. Tests were performed on a loop of 1033.5 kcmil
(527 mm
2
), with a 45/7 Stranding, Ortolan SSAC con-
ductor with various fittings installed. In all cases, the
temperatures measured at the conductor hardware were
well below the conductor operating temperature. At a
200
o
C conductor operating temperature, the measured
2-51
full-tension splice temperatures ranged from 120
o
to
130
o
C. The operating temperature of the partial-tension
splices was observed to be significantly higher, ranging
from 160
o
to 170
o
C. On the contrary, the operating tem-
perature at the suspension clamp was 84
o
C, and the
temperature measured at the first insulator pin was
37
o
C. Operating temperatures at the AGS unit were
102
o
C on the armor rods, 46
o
C on the clamp surface,
and 106
o
C beneath the neoprene sleeve.
High-Temperature Tests of ACSR Conductor Hardware
Detroit Edison conducted tests of ACSR conductors
and hardware to determine if high-voltage power line
hardware could be safely operated at temperatures up to
200
o
C. Tests were conducted over a temperature range
from 75
o
C to 200
o
C. Hardware tested included alumi-
num body bolt-type strain clamps, aluminum body sus-
pension clamps, armor rods, parallel groove clamps, line
dampers, and full-tension splices. Hardware was tested
with 477 kcmil (243 mm
2
) ACSR Hawk, 795 kcmil (405
mm
2
) ACSR Drake, 954 kcmil (487 mm
2
) ACSR Cardi-
nal, and 1431 kcmil (730 mm
2
) ACSR Bobolink con-
ductors. The aluminum strain clamp temperature
ranged from 35 to 61% of the conductor operating tem-
perature for the four conductors tested, while the sus-
pension clamp temperature ranged from 29 to 43% of
the conductor temperature. On the contrary, armor rod
temperatures were 68 to 80% higher than the conduc-
tors temperature.
High-Temperature Tests on Polymer Insulators
Test results on polymer insulators indicate that most
insulators are well equipped to handle high operating
temperature loads. An EPRI report (EPRI 2002)
describes the results of a typical conductor suspension
assembly using a polymer insulator. In the test, temper-
atures were measured on the conductor, suspension
clamp, shackle, link, insulator pin, and polymer insula-
tor, as well as the ambient temperature. The tempera-
tures of the test components at the maximum conductor
operating temperature of 292
o
C (not referenced to
ambient temperature) ranged from 163
o
C at the suspen-
sion clamp, to 88
o
C at the shackle, to 43
o
C at the link,
to 32
o
C at the insulator pin, to 28
o
C at the insulator.
The ambient temperature at the time of the test was
measured at 17
o
C.
Another set of tests on polymer insulators had similar
results. In these tests, temperatures were measured at the
insulator end fitting near the fiberglass rod end, near the
suspension fitting end, on the suspension fitting conduc-
tor keeper, on the suspension fitting body near the con-
ductor groove, and on the conductor itself. The testing
included the mechanical loading of the suspension
assembly and current cycling of the conductor to a max-
imum conductor core operating temperature of 250
o
C
(not referenced to ambient temperature). Tests were
conducted on high-voltage polymer suspension insula-
tors of five different manufacturers. The measured insu-
lator end fitting temperature, nearest the fiberglass rod,
ranged from 47
o
to 61
o
C depending on the type of poly-
mer insulator. The end fitting temperature nearest the
conductor was slightly higher, ranging from 54
o
to 67
o
C.
After the thermal testing, each insulator was tested
mechanically to determine the residual strength. Results
showed no measurable degradation of the mechanical
performance of the five polymer insulators regardless of
the manufacturer.
A third set of tests reinforces the same conclusions for
polymer suspension insulator assembly. In each test,
thermocouples were used to measure the conductor
temperature, the insulator end fitting temperature, and
the ambient temperature. Sufficient current was applied
to the conductor to generate a nominal operating tem-
perature of 200
o
C. Tests generated end-fitting tempera-
tures ranging from 38
o
to 46
o
C at ambient temperatures
ranging from 19
o
to 27
o
C.
2.5 UPRATING WITHOUT RECONDUCTORING
2.5.1 Introduction
There are two overall strategies to uprating overhead
transmission lines without reconductoring: using deter-
ministic methods or probabilistic methods.
With deterministic methods, line rating calculations are
done usi ng tradi ti onal tool s, such as the EPRI
DYNAMP program, and assumed ambient conditions.
Physical alterations to the line can be made (such as
retensioning the conductors or raising their support
points) to increase the maximum allowed conductor
temperatures, thereby increasing the ratings. Or, the
assumed ambient conditions themselves can be changed
(such as assuming a higher wind speed).
With probabilistic methods, actual weather data is ana-
lyzed statistically to determine the most viable assump-
tions and the associated risks.
2.5.2 Deterministic Methods
This subsection focuses on uprating methods that avoid
reconductoring, involve minimal capital investment,
and do not require field monitors. Typical increases in
power flow resulting from these options range from 5%
to 50%, depending on the original design conditions, the
present rating assumptions, and the type of structure
and conductor used in the existing line. These uprating
choices are particularly effective for older lines, perhaps
2-52
designed under the old 120
o
F (49
o
C) clearance tempera-
ture.
Si nce the transmissi on conductors are not to be
replaced, the result will be operation of the lines exist-
ing conductors at increased temperature levels. Conse-
quently, the conductor, its hardware, and its connectors
need to be carefully inspected prior to uprating. Any
questionable elements need to be replaced.
In addition to a physical inspection, the line uprating
process must verify that adequate electrical clearances
will be maintained after the uprating is complete. Typi-
cally, this verification should consist of three steps:
1. Measurement of everyday sag clearance and sup-
port point locations by use of conventional survey
equipment, or airborne digital imaging equipment,
with the line out of service or at relatively low electri-
cal load.
2. Calculation of maximum sag (minimum ground
clearances) under electrical load equal to the pro-
posed new thermal rating.
3. Experimental verification of electrical clearances
under a combination of rated load and worst-case
weather conditions through the use of sag or tension
monitors.
Since the maximum allowable temperature is to be
increased, these steps are necessary to be certain that the
additional sag does not violate minimum electrical
clearance requirements and that any increased anneal-
ing of the conductors aluminum or copper strands does
not reduce the lines loading safety factors to an unac-
ceptable level.
Evaluating Sag Clearance Under Everyday Conditions
As discussed in Section 2.3, new lines are designed to
meet certain minimum electrical clearances under all
weather conditions at electrical loads less than or equal
to their thermal rating. They are also designed to limit
the maximum tension under maximum ice and wind
loads to the structure design values. To do this, initial
unloaded conductor sags are specified such that the
final sags at high temperatures and the maximum ten-
sions under ice and/or wind loading are within these
limiting values. By adjusting the initial sags to these
stringing sags at the time of construction, minimum
clearance and maximum tension limits are maintained
throughout the life of the line. The final sags include
an allowance for permanent elongation due to creep
elongation of aluminum and to plastic elongation due to
ice and wind loads. In addition, because of uncertainties
in these calculations, new lines are typically designed
with clearance buffers of at least 3 ft (1 m).
With existing lines that have been in service for 10 years
or more, the measured sag may be assumed equal to
final conditions. That is, the conductor has experi-
enced most of the permanent elongation allowed in the
original design. Unless the line is retensioned, the error
in estimating final sag at the maximum design tempera-
ture is limited to questions about thermal elongation at
high temperature. Also, since the structures have been
located and the support point elevations determined,
the initial sag buffer requirement may be less than that
required for a new line.
Many different techniques are available to determine the
electrical clearance under everyday loadings where the
conductor temperature is quite close to air temperature.
These techniques range from the use of survey crews at
selected spans, to flying the span by airplane or helicop-
ter with digital recording devices. The latter provides
more data than required and costs more. The former
provides less data than one might wish for and costs less.
Particularly with digital recording from the air, the data
can be loaded directly into line profiling and design pro-
grams like PLS-CADD and TL-CADD. This allows
a span-by-span verification of sag and a relatively
straightforward calculation of conductor sag at higher
temperatures.
While the accuracy of these measurements is in the
range of a few inches (several centimeters), the determi-
nation of the corresponding conductor temperature at
the time of measurement is less accurate. Generally, the
conductor temperature is determined by use of a heat
balance equation such as IEEE738 or DYNAMP with
the line electrical load and local weather data. If the
electrical load of the conductor is less than 0.25 A/kcmil
(0.5 A/mm
2
), then the difference between calculated
conductor temperature and actual should be less than
5
o
C. If the electrical load is higher, then the difference
can also be higher, depending on how the calculation is
done and how the weather data (see Section 2.3) is
determined.
Raising Conductor Support Points
The thermal elongation of stranded conductors is essen-
tially the same as that of its component strands. There-
fore, for an all-aluminum or copper conductor, once the
sag at final everyday conditions is established, the sag
at high temperatures can be calculated with relatively
small uncertainty.
For example, consider a line section of an all-aluminum,
37 strand (Arbutus) conductor having a ruling span of
600 ft (183 m) installed to meet the following con-
straints: maximum tension of 50%, 33% initial unloaded
2-53
at 15F (9.4), and 25% final unloaded at 15F (-9.4 C).
An equally typical SAG10 program line design sag-ten-
sion run is in Table 2.5-1.
As an example, consider that the line was originally
designed for a maximum conductor temperature of
120F (49 C) and the line structures were placed such
that the minimum ground clearances are met with the
ruling span final sag of 15.5 ft (4.7 m).
To operate the existing line at 167F (75 C), the attach-
ment points must be raised approximately 2.2 ft (0.67
m). To operate at 212F (100 C), the attachment points
must be raised approximately 4.2 ft (1.28 m).
Given rating conditions of 2 ft/sec crosswind, sun, and
an air temperature of 35
o
C, the rating of the original
line is 345 A. By raising the attachment points by 2.2
and 4.2 ft (0.67 and 1.28 m), to allow operation to 75
o
C
and 100
o
C, the lines thermal rating is increased to 775
and 1010 A, respectively. Note the large increase in ther-
mal rating corresponding to modest increase in attach-
ment height. Many lines built prior to 1970 originally
utilized 49
o
C as the maximum conductor temperature
for clearance estimates.
With wood pole H-frame structures, increasing attach-
ment height in order to increase the line thermal rating
may be particularly attractive. Crossarm attachment
points can be raised, hardware replaced, and shield
wires placed on metal pole top extensions at minimal
cost. Figure 2.5-1 is a photograph of a wood pole struc-
ture where the conductor attachment height has been
raised.
In existing lines having longer ruling span sections, there
are fewer structures per mile (km) to modify but greater
sag increases required to offset increased sag at higher
operating temperatures, as shown in Figure 2.5-2.
Retensioning Existing Conductors
Rather than replacing existing conductors with new, it
may be possible to increase the everyday tension of
existing lines in order to reduce sag at high temperature
and therefore increase the line rating. For example, con-
sider the following case where an existing line with Mal-
lard ACSR is to be uprated.
Table 2.5-2, taken from the SAG10 program, shows the
sag and tension (total, aluminum, and steel component
tensions) for initial and final conditions for 30/19, 795
kcmil (405 mm
2
) ACSR (Mallard) initially sagged so as
Table 2.5-1 Sag-tension Calculations for 37 AAC (Arbutus)
Conductor Arbutus 795.0 kcmil 37 Strands AAC Area = 0.6234 sq in. Dia + 1.026 in.
Temp (F) Ice (in.) Wind (psf) K (lB/F)
Weight
(lb/F)
Sag (ft) Tension (lb) Sag (ft) Tension (lb)
15. .25 4.00 .20 1.451 12.02 5446. 10.65 6140
32. .25 .00 .00 1.143 12.00 4294. 10.06 5118
0. .00 .00 .00 .746 8.77 3833. 6.63 5067.
15. .00 .00 .00 .746 9.67 3475.
a
7.27 4621.
30. .00 .00 .00 .746 10.58 3179. 7.98 4212.
60. .00 .00 .00 .746 12.34 2727. 9.54 3524.
90. .00 .00 .00 .746 13.99 2406. 11.19 3006.
120. .00 .00 .00 .746 15.54 2167. 12.82 2624.
167. .00 .00 .00 .746 17.78 1897. 15.24 2210.
212. .00 .00 .00 .746 19.73 1711. 17.37 1941.
257. .00 .00 .00 .746 21.54 1570. 19.33 1746.
302. .00 .00 .00 .746 23.22 1457. 21.15 1598.
Figure 2.5-1 Photograph of a wood pole H-frame structure
with raised crossarm and pole-top extensions for shield
wires.
2-54
not to exceed a final unloaded tension of 15% of Mal-
lards Rated Breaking Strength at 60
o
F (15.5
o
C). NESC
Medium Loading conditions and conductor tempera-
tures up to 302
o
F (150
o
C) are included, but the original
line is rated at 75
o
C, at which temperature electrical
clearance to ground is near the NESC minimum values.
Notice that the knee-point temperature, where the alu-
minum tension goes to zero, under final conditions,
occurs at only 90
o
F (32
o
C).
It is assumed that the existing conductor has been
inspected, and that the increased vibration that will
result from pulling the conductor to a higher everyday
tension (up to 25% rather than 15% of RBS at 60F
[16C]) can be offset by the addition of dampers.
From Table 2.5-2, it can be seen that increasing the ten-
sion from 15% to 25% RBS will reduce the high-temper-
ature sags at temperatures of 75
o
C to 150
o
C by about 4
ft (1.2 m). Therefore, the maximum design temperature
can be increased to 150
o
C or more without violating
clearance limits.
At the original lines maximum allowable conductor
temperature of 75
o
C, with an air temperature of 40
o
C,
full summer sun, and a wind blowing perpendicular to
the conductor axis at 2 ft/sec, the thermal rating was 735
A. If the retensioned line is rated at 150
o
C with the same
weather conditions, the new thermal rating is 1345 A
(83% higher).
Limitations on Retensioning Existing Conductor
There are several concerns about this method of thermal
uprating:
1. The maximum tension exerted on strain structures
during maximum wind and ice loading has increased
from 7800 to 11,300 lbs. (34.8 kN to 50.4 kN). As a
result, it is likely that these structures would need to
be replaced. An alternative solution may be to limit
the increase in everyday tension so that the increase
in tension loading of existing structures was accept-
able without having to rebuild. This would, of course,
reduce the allowable conductor temperature and the
resulting increase in thermal rating.
Figure 2.5-2 Change in sag for all aluminum conductor
as a function of span length.
Table 2.5-2 Sag-tension Calculations for 30/19, 795 kcmil ACSR (Mallard)
a
Conductor Mallard 795.0 kcmil 30/19 ACSR Area = 0.7669 sq in. Dia + 1.140 in.
Design Points Re-tensioned Original Line
Temp
(F)
Ice
(in)
Wind
(psf)
K
(lb/F)
Weight
(lb/F)
Sag
(ft)
Tension
(lb)
Sag
(ft)
Tension
(lb)
15. .25 4.00 .20 1.955 7.80 11283. 11.26 7823
32. .25 .00 .00 1.667 7.68 9773. 11.39 6600
0. .00 .00 .00 1.235 5.30 10495. 8.97 6202
15. .00 .00 .00 1.235 5.79 9600. 9.66 5760
30. .00 .00 .00 1.235 6.34 8775. 10.35 5377
60. .00 .00 .00 1.235 7.56 7357. 11.71 4755
90. .00 .00 .00 1.235 8.65 6432. 12.57 4433
120. .00 .00 .00 1.235 9.26 6010. 13.26 4204
167. .00 .00 .00 1.235 10.27 5422. 14.33 3891
212. .00 .00 .00 1.235 11.27 4939. 15.33 3637
257. .00 .00 .00 1.235 12.30 4528. 16.32 3419
302. .00 .00 .00 1.235 13.34 4178. 17.28 3230
2-55
2. While the conductor in the existing line had reached
its final sag condition, increasing the tension will
cause additional creep elongation and must be
allowed for in the uprating study (see Section 2.4).
3. The calculation of sag and tension for this Mallard
ACSR ignored the problem of compression in the
aluminum strands. If compression is considered, the
sag at high temperature will be higher than that indi-
cated in Table 2.5-2. This has two effects: the sag
clearance of the original line may not have been ade-
quate at 75
o
C, and the change in sag with tempera-
ture after retensioning will also be greater than
indicated.
4. Conductors and structures that have been in place for
many years will be mechanically and electrically
stressed to increased levels. Unless the existing con-
ductor, structures, hardware, and connectors are
thoroughly inspected, there is a possibility that the
reliability of the line will be less than that of a new
line or one that was reconductored.
Redefining Weather Assumptions
Through a careful review of weather conditions, it may
be possible to use less conservative weather assumptions
for rating calculations. This increases the rating without
the need for physical modification of the line. The tech-
nique is limited by the increased risk necessarily
assumed by allowing higher current operation without
increasing everyday clearances.
Figure 2.5-3 illustrates the difference between actual line
ratings and static ratings. The rightmost bell-shaped
curve represents the probability distribution of line ther-
mal ratings calculated based upon real-time field mea-
surements of weather data. Note that the ratings of the
line typically vary over a range of more than 2:1. The
very lowest ratings correspond to still air, maximum air
temperatures, and full sun. A typical static thermal rat-
ing of 800 A is shown at the left tail of the rating distri-
bution. A less conservative static rating of about 900 A
is also shown. Clearly, the higher the static rating, the
more frequently the actual line rating is less than the
static.
The leftmost distributions are line loadings (which vary
as a result of varying customer load levels and system
configuration changes) appropriate to each of the static
rati ngs. Note that in thi s case, the l i ne loadi ngs
approach, but do not exceed, the static ratings, and for
each loading curve, the load may occasionally exceed
the dynamic rating. This happens more frequently
(larger overlap area) for the higher load distribution.
Unless a dynamic line rating system is in use, the opera-
tor cannot know when the actual line rating is lower
than the line load and therefore is unable to avoid occa-
sional clearance infringements if the load distribution is
high enough.
Actual Line Ratings are Normally Higher than Static
Ratings
It can be seen from Figure 2.5-3 that most of the time
the line rating is well above the static rating and that
under these conditions, the line current could safely be
higher than the static limit. As long as line loads only
rarely approach the line rating, it has been argued that
the static rating can be increased with a negligible effect
on electrical safety. This approach is investigated more
thoroughly in Section 2.6. The method discussed here is
much less scientific and much more common. Lines are
sometimes uprated simply by using less conservative
weather assumptions.
Concerns About Using Less Conservative Weather
Conditions
In a regulated business environment, under ordinary
operating conditions, power equipment was lightly
loaded. High electrical loadings were rare; hence, the
precise determination of high-temperature behavior was
not critical. Some years ago, however, as the regulation
of utility business began to decrease, many older high-
voltage lines have been operated at higher and higher
load levels. This might lead to increased failure rates
and consequent service outages unless the mathematical
models used to specify load limits were refined, and crit-
ical equipment parameters verified by measurement.
Driven by the advent of open transmission access and
deregulation of the utility business, there has been a dis-
tinct trend toward the use of less conservative rating
assumptions with little or no basis in science. Field test-
Figure 2.5-3 Probability density distributions for a typical
circuit load and dynamic rating.
2-56
ing of dynamic thermal methods offered an opportunity
both to evaluate the possible increase in ratings, and to
detect the frequency of occurrences where existing
equipment might be damaged due to less conservative
rating assumptions.
The calculation of thermal ratings for overhead lines is
typically based upon heat balance methods such as that
found in IEEE 738-1993. Given a maximum allowable
conductor temperature, the corresponding maximum
allowable current (the thermal rating) is determined for
worst-case weather conditions. Maximum allowable
conductor temperatures typically range from 50C to
150C. Typical worst-case weather conditions are a
wind speed of 2 ft/sec (0.61 m/sec) perpendicular to the
conductor, with full solar heating and an air tempera-
ture of 30C to 40C.
Table 2.5-3 illustrates the advantage of assuming a
higher wind speed and the consequence of doing so. Use
of a higher wind speed for thermal rating calculations
yields an increase in the line rating, even though the
maximum conductor temperature (100C) remains the
same. For example, an increase in assumed wind speed
from 2 to 3 ft/sec (0.61 m/sec to 0.91 m/sec) yields an
increase in the rating from 990 to 1080 A and, since the
assumed conductor temperature remains the same, no
line modifications are required.
The major advantage of this method of uprating is
clearit is very inexpensive. Since the maximum allow-
able conductor temperature remains the same (100C),
the corresponding maximum sag is unchanged and no
The major disadvantage of this approach is also clear
from the rightmost column of Table 2.5-3. This column
shows the temperature attained by the conductor for
still air conditions , with a line load equal to the calcu-
lated rating shown in column 2. Historically, the joint
probabi li ty of maximum l oading and worst-case
weather was considered a rare event. Recent field studies
indicate that, in certain areas, the probability of still air
may be in excess of 10%. Combined with the previously
noted increase in normal and emergency line loading,
the temperatures indicated in the last column of Table
2.5-3 may be a real concern, and the use of a less conser-
vative wind assumption may impact line reliability.
2.5.3 Probabilistic Methods
The probabilistic approach uses the actual weather data
and conditions prevailing on the line to determine the
likelihood or probability of a certain condition occur-
ring. Such a condition could be, for example, the con-
ductor temperature rising above the design temperature
(the maximum allowed conductor temperature). These
methods have been developed to include a measure of
safety of the line. This can be used as a means of com-
parison of practices between utilities in all countries.
The pros of probabilistic methods are that the risk is
better known and can be quantified and defended if nec-
essary. The designs can be consistent from a risk point
of view in that, if localized weather conditions are used
for different lines in different geographical areas, the
lines can have different ratings, even though their design
temperature and conductor types are the same.
If there is a thermal rating limitation on a line, the prob-
abilistic approach can ensure minimal or zero capital is
used to uprate the line. The line designers also have far
more parameters to vary in increasing the rating of a
line. They could use the load profile, surge, or traffic
patterns to increase the thermal rating of lines. This is
not possible in the deterministic method.
The cons of the probabilistic method are that significant
amounts of data are required to determine the rating.
This includes weather data as a minimum. If more
sophisticated methods are required, the data needs to be
determined on traffic patterns, surge patterns, and load
profiles etc. Regarding the weather data, the variation of
the data with time as well as with area needs to be taken
into account. While the ambient temperature and solar
radiation may be consistent over large areas, the wind
speed and direction may vary within a few meters. It is
also necessary to know how the parameters will vary
into the future. This is particularly important with
regard to the more complicated probabilistic methods
where, for example, load profile is used. It is thus neces-
sary to be aware of these issues in the determination of
ratings.
Table 2.5-3 Effect of Assumed Wind Speed on Thermal
Rating for Drake 795 kcmil ACSR at 100C, Assuming Full
Sun and an Air Temperature of 40C.
Assumed Wind
Speed for Line
Rating Calcula-
tion
Line Rating for
795 kcmil ACSR
@ 100C
When Current =
Rating and Wind Speed
= 0 ft/sec (0 m/sec)
(ft/sec) (m/sec) (A) (C)
0 0 750 100
2 0.61 990 130
3 0.91 1080 145
4 1.22 1160 160
2-57
Three main probabilistic methods are available at
present:
1. Absolute Method
The probability of an unsafe condition occurring can
be quantified in this method. The benefit is that an
absolute measure of safety can be achieved. The
drawback is that the nature of the parameters is
extremely difficult to determine. In addition, the cor-
relation between the parametersfor example, the
weather parametersmust be determined. This
could vary from country to country.
2. Exceedence Method
Historical weather data is used to determine the tem-
perature of the line conductors for a given current
flow. The amount of time that the temperature
exceeds the line design temperature can be determined
for each current level. The utility can then decide on
the current level to use based on the percentage of
excursion or exceedence. The advantage of this
method is that it is relatively easy to determine the
percentages and decide on a level by which to operate.
The disadvantage is that there is no way of determin-
ing the difference in safety (to the public) between, for
example, the 5% and 6% exceedence levels.
An adaptation of the above method is to simulate the
weather data and the current flow to determine the
cumulative distribution of the conductor temperature
as a function of current. This curve could be used to
determine the current and excursion level.
3. Modified Exceedence Method
The safety of a transmission line is estimated by
incorporating all relevant factors. From this method,
a measure of safety can be developed whereby the
practices in different countries can be compared on
an objective basis. The advantage of this method is
that all factors are considered. The variation of the
occurrence of objects under the linefor example, a
traffic patterncan be related to the safety of a line.
Designers can use a wider range of methods, not gen-
erally used before, to increase the thermal rating of
the line. Examples of this are the reduction of surge
magnitudes or the number of surges per year, both of
which can be used to increase the current-carrying
capacity of a line.
By using the measure of safety, system planners and
line designers are in a position to determine the con-
sequences of decisions in a more objective, rather
than a subjective, way. Similarly by using the measure
in conjunction with real-time monitoring systems to
determine exact conductor temperature, operators
can take more objective decisions. Utilities worldwide
would also be in a position to determine the safety of
their lines in relation to those of other utilities.
The Absolute Method of Determining the Probability of
an Accident
Research to date has primarily been confined to
attempts at determining the probability of an unsafe
condition arising. This is determined by ascertaining the
probability of each factor occurring and multiplying the
probabilities. This is represented in Equation 2.5-1.
2.5-1
Where:
P(acc) is the probability of the accident occurring.
P(CT) is the probability of a certain temperature
being reached by the conductor, and is cal-
culated from existing weather conditions,
conductor types, and an assumed current.
P(I) is the probability of the assumed current
being reached, and is determined from the
actual current being measured on a system.
P(obj) is the probability of the electrical clearance
being decreased by an object or person.
P(surge) is the probability of a voltage surge occur-
ring in the line and may be determined
from fault records kept by the power utility
as well as simulations on switching surge
overvoltages on the system. If the surge
occurs simultaneously with an object being
under the line, the likelihood of a flashover
is increased.
Each of the above is considered to be determined inde-
pendently.
P(CT) is determined by the Monte Carlo simulation
technique sampling from distributions of ambient tem-
perature, wind speed, wind direction, and solar radia-
tion to calculate the probability of a certain temperature
being reached given a current transfer. The ambient
temperature, solar radiation, wind speed, and wind
direction are sampled independently to form a set of
parameters from which the temperature of the conduc-
tor is determined.
The problem with the above method is that it assumes
no correlation between weather parameters or the cur-
rent, object, and surge occurrences. This may not be
correct in all cases. The correlation between the individ-
ual weather parameters, as well as the weather parame-
ters and the surge occurrences and objects being under
the line, needs to be ascertained.
This problem can be partly overcome by using sets of
weather parameters, nmeasured at the same time. These
sets will be used to determine the P(CT). Since each set
used is determined from actual recordings of ambient
P(acc) = P(CT) P(I) P(obj) P(surge)
2-58
temperature, solar radiation, wind speed, and wind
direction taken at the same time, the correlation
between the parameters is taken care of.
Determination of Line Rating by the Standard
Exceedence Method
This method assumes that the design temperature of the
line may be exceeded for a small percentage of time, but
that the line remains safe because of the low probability
of other factors, such as high current levels and switch-
ing surge voltages. The amount of time the conductor
exceeds the design temperature expressed as a percent-
age of total time is termed the exceedence level. By
determining the exceedence level, and keeping it con-
stant, the line ratings can be determined using different
weather conditions or different geographical locations.
This method is simpler to use than the absolute probabi-
listic method described earlier and does ensure consis-
tent risk. The standard use of this method assumes that
the full load current will flow at all times. A modified
exceedence method uses the load profile of the line in
question to increase the thermal rating above that of the
standard method.
This (standard) method uses the weather data as well as
the current and conductor characteristics to determine
the frequency of occurrence of each temperature range
attained for a given conductor current. Alternatively the
current that would result in the templating or design
temperature being reached can be calculated for each set
of weather conditions. The frequency of occurrence of
each current range can be determined. The weather data
used can be hourly readings, although the accuracy will
increase with more frequent readings. The steady-state
model can be used for the determination of the conduc-
tor temperature or the current required to reach a cer-
tain temperature.
With reference to Figure 2.5-4, the exceedence levels
have been calculated for the same conductor (Bersfort)
in two separate locations in South Africa. The solid line
represents the exceedence levels experienced in Bloem-
fontein, a hot semi-arid climate, compared to Volksrust,
a small town in a more moderate climatic region. The
number of lines for Volksrust indicates the differences in
the weather data from year to year. The dashed line rep-
resents the average of the years. The graph indicates an
interesting fact in that the operators utilize a single rat-
ing for both sites. The risk or exceedence experienced at
both sites is likely to be different for the same current.
The operators are therefore operating blind with
regard to the risk. The exceedence method has the bene-
fit of applying a uniform risk, or probability of exceed-
ing the design temperature, by utilizing different ratings
at the different locations.
Difficulty often arises in setting the exceedence limit.
One approach could be to determine the exceedence
limits experienced at present for various geographical
areas at different times of the year. This can be achieved
by plotting graphs similar to those of Figure 2.5-4. The
present current limit is used to determine the exceedence
level. This level can then be used to determine the
Figure 2.5-4 Graph showing the results of the use of the exceedence method.
2-59
ampacity of different conductors at different design or
line design temperatures.
Benefits of Using Standard Exceedence Method
The Standard Exceedence Method has been in use by
the National Grid Company (UK) for some years. Table
2.5-4 gives an example for rating its transmission sys-
tem. The percentage excursion time, now called
exceedence, corresponds to the portion of time that
the conductor would exceed the stated design tempera-
ture if it were operated continuously at the correspond-
ing current. These values of current are compared in the
table with their previous ratings assigned on a determin-
istic basis. The exceedence to which the deterministic
ratings actually corresponded, which could not be iden-
tified or acknowledged previously, are seen to vary from
approximately zero to 6%, depending on the rating, sea-
son, and the design temperature.
Adopting a consistent 10% exceedence throughout the
year led to increases in rating from 3% to 30%. More
accurate meteorological studies may allow a rating
increase without increasing the exceedance. The method
also allows greater consistency compared to the other
approach, since the exceedence can be made constant
for all seasons.
Note that the ratings given in Table 2.5-4 are post-con-
tingency values, which are used only in emergency con-
ditions. The risk of flashover depends on the coincident
risk of a maximum voltage surge occurring while worst-
case cooling conditions exist in the span with the critical
clearance. This risk is many orders of magnitude less
than the value of exceedence chosen. In addition to this,
the current is not at the maximum level all the time.
These factors vary from area to area and line to line.
Instead of looking to the probability of an unsafe condi-
tion arising that relates directly to safety, it has been
assumed that to exceed the design temperature is unsafe.
No knowledge exists of exactly how unsafe it is. The
next method, although more complex, goes some way to
solve this problem.
Modified Exceedence Method
The Standard Exceedence Method assumes that the line
current is constant. For lines having a reasonably pre-
dictable electrical load cycle, the load cycle shape can be
used. The calculation is then referred to as the Modified
Exceedence Method.
Effect of the Load Profile on Thermal Rating
The effect of a load profile on thermal rating can be
quite pronounced. This can be shown in the following
set of graphs (Figure 2.5-5) developed for conditions
prevailing in Cape Town (moderate climate with strong
winds).
This load profile was used to generate the exceedence
and modified exceedence curves. The conductor used in
this case was 196 kcmil (100 mm
2
), 6/1, Hare.
Figure 2.5-6 shows the increase in the ampacity of the
conductor using the load profile and local weather data.
The graph shows that the ampacity can be increased, if
we are using an exceedence level of 5%, for example,
from 230 A to 340 A. The two graphs using the load
Table 2.5-4 Comparison of Deterministic and Probabilistic Ratings for 4% - A1/S1 - 54/7 ZEBRA
Design
Temp
Exceedence
% Summer Spring/Autumn Winter
Previous
Deterministic
Rating (A)
Probabilistic
Rating (A)
Previous
Deterministic
Rating (A)
Probabilistic
Rating (A)
Previous
Deterministic
Rating (A)
Probabilistic
Rating (A)
0.1 610 683 770 789 950 847
50 3 745 861 925
6 769 888 954
10 790 912 980
0.1 795 826 896 910 1019 959
65 3 901 994 1046
6 930 1025 1079
10 955 1053 1109
0.1 912 906 1000 981 1090 1025
75 3 989 1071 1118
6 1020 1105 1153
10 1048 1135 1185
2-60
profile for HARE and HARE 95 result from different
years data being used. This shows that in this type of
evaluation it may not be necessary in Cape Town to use
many years data to determine ampacity as each year is
sufficiently variable.
2.5.4 Development of a Measure of Safety as a
Basis for Line Rating
The above probabilistic methodsthe absolute method,
the standard exceedence method, and the modified
exceedence methoddo not make use of all the factors
that affect the thermal rating of a line.
The factors that affect the safety are:
1. Whether or not a surge exists on the line.
2. The magnitude of the surge should one be present.
3. Whether or not an object is present under the line.
4. The size of the object should one be present.
5. The position of the conductor.
6. The probability of flashover as a function of spacing
and shape.
Types of Accidents Occurring Relating to Transmission
Lines
Based on the research conducted to date in Eskom
(South Africa), the area in which this probabilistic
approach to the determination of conductor ampacity
would be most beneficial is on lines above 132 kV. The
objects that would result in an unsafe condition arising
in this case are mainly vehicles. It should be noted that
these findings may vary from country to country, and
similar research may be required.
Simulations of transmission-line flashovers indicate that
the correlation between the various factors make the
calculation of probable safety difficult, if not impracti-
cal. It is not merely a matter of multiplying the probabil-
ities of each of the factors together, but to include the
correlation functions, which are extremely difficult to
determine.
Figure 2.5-5 Load profile for exceedence rating example.
Figure 2.5-6 Results of exceedence rating probability
example.
2-61
Simulation using actual weather and data for each fac-
tor that affects the safety of a line measured at the same
time eliminates the need for correlation evaluation.
Simulation of Unsafe Conditions
The simulation combines different models to determine
whether a flashover occurs to a vehicle or how close the
situation was to a flashover. At Eskom this is done by
means of four modules. Each of the modules simulates
the behavior of the particular parameter. These modules
are:
Surge Module. This module determines the month,
day, hour, minute, and second in which a surge
occurs. It then determines the magnitude of the surge.
Object Module. This simulates the size and time that a
vehicle will be under the line. If the vehicle is under
the line at the same time as the simulated surge, the
data containing the surge time and magnitude, as
well as the vehicle data, is stored.
Position Module. The position module determines the
position of the conductor at the time of the surge.
Accident Module. Using the gap-surge relationship, it
is possible to determine the surge required to create a
flashover. This module determines the difference
between the surge required to cause a flashover with
the actual simulated surge on the line.
Developing a Measure of Safety
Parameters Independent of Current and Line Design
Temperature
The number of surges and vehicles in each category
were analyzed based on the generation of 572 counts,
or times at which the surge and the truck were simulta-
neously under the line.
Average counts per year 12.43
Number of cars 26.4% of vehicles
Number of small trucks 49.8%
Number of large trucks 23.8%
Surges of magnitude 1.6 30.6% of surges
Surges of magnitude 1.7 40.6%
The probability of surges follows the expected distribu-
tion, but the probability distribution for vehicles under
the line is affected by the time span that the different
vehicles remain under the line. Incorporating the surges
generated on the line into the measure of safety is diffi-
cult. One method investigated was to subtract the surge
required to cause a flashover from the surge generated
at the time. The more positive the mean of the differ-
ence, the safer the line. This measure is felt to be the
most valid since it takes into account not only the surges
generated, but also the surges generated at the time the
vehicle is under the line. The designer can readily deter-
mine the probability of an accident occurring since the
integral (number of incidents) of the curve below zero
would represent the number of accidents expected. The
distribution of the new safety measure as a function of
line design temperature is shown in Table 2.5-5.
It appears that the statistical description of the safety
measure as listed above is a valid and reliable means of
determining the safety of a transmission line since the
measure is valid for any line design temperature and
takes into account every parameter dealt with in deter-
mining the likelihood of an accident.
Uses of the Established Measure of Safety
A statistical signature can now be established describ-
ing the safety of a particular line. This can be used as
described in the above example to increase current by
line design temperature or by reducing surge magnitude.
Utilities are now able to quantify the safety of a trans-
mission line. This will enable presentation of the ratio-
nale behind any action taken. Uprating an existing line
or establishing a power transfer rating for a new line can
be justified in an objective rather than a subjective way.
The advantage of the above method is that the probabil-
ity of an accident occurring can be determined. The
safety of the line can be quantified and compared with
other risks such as that of nuclear facilities. This will
assist with objective discussions with regulators.
The model can be used to determine the reliability of
lines by comparing statistical signatures of the difference
between the surge present on the line and that required
to cause a flashover with the insulator string swinging
due to wind. The wind data as well as the surge modules
can be used in this case. Lines in similar areas with dif-
Table 2.5-5 Probability Distributions versus Line Design
Temperature
Line Design Temperature
o
C
40C 50C 60C 80C
Mean
Std. Dev
Max.
Min.
% below 0
1.58
0.84
1.95
-1.9
4.3%
1.64
0.78
3.2
-0.37
4.0%
1.89
0.70
3.09
-0.19
0.8%
2.34
0.56
3.32
0.91
0%
2-62
ferent designs can be compared, and a measure of reli-
ability relating to conductor swing can be developed.
2.5.5 Comparison of Probabilistic Rating
Methods
Table 2.5-6 summarizes the main points of each proba-
bilistic rating method.
2.5.6 Device for Mitigating Line Sag - SLiM
A new class of line hardware, SLiM (Sagging Line Miti-
gator) (Figure 2.5-7), can help solve problems where there
may exist excessive line sag. SLiM reduces excessive sag
in transmission lines and allows transmission owners to
realize higher line ratings and increases transmission sys-
tem performance, reliability, and asset utilization. The
passive design of SLiM and its ruggedness allow trans-
mission owners to treat SLiM like typical transmission
line hardware such as insulator strings. Its installation
procedure is a relatively simple O&M activity.
SLiM reduces excess sag and allows transmission own-
ers to realize higher line ratings, permitting them to
increase asset utilization and maintain safety and reli-
ability in a very cost-effective manner. SLiM has multi-
ple applications for both new and existing transmission
lines to address a host of challenges for transmission
owners. SLiM provides transmission planners, engi-
neers, and asset managers another tool to help them
manage transmission systems in an increasingly chal-
lenging environment.
The SLiM can serve as an economic alternative to con-
ventional solutions, such as:
Replacing the existing conductor with a premium
conductor that can operate at high temperatures
without increased sag.
Reinforcing line structures and foundations for
increased mechanical loading, and either reconduc-
tor with a larger conductor or bundle with the exist-
ing size conductor.
Raising towers and improving foundations at key line
locations to provide for increased clearance.
Adding intermediate towers at key line locations to
increase ground clearances.
Table 2.5-6 Comparison of Probabilistic Rating Methods
Absolute Method Std Exceedence Method Modified Exceedence Method Safety Method
Establishes the absolute proba-
bility of an unsafe condition
arising. Can be used to com-
pare with other industries such
as nuclear safety standards.
Uses relative comparison of
risk. Cannot relate to an abso-
lute probability.
lute probability.
lute probability.
Complex method that requires
explicit equations for the proba-
bilities of current, surges, etc.
Uses simulation. Little
advanced probabilistic theory
required.
required.
required.
Takes into account all factors
that affect thermal rating of
lines.
Takes only the weather data
into account.
Uses weather data and load
profile.
Takes into account all factors
that affect thermal rating of
lines.
Requires analysis of correlation
between weather data and
weather data and other factors
such as surges and load.
Automatically takes care of
weather data correlation by
using actual wind speed, solar
radiation, wind direction, and
ambient
Uses simulation that automati-
cally takes care of correlation if
sets of weather data and load
profile data are used.
Uses simulation taking care of
correlation if sets of weather
data are used.
Can be used on its own.
Can be used in isolation, but
may encounter difficulties in
determining the exceedence
limit to use without reference to
an absolute risk
Can be used in isolation, but
may encounter difficulties in
determining the exceedence
limit to use without reference to
an absolute risk
Can be used as a comparison
between two lines, but needs to
refer to the risk level found in
the absolute method to deter-
mine a reference level.
Analysis of data is required to
determine the probabilistic func-
tions.
Data can be used directly in the
method without statistical analy-
sis of the data itself.
Figure 2.5-7 The SLiM device.
2-63
Installing a sag monitoring system and the infrastruc-
ture to process this information.
How SLiM Works
SLiM uses a shape-memory alloy actuator that is acti-
vated by the same temperature changes that cause a
conductor to change sag. The device is passive, with no
motors or electronic controls. As increasing temperature
increases conductor length, and hence its sag, SLiM
changes its geometry to decrease line length. As conduc-
tor temperature returns to normal, SLiM returns to its
original shape, keeping the conductor always within
acceptable sag and tension limits (see Figures 2.5-8 and
2.5-9).
Some Potential Applications
Here is a bulleted list of situations where a SLiM device
may offer a viable solution:
A system contingency situation can cause loading on
parallel transmission lines to exceed their thermal
limits. These limits are often established to maintain
conductor line-to-ground clearances. Thus, the
action of SLiM, which mitigates the excess sag
caused by high-temperature operation, can allow for
safe line operation during these contingencies. Line
capacity is increased by allowing operation beyond
conventional thermal limits, and expensive line modi-
fication projects may not be required.
Many older lines were constructed to 120F maxi-
mum conductor temperature operation. Studies have
shown that SLiM can enable operation of such lines
at a conductor temperature of about 200F without
compromise of line clearances or tensions. This can
represent a substantial increase of rated line capacity.
System planning may project that certain lines will
become overloaded as local growth increases
demand. In this instance SLiM can delay the need for
either a new line or considerable line modifications
while the anticipated load materializes. Installation of
SLiM is a simple O&M activity and can help bridge
needs.
Line routing or line modifications near airports or
other unique situations quite often require structures
to be as low-profile as possible. SLiM can be
employed in a cost-effective fashion to minimize
tower height for such installations while maintaining
required ground clearances and higher power flows.
Limitations on line ratings to maintain clearances
over road or river crossings can be lifted by using
SLiM while maintaining ground clearances and
higher power flows.
SLiM can be used on new line construction and allow
for use of lower towers, mitigating visual impacts and
community objections.
SLiM Specifications
Electrical Connection
The SLiM device carries the full line current, splitting
the current between the actuator and the body of the
device. Standard flexible connectors carry current
between the transmission line and the SLiM device. The
electrical connectors on the device terminate with 4-bolt
NEMA paddles for easy connection to standard line
hardware.
Mechanical Connection
The device is installed in series with the transmission
line. Either end of the device is equipped with standard
oval-eye end fittings. The mating attachment on the con-
ductor is the choice of the utility. The device accepts any
industry standard dead-end attachment with a 1 in.
(2.54 cm) clevis pin. Options include dead-end compres-
sion fittings, preformed dead ends, and wedge dead ends.
Installation
The device can be installed at a dead end, or anywhere
along a span using procedures similar to line-splicing
techniques. During installation, a piece of conductor
approximately the length of the SLiM device is removed
and replaced by the device. The length of conductor to be
removed as well as the number and locations of devices
along a section of transmission line can be determined
using line-sagging software for optimum performance.
Figure 2.5-8 SLiM is strong, maintenance-free, and designed to have a very long life. It is designed for easy
installation using hot or cold procedures. Industry standard connectors can be used to attach SLiM to the
line. Its operation is configurable to match specific line requirements, and has no negative effects on line
electrical performance.
2-64
Operating Parameters
Tables 2.5-7 and 2.5-8 list some of SLiMs basic parame-
ters and operating characteristics. The standard unit is
targeted primarily to 115- and 230-kV lines, but can
operate in 60- and 70-kV class, as well as 345-kV class
lines. Standard units can also be configured to actuate at
different temperature points based on individual
requirements.
Development and Testing of SLiM
SLiM has been through several years of extensive R&D
and testing including laboratory functionality and reli-
ability testing as well as actual field demonstration.
Functionality testing was performed at PG&Es training
facility in Livermore, California. The sag differential
between the test and control spans at the maximum tem-
perature of 100C was 44 in. (Figure 2.5-10), which
closely matched the predicted effect of SLiM. Prior to
and following the full functionality testing at PG&E,
SLiM and its components were extensively tested for
load and current-carrying capacity, fatigue, corrosion,
and repeatability of performance.
SLiM was also tested at Kinectrics and was subjected to
a number of fault current events ranging from 21 kA to
40 kA (rms) for a minimum of 10 cycles. After current
testing, SLiM was subjected to increasing load until its
break links failed, as designed, at about 49 Kips, exceed-
ing the target range (110% of the conductor breaking
load).
The SLiM device was installed on the bottom phase of a
69-kV transmission line, in SDG&Es service area in
Escondido, California (Figure 2.5-11), and monitoring
equipment was installed directly below the test span on a
12-kV distribution pole. The field demonstration was
intended to prove both ease of installation and function-
ality and reliability. SDG&E crews called the installation
straightforward. SLiM has successfully completed
over a year of field demonstrations.
Figure 2.5-9 The major components of the SLiM device in its Dead-End Configuration and its In-Line
Configuration.
Table 2.5-7 Operating Parameters for Standard
a
SLiM
a. Custom sizes for special applications available.
Criteria Application
Voltage Rating
230 kV and below. Higher voltages
(345 and 400 kV) possible.
Target Conductor
Conductors with a breaking load
40,000 lb (180 kN)
Range of Motion Up to 8 in. (200 mm)
Line Tension @ 110F Up to 5,000 lbs (22.5 kN)
Functional
Temperatures
~120212F (50-100C)
(conductor temperature)
Mechanical Failure
Load
> 49,000 lbs (218 kN)
Tested per IEC at Kinectrics
Electrical Current
Capacity
> 1400 A
Short Current Rating
40 kA (rms)
Tested per IEC at Kinectrics
Total Weight ~ 85 lbs (380 N) (Production Version)
End-to-end
Dimension
~ 5 ft, 3 in. (1610 mm) open, 4 ft, 7 in.
(1410 mm) closed
End Connectors
Any standard connector with 1 in. clevis
pin (utilitys choice)
Installation
Cold using standard procedures or
Live using live-line hand procedure
(similar to splicing procedure). Detail pro-
cedure available upon request.
Table 2.5-8 Example Sag Mitigation for a Drake Conductor
Span
ft / m
a
a. Tension at 40F considered equal to 20% of tensile strength of conductor.
Excess Sag due to Heating Sag
Initial
F / C
Final
F / C
Without SLiM
ft / cm
With SLiM
ft / cm
Reduction
ft / cm
750 / 230 110 / 43 212 / 100 5.2 / 158 0.2 / 6 5.0 / 152
1000/300 110 / 43 212 / 100 6.0 / 183 1.9 / 58 4.1 / 125
2-65
2.6 RECONDUCTORING WITHOUT
STRUCTURAL MODIFICATIONS
2.6.1 Introduction
By replacing the original line conductors, it is possible
to employ modern conductors having lower resistance
for the same diameter and/or having the same maximum
sag as the original conductor but at greatly increased
temperature. The use of certain new conductors can
yield an increase in thermal capacity of as much as
100% at a cost of less than half that of a new line. This
section concerns the application of commercially avail-
able conductors, but the analysis methods are applicable
to new conductor types.
Existing lines may be uprated using methods discussed
in preceding sections of this chapter, all of which involve
using the original conductor. In certain applications, it
may make sense to replace the original conductor with a
new one, usually having a diameter near the original and
often capable of operation at higher temperatures. This
section reviews the various reconductoring choices, and
provides some insight into those line uprating applica-
tions where reconductoring is preferred.
Given the low cost, high conductivity, and low density
of aluminum, no other high conductivity material is
presently used. Therefore, any low-resistance replace-
ment conductor must have increased cross-sectional alu-
minum area, and increased wind/ice and tension loads
on the existing structures.
Figure 2.6-1 shows how the thermal rating of an existing
line may be increased by about 50% by using a replace-
Figure 2.5-10 The sag differential being measured on a control span (bottom) and a test span (top).
Figure 2.5-11 Field demonstration of the SLiM device
at SDG&E.
2-66
ment conductor with twice the original conductors alu-
minum area. Note that this doubles the original strain
structure tensi on l oads and i ncreases transverse
wind/ice conductor loads on suspension structures by
about 40%. Such large load increases would typically
require structure reinforcement or replacement.
High-Temperature, Low-Sag (HTLS) conductors,
capable of continuous operation at temperatures above
100
o
C with stable tensile strength and creep elongation
properties, are commercially available or under develop-
ment. Practical temperature limits of up to 200
o
C have
been specified for some conductors. As is also shown in
Figure 2.6-1, use of an HTLS conductor (having the
same diameter as the original) at 180
o
C also increases
the line rating by 50%, but without any significant
change in structure loads. Raising the structures may
also be avoided if the replacement conductor has a
lower thermal elongation rate than the original.
EPRI is presently engaged in a project to field-test long
samples (several spans) of HTLS conductors in operat-
ing lines. This project is in its early stages at the printing
of this guidebook, but some preliminary observations
are discussed in the following sections. Future versions
of this guidebook will be updated with the latest pub-
lishable results.
The use of a lower-resistance conductor has two advan-
tageslosses are reduced, and operating temperatures
remain modest. The use of HTLS conductor has the pri-
mary advantage that structure modifications are mini-
mized. In either case, the replacement of existing
conductor should improve the mechanical reliability of
the line since conductor, connectors, and hardware are
all new.
Sag Constraints for Reconductoring
As shown in Figure 2.6-2, the original conductors Ini-
tial installed sag increases to a final everyday sag
condition (typically at 60F [16C] with no ice and
wind) as a result of both occasional wind/ice loading
events and the normal creep elongation of tensioned
aluminum strands over time. This everyday final sag
increases further (but reversibly) due to ice/wind loading
or high electrical loads. For most transmission lines, as
shown, maximum final sag occurs as a result of electri-
cal rather than mechanical loads.
Any replacement conductor must be installed such that,
over time, its final sag under maximum electrical or
mechanical load does not exceed the original conduc-
tors maximum final sag. Otherwise, existing structures
will have to be raised or new structures added. HTLS
replacement conductors are usually applied to existing
lines where structure reinforcement or replacement is to
be avoided.
2.6.2 TW Aluminum Wires ACSR/TW or
AAC/TW
While this conductor is limited to operation at moderate
temperatures, the use of compact trapezoidal strands
results in a resistance reduction of about 20% for the
same diameter as the original conductor. The use of alu-
minum trapezoidal wire (TW) wires in place of round
wires potentially increases the cross sectional area of a
round wire conductor of the same diameter by approxi-
mately 20%. Therefore, the use of TW conductor in
uprating offers a reduction in conductor resistance of
20% with no increase in structure transverse loading. If
some increase in conductor diameter over the original is
possible with limited structural reinforcement, the resis-
Figure 2.6-1 Thermal rating as a function of conductor
area and maximum temperature.
Figure 2.6-2 Typical transmission line sag as a function of
time, load, and temperature.
2-67
tance reduction can be in excess of 20%, and the
increase rating can be 20% or more.
For example, consider an existing line with a 477 kcmil
(243 mm
2
), 26/7 Aluminum Conductor Steel-Reinforced
(ACSR) (Hawk) conductor operated to a line design
temperature of 90
o
C. Given reasonably conservative rat-
ing weather conditions, the rating of the line is 650 A.
If this original conductor is replaced by a Calumet/TW
conductor (which has the same diameter) and operated
to the same maximum temperature of 90
o
C, it would
have approximately the same sag at 90
o
C, but its lower
resistance would result in a rating that is on the order of
710 A (9% higher). The maximum tension on strain
structures and for broken wire calculations would be
about 10% higher.
To get a larger increase in thermal capacity, a larger-
diameter replacement conductor could be used, but this
would require the existing structures to handle increased
ice and wind transverse loads as well as even higher ten-
sion loads than Calumet/TW. Other advantages to this
replacement conductor include reduced electrical losses
over the life of the line, and since the conductor and
connectors are new, one might argue that the reconduc-
tored line is capable of safe operation at temperatures
above 90
o
C and that i nstal l ed tensions coul d be
increased if dampers are used.
In summary, reconductoring with ACSR/TW requires
that structures, , especially strain structures, be rein-
forced. The probable increase in line rating will be mod-
est, but the electrical losses over the life of the line will
be less. These conductors are intended to be operated at
usual temperatures, and are not part of the EPRI HTLS
project.
2.6.3 ACSS and ACSS/TW
Many millions pounds of Aluminum Conductor Steel
Supported (ACSS) have been installed and are operat-
ing successfully in the United States. However, for many
it is still considered a relatively new conductor, and its
performance is not well understood. As such, it is an
integral part of EPRIs HTLS conductor field test
project. Most of the initial concerns about installation
and surface roughness problems due to the use of
annealed aluminum strands have passed. The main limi-
tation with ACSS is its relatively low strength and mod-
ulus that may limit its application in regions
experiencing high ice loads. The use of ACSS/TW can
offset this problem to some extent, as can the use of
extra-high-strength steel core wires. The conductor and
special connectors designed for it allow continuous
operation at temperatures up to 200
o
C with conven-
tional galvanized steel core wires. The conductor can be
operated above 200
o
C if Alumoweld or special zinc
Galfan coated steel is used.
ACSS is described in ASTM B 856-95. It consists of
fully annealed strands of aluminum (1350-H0) stranded
around stranded steel core. The steel core wires may be
aluminized, galvanized, or aluminum clad, and are nor-
mally high strength, having a tensile strength about
10% greater than standard steel core wire. In appear-
ance, ACSS conductors are essentially identical to stan-
dard ACSR conductors (see Table 2.6-1).
By using annealed aluminum, the rated strength of
ACSS is reduced by an amount dependent on the strand-
ing (e.g., 35% for 45/7, 18% for 26/7, and 10% for 30/7).
In fact, a 45/7 ACSS conductor has about the same rated
breaking strength as a conventional all-aluminum con-
ductor (e.g., 16,700 lbs (71.4 kN) for 954 kcmil (487
mm
2
) 45/7 ACSS versus 16,400 lbs (73.2 kN) for 954
kcmil (487 mm
2
) 37 strand AAC [Magnolia]). The ther-
mal elongation coefficient, creep rate, and maximum
operating temperature are, however, quite different.
ACSS Conductor Designs
ACSS is typically available in three different designs:
Standard Round Strand ACSS, Trapezoidal Wire of
Equal Area, and Trapezoidal Wire of Equal Diame-
ter. In addition, it is possible to obtain all three ACSS
conductor designs with any of the standard types of steel
core wire (galvanized, aluminized, and Alumoweld).
Advantages and Disadvantages of ACSS
ACSS provides a number of advantages in reconductor-
ing. The combined effect of these factors can make it
economically attractive in thermal uprating applica-
tions. It has higher self-damping than conventional
ACSR. It has lower thermal elongation over a wide
range of conductor temperatures. It can be operated at
temperatures as high as 250 C without damage. It can
be installed at smaller initial sags without dampers if it
is prestressed. With reference to the preceding discus-
Table 2.6-1 ACSS Equivalents to Standard Type 16,
795 kcmil, 26/7 ACSR (Drake)
Conductor
Name OD
Alum
Area AC Resistance
(in.) (mm) (kcmil)
(/
mile)
(/km) (%)
Drake ACSR 1.108 28.14 795 0.1170 0.0727
Drake/ACSS 1.108 28.14 795 0.1137 0.0707 -3%
Suwannee/
ACSS/TW
1.108 28.14 960 0.0939 0.0584 -17%
Drake/ACSS/TW 1.010 25.65 795 0.1132 0.0704 -3%
2-68
sion of sag clearance, the conductor properties make it
attractive for reconductoring applications as well as cer-
tain new line designs.
In reconductoring existing lines, in comparison to con-
ventional ACSR conductors, ACSS can yield a much
larger increase in thermal capacity while minimizing the
need for expensive structure modifications. In new lines,
this conductor can yield designs with less environmental
impact (shorter and/or fewer structures) with greatly
increased thermal capacity for essentially no increase in
cost. As discussed in the following, the key advantages
of ACSS are:
Operate to 250 C with no loss in strength
No creep elongation over time
High self-damping (which yields low levels of Aeo-
lian vibration)
Lower thermal elongation than conventional conduc-
tor
63% IACS conductivity, not 61.2%
Equal OD and equal AREA options.
Higher Maximum Temperature
Typically aluminum stranded conductors can be oper-
ated at temperatures up to 95C without significant loss
of tensile strength. Aluminum conductors with a steel-
reinforcing core can be operated at temperatures of up
to 150 C for limited periods. Because the aluminum in
ACSS is fully annealed at the factory, it can be operated
continuously at temperatures up to 250C or, with spe-
cial high-temperature-tolerant galvanizing coatings
such as Galfan, even higher.
Table 2.6-2 shows a comparison of continuous ampacity
(with 2 ft/sec (0.61 m/sec) crosswind, 40 C air tempera-
ture, and full sun) for ACSS and ACSR conductors.
Note that the ampacity of an ACSS conductor operat-
ing at 250C is nearly twice that of an ACSR of the same
cross-sectional area operating at 100C.
Thermal Elongation
Aluminum strands elongate thermally at twice the rate
of steel. The sag increase of ACSR conductor is, there-
fore, less than it is for AAC. In the case of ACSS, the
tension level in the aluminum strands is very small, and
the conductor elongates thermally as though it were
steel. Thus, the sag increase in going from 15C to
150C with ACSS may be the same as the sag increase
from 15C to 95C with ordinary ACSR.
As an example of this lower thermal elongation of
ACSS, consider the data in Table 2.6-3. The ACSS con-
ductor has the same sag at 150C as the ACSR conduc-
tor of the same diameter has at 100C. Therefore, for a
clearance-limited line, by reconductoring with ACSS,
the thermal capacity of the line increases by about 30%
without the need to raise or reinforce structures.
Self-Damping
The tension of conductors in overhead lines is normally
determined by concern about Aeolian vibration-
induced fatigue. It is normal to limit initial tension to no
more than 20% of the rated breaking strength in order
to limit vibration levels. Because it has higher self-
damping than ordinary ACSR, ACSS may be installed
to smaller initial sags, and because it has a lower modu-
lus, it yields lower maximum tensions than ACSR.
Low Creep Elongation
When reconductoring, one must allow for creep elonga-
tion over time with ordinary ACSR. In addition, except
for ACSR conductors with a high steel content, one
must consider the possibility of accelerated creep at high
operating temperatures. ACSS does not creep at any
temperature, high or low. Thus, its final and initial sags
are the same as shown in Figure 2.6-3.
Not only is there little or no difference between the ini-
tial and final sag, but also the initial sag is less, and the
change in sag due to temperature is less than it is for
standard ACSR.
Table 2.6-2 Continuous Ampacity of Equivalent ACSR and
ACSS Conductors as a Function of Maximum Allowable
Conductor
Temperature
(
o
C) Drake ACSR
a
Suwannee
ACSS/TW
Drake/ACSS
or
Drake/ACSS
/TW
75 730 820 720
100 990 1110 980
150 -- 1490 1320
200 -- 1770 1560
250 -- 2000 1740
a. For continuous loads, ACSR is normally limited to
about 100
o
C to avoid annealing of the aluminum
strands.
Table 2.6-3 Illustration of the Lower Thermal Elongation of
ACSS Conductor
Conductor
Temp
Sag of Drake
ACSR
Sag of
Drake/ACSS Ampacity
(
o
C) (ft) (m) (ft) (m) (A)
15 31.0 9.4 31.0 9.4 --
100 37.6 11.5 35.3 10.8 1110
150 -- -- 37.8 11.5 1490
2-69
The novel characteristics of ACSS make it attractive as
a replacement conductor for HV lines where thermal
capacity is inadequate. ACSS can be substituted for
existing ACSR of the same diameter. Although having
nearly the same resistance and diameter as the conduc-
tor it replaces, ACSS can be operated at a much higher
temperature without exceeding the original high-tem-
perature sag levels. Since the aluminum strands of ACSS
are fully annealed, it has a somewhat lower rated
strength than the same stranding in ACSR. In areas
where ice and wind loads permit, ACSS may be speci-
fied with a reduced steel content. The result is that, with
ACSS, the maximum tension loads on angle and dead-
end structures may be no higher than those generated by
the ACSR conductor that it replaces.
As an example of the advantages of ACSS in reconduc-
toring, consider Figure 2.6-4, which shows ampacity
and sag as a function of maximum allowable tempera-
ture. The original conductor in the existing line is
assumed to be 477 kcmil (243 mm
2
) ACSR (Hawk). The
proposed replacement conductors are 565.3 kcmil (288
mm
2
) ACSS/TW (Calumet), which has the same diame-
ter as the original and 795 kcmil (405 mm
2
) ACSR
(Drake), which has a diameter that is 30% higher. For
continuous operation, the 565.3 kcmil (288 mm
2
)
ACSS/TW (Calumet) conductor at 200C has an
ampacity about 25% higher than Drake at 100C and
lower maximum sag than the original or replacement
ACSR conductors.
ACSS/TW: Field Trial in the EPRI HTLS Conductor
Project
As part of the EPRI HTLS conductor project, an
ACSS/TW conductor was spliced into a line segment of
an operating 138-kV transmission line. This test seg-
ment consists of four spans, and is approximately 2880
ft in length, and includes five structures (two dead-end
and three suspension towers). The test conductor was
spliced into all three phases of one circuit of a double-
circuit vertical line. Various field data associated with
conductor performance are intended to be collected
over an extended period of time (about three years).
The conductor is classified as Trapezoidal Shaped Wire
Concentric-Lay Aluminum Conductor Steel Supported
(ACSS/TW). It is designated by the name Suwannee,
and is 1.108 in. (2.814 cm) in diameter. Figure 2.6-5
shows photos of the conductorthe outside aluminum
strands and steel center strands are indicated.
Figure 2.6-3 Typical behavior of ACSS conductor,
illustrating that initial and final sags are nearly identical.
Figure 2.6-4 Ampacity and sag of original Drake ACSR
and Calumet ACSS/TW replacement conductor as a
function of maximum allowable temperature.
Figure 2.6-5 ACSS/TW cable, manufactured by
SouthWire, installed on operating test line.
2-70
During the course of the project (ongoing for just over
two years at this time), measurements and observations
were made of the following quantities:
Sag and Tension
Weather Parameters
Average Conductor Temperature
Current
Splice Resistances
Hardware Temperatures via Infrared Measurements
Corona
Electric and Magnetic Fields
Visual Inspections
This project is still ongoing, and final results are not yet
available
2.6.4 High-Temperature Aluminum Alloy
Conductors
T-Aluminum Conductor Steel Reinforced (TACSR) is a
conductor widely used in Japan. A special type of steel
core (Invar) adds expense but reduces thermal elonga-
tion. There is extensive laboratory test data on the Zir-
conium aluminum alloy wire materials (TAL and
ZTAL). There appear to be no special problems with
installation and termination of (Z)TACSR. TAC can be
operated continuously up to 150
o
C and ZTAC to 210
o
C
without loss of strength.
The various Japanese manufacturers (e.g., Fujikura
Ltd., Sumitomo Electric Industries, Ltd.) have devel-
oped a whole range of special high-temperature conduc-
tors. These conductors consist of special temperature-
tolerant aluminum alloy wires combined with ordinary
steel or a special low-thermal-elongation steel wire
called Invar. The acronyms for these conductors indi-
cate the type of aluminum alloy (TAC, GTA, UTA,
XTA, and ZTA); the type of steel core wire (SR or IR);
and whether the aluminum strands are trapezoidal; and
whether there is a gap between the inner layer of alumi-
num and the steel core (e.g., GACSR or GTACIR).
A partial list of the most common types includes the fol-
l owi ng names : TACSR, GTACSR, UTACSR,
GTACSR, UTACIR, XTACSR, XTACIR, ZTACSR,
and ZTACIR. The acronyms refer to the type of high-
temperature alloy, whether the conductor is gapped,
and the type of steel core material.
As a simple comparison, consider Table 2.6-4, a sum-
mary of the maximum operating temperatures of the
various Japanese heat-resistant conductors.
High-Temperature Alloys of Aluminum
Table 2.6-5 is a description of the heat-resistant alloys of
aluminum.
The TAL alloy was developed in the 1960s. The other
alloys were developed in a continuing attempt to keep
the conductivity near that of ordinary electrical conduc-
tor grade aluminum (1350-H19). The relationship
between conductivity and maximum continuous tem-
perature is shown in Figure 2.6-6.
2.6.5 Special Invar Steel Core
ACSR conductors are manufactured with a variety of
steel wire coatings to prevent corrosion. Normal steel
core wire has a tensile strength of 170 to 190 psi (1170 to
1310 Mpa). Invar steel wires have a 15-20% lower tensile
strength but also have a much lower coefficient of ther-
mal expansion than conventional galvanized steel wire.
The thermal expansion coefficient of conventional steel
is 11.5 10
-6
per-degree-C, whereas the thermal coeffi-
cient of Invar steel is only 2.8 10
-6
per-degree-C.
At high operating temperatures, the aluminum strands
of any high-temperature conductor unload tension
almost entirely to the steel core. With Invar, this hap-
pens at a lower (knee point) temperature. In addition,
Table 2.6-4 Maximum Operating Temperatures (C) for
High-Temperature Alloys Made in Japan
Description Symbol
Max Temp
Continuous
Max Temp
Emergency
Super Heat
Resistant
UTACSR 200 230
Super Heat
Resistant
ZTACSR 210 240
Super Heat
Resistant
XTACSR 230 310
Heat Resistant TACSR 150 180
Normal ACSR 95 125
Table 2.6-5 Conductivity of High-Temperature Alloys Made
in Japan
Aluminum
Alloy % Conductivity
Max Temp
Continuous
Min. Tensile
Strength
(IACS)
(C)
(kgf/mm
2
)
UTAL 57.0 200 16.2 to 17.9
ZTAL 60.0 210 16.2 to 17.9
XTAL 58.0 230 16.2 to 17.9
TAL 60.0 150 16.2 to 17.9
1350-H19 61.0 95 16.2 to 17.9
2-71
the rate of increase in sag with further increases in con-
ductor temperature is less with Invar steel cores. This is
demonstrated in Figure 2.6-7.
The EPRI HTLS project has begun the field testing of an
Invar Drake conductor spliced into a five-span section
of a 230-kV line. The installation process went well, but it
is too early in the project to make any further meaningful
conclusions at this time. The data shown in Table 2.6-6
show the physical properties of this conductor.
2.6.6 Gapped Construction
Gapped ACSR has been used both in Japan and
England. The conventional steel core is surrounded by a
layer of trapezoidal aluminum wires, and the gap filled
with grease. Through the use of special terminations
and suspension clamps and by preloading the steel core,
the thermal elongation of the conductor is less than that
of conventional ACSR, while maintaining the full
strength of a conventional ACSR conductor under
heavy ice conditions.
The lower temperature range aluminum alloys are
optionally supplied in a gapped construction, as
shown in Figure 2.6-8a picture taken from a Sumit-
omo Technical Data Sheet.
In the gapped construction, the space between the steel
core and the inner layer of the aluminum alloy strands is
filled with high-temperature grease to prevent corrosion.
In addition, the gapped construction conductors are
installed with full tension on the steel core (and little or
no load on the aluminum strands).
It was noted in the preceding comparison of Invar with
conventional steel wire that Invar has a reduced tensile
strength. While it is conceivable that a gapped construc-
tion conductor could be made with an Invar steel core
for use in a light-loading region such as Arizona, it is
not commonly done in Japan, where heavy ice and wind
loads commonly occur. Thus, as shown in Figure 2.6-8,
Figure 2.6-6 Plots of conductivity (top) and loss of
strength (bottom) for high-temperature Japanese
aluminum alloys.
Figure 2.6-7 Comparison of ACSR-type conductors with Invar and conventional steel cores.
2-72
Table 2.6-6 Physical Properties of a Drake Invar Conductor
Item Unit Specifications
Cable designation -- Hi-STACIR/AW 795kcmil (Drake)
Stranding wire composition
Super thermal-resistant aluminum-alloy wire
High tensile strength aluminum-clad invar wire
Nos./mm
Nos./mm
26/4.44
7/3.45
Minimum rated tensile strength kgf 13,630
Calculated cross-section area
a
Super thermal-resistant aluminum-alloy wire
Complete conductor
a. Tabulated values are for reference calculated on standard diameter and density.
mm
2
mm
2
mm
2
402.56
65.44
468.0
Calculated overall diameter
a
Complete conductor
mm
mm
10.35
28.11
Calculated nominal weight
a
kg/km 1,582
Calculated D.C. resistance at 20C
a
ohms/km 0.0706
Typical modulus of elasticity
a
Up to transition point temperature
Above transition point temperature
kgf/mm
2
kgf/mm
2
7,590
15,500
Typical coefficient of linear expansion
a
Up to transition point temperature
From transition point temperature to 230C
From 230C to 290C
1/C
1/C
1/C
17.5 x 10
-6
3.7 x 10
-6
10.8 x 10
-6
Maximum
operating temperature
Continuous C 210
for emergency C 240
Calculated current carrying
capacity
b
b. Current carrying capacity is calculated on the following conditions:
Ambient air temperature (C) 40
Maximum temperature (C) 210
Frequency (Hz) 60
Wind velocity (m/s) 0.61
Total solar and sky radiated heat flux at sea level (W/cm) 0.10
Emissivity 0.50
Solar absorptivity 0.50
Continuous A 1,628
for emergency A 1,752
The direction of lay of the outer most layer -- Left-hand (S)
Standard length per reel m 200 -0%/+1.0%
Figure 2.6-8 Summary table showing gapped and conventional constructions for
Japanese high-temperature conductors.
2-73
Gapped conductors are designed with conventional
high-strength steel core wires.
Gapped Conductor: Field Trial in the EPRI HTLS
Conductor Project
The EPRI HTLS project has begun the field testing of a
Gap conductor spliced into a four-span section of a 230-
kV line. The installation process went well, but it is too
early in the project to make any further meaningful con-
clusions at this time. The data shown in Table 2.6-7
show the physical properties of this conductor.
In addition to a field trial on an operating transmission
line, the EPRI HTLS project included a demonstration
and training session on the installation of a Gap con-
ductor on a full-scale test line at the EPRI Engineering
and Test Center in Lenox, Massachusetts. The installa-
tion of a Gap conductor is somewhat unique, particu-
larly in the need to strip back a sizeable amount of the
aluminum outer strands in order to expose and splice
the steel core separately from the aluminum layers in a
two-step splice arrangement.
There are presently no other installations of this con-
ductor in North America; however, there are many
applications in other countries (e.g., Japan, UK, Saudi
Arabia). Figure 2.6-9 shows a photo of a lineman from a
North American utility being trained by a specialist
from Japan on the splicing technique. The participants
Table 2.6-7 Physical Properties of a Drake Invar Conductor
Item Unit Proposal
Construction Nos./mm
16/4/4-ZTAI
10/TW
a
-ZTAI
7/3.2-Est
a. TW Trapezoid wire
Direction of outer lay -- Z-Strand
Minimum breaking strength kN 149.2
Lay ratio
(length/diameter)
Aluminum layer
Outer layer
Inner layer
times
10-14
8-16
Steel core 16-26
Maximum D.C. resistance at 20C /km 0.0714
Calculated cross-sec-
tional area
Super thermal resistant
aluminum alloy
mm
2
413.2
Steel 56.29
Total 469.5
Outer diameter mm 27.8
Weight kg/km 1614
Modulus of Elasticity
Conductor
Steel Core
GPa
79.1
205.9
Coefficient of linear
expansion
Conductor
Steel Core
x 10
-6
/C
19.4
11.5
Cross sectional view
Figure 2.6-9 A lineman being trained on the installation of
a Gap conductor.
2-74
of the training session generally believed that, although
the installation and splicing technique is different than
what they are accustomed to, it was easy to learn and
very do-able. However, it is recommended that prior
to an installation, linemen unfamiliar with the methods
should be provided training on the techniques involved.
2.6.7 ACCR Conductor
The Aluminum Conductor Composite Reinforced
(ACCR) is commercially available in limited quantities
from the 3M Company. Reasonably extensive tests have
been performed on several sizes of this conductor under
laboratory conditions, and terminations and suspension
clamps are available from Preformed Line Products.
Xcel Energy, Hawaiian Electric Company, and Western
Area Power Administration have successfully completed
test installations. The installation of this conductor
appears to be fairly straightforward, but may require
special large blocks and careful handling.
The key advantage of ACCR is that the composite core
strands have a conductivity of about 40% International
Annealed Copper Standard (IACS) and have the modu-
lus and tensile strength of steel but are approximately
the same density as aluminum.
The ACCR conductor is about to be field tested in the
EPRI HTLS conductor project. Future updates to this
guidebook will provide more details as they become
available.
2.6.8 Conductors with Exotic Cores
There are other less well-known conductors that are
either still in the development stage, or in early trials
compared to the other HTLS conductors. One is a fiber-
glass core conductor that was a popular topic of discus-
sion in the power industrys research community, but
interest seems to have dwindled recently. Another has a
carbon fiber core with a slightly negative coefficient of
thermal expansion. Development of this conductor was
supported by the National Science Foundation, and a
patent has been issued. However, no significant labora-
tory or field tests have been reported thus far.
Another conductor that involves a carbon and polymer
fiber core, referred to as ACCC (aluminum conductor
composite core), has received some press coverage
recently. Very little laboratory test data have been pub-
lished, and there is not much expertise in the industry
on it. Apparently, there are other business-related issues
about the ability to procure it at this time.
The EPRI HTLS project has begun field testing a sam-
ple of the ACCC conductor in four spans of a 69-kV
line. The installation process went well, but it is too
early in the project to make any further meaningful con-
clusions at this time. Figure 2.6-10 shows a photo of the
conductors cross section. The core has the look and feel
of a rod that is somewhat rigid, but with some flexibility.
This conductor requires some special hardware and
installation techniques, and linemen would require some
degree of training prior to installation. Also, as with any
core that has a polymer component, there is some con-
cern about long-term performance.
2.6.9 Comparing ACSS and High-Temperature
Alloy Conductors
The major advantage of using ACSS is its cost (typically
sold at a premium of less than 50%) and its wide avail-
ability outside of Japan. Also, ACSS has been used
extensively, and most of the handling and installation
difficulties are well understood.
The major advantage of the High-Temperature Alloy
conductors is that they can be used in regions experienc-
ing heavy ice and wind loads (ACSS may not), and they
are applicable to EHV lines, where surface roughness of
ACSS may yield higher corona noise and radio noise
levels. The cost of these conductors, however, appears to
be relatively high (probably a premium in excess of
100% over conventional ACSR). The availability of
these high-temperature alloy conductors outside of
Japan is uncertain at this point, and shipping costs
would simply worsen the cost issue.
The selection process for HTLS replacement conductor
is unique to each line uprating, but the most important
aspect is sag as a function of temperature. Consider an
existing line with 795 kcmil (405 mm
2
) 26/7 Drake
ACSR installed in a 1000 ft (305 m) ruling span to an
initial unloaded tension equal to 20% of its rated break-
ing strength at 60F. The everyday initial sag of 21.8 ft
Figure 2.6-10 Cross section of an ACCC
conductor.
2-75
(6.6 m) increases to 25.7 ft (7.8 m) over the life of the
l i ne. The i ni ti al conductor tensi on i s 15, 300 l bs
(68.3 kN) under maximum ice and wind load.
At the original maximum conductor temperature of
100C (212F), the ruling span sag is 31.7 ft (9.7 m). To
avoid raising the existing structures, any HTLS replace-
ment conductor will also be limited to a sag of 31.7 ft
(9.7 m) at its maximum temperature.
The thermal rating of DRAKE at 100C is 990 A for
2 ft/sec (0.61 m/sec) crosswind, 40C air temperature,
and solar heating for summer at noon. The sag behavior
of HTLS replacement conductors (ACSS, ACSS/TW,
ACCR(3M), GTACSR, and TACIR) are compared in
Figure 2.6-11, where each of the HTLS alternatives has
the same di ameter as Drake and the same fi nal
unloaded sag at 60F as the original conductor.
From this figure, one can see that the ACCR conductor,
with its very low thermal elongation, attains the highest
operating temperature of 370F (190C). Given the rela-
tively high conductivity of its composite core, the ther-
mal rating is 1550 A.
ACSS/TW reaches the maximum sag (31.7 ft (9.7 m)) at
120C, given the thermal elongation of its steel core,
which yields a thermal rating of 1270 A with its some-
what higher aluminum cross-sectional area.
T-Aluminum Conductor Invar Reinforced (TACIR)
reaches the maximum sag at 270F (132C) for a ther-
mal rating of 1220 A.
Gapped T-Aluminum Conductor Steel Reinforced
(GTACSR) reaches the sag limit at 124C for a thermal
rating of 1170 A.
For all of these replacement conductors, the conductor
temperature at the maximum sag of 31.6 ft (9.6 m) is
well below their continuous operating temperature limit.
The thermal rating comparison could be quite different
if the line were not clearance limited or limited to a
higher sag. Similarly, the results could be quite different
if the final everyday sags of the HTLS conductors were
different due to differences in vibration damping or
structure tension load limits.
2.7 DYNAMIC MONITORING AND LINE
RATING
2.7.1 Introduction
If dynamic rating methods are applied to increase the
effective rating of an overhead line, real-time weather
data and, optionally, line temperature or sag-tension
data must be communicated from multiple remote loca-
tions to the operations center where the line rating cal-
culations are performed. In all such cases, the line rating
is no longer constant but varies with weather conditions.
This technology has been implemented at a number of
EPRI member utilities and is worth considering in cases
where there is a need for a modest increase in rating at
minimum capital investment. The technology requires
operational flexibility and available SCADA/EMS com-
munications.
2.7.2 Dynamic Ratings Versus Static Ratings
found in IEEE 738-1993 (see Section 2.3). For static rat-
ings, given a maximum allowable conductor tempera-
ture, the corresponding maximum allowable current (the
thermal rating) is determined for worst-case weather
conditions.
As discussed in Section 2.2, for most existing transmis-
sion lines, the maximum allowable conductor tempera-
tures typically range from 50C to 150C. In most cases,
the maximum temperature is limited in order to avoid
excessive conductor sag (referred to as a clearance lim-
ited line), or, in some cases, a loss in conductor
strength (referred to as a thermally limited line).
Also, as discussed in Section 2.5, most power utilities
(both domestic and foreign) assume worst-case
weather conditions that are not really worst, but rather
conservative. Worst-case line rating conditions would be
the peak 1-hour air temperature and still air with full
solar heating. Conservative line rating conditions
assume a wind speed of 2 to 3 ft/sec (0.6 to 1 m/sec) per-
pendicular to the conductor with full solar heating and
Figure 2.6-11 Typical plot of sag versus temperature for
various HTLS conductor types.
2-76
a reasonably high air temperature of 30C to 40C. As a
result, dynamic line ratings are usually higher than
static line ratings but can occasionally be less.
Table 2.7-1 illustrates the advantage and consequence of
various wind speed assumptions. Use of a higher wind
speed for static thermal rating calculations yields an
increase in the line rating, even though the maximum
conductor temperature (100 C) remains the same. For
example, an increase in assumed wind speed from 2 to 3
ft/sec (0.61 m/sec to 0.91 m/sec) yields an increase in the
rating from 990 to 1080 A and, since the assumed con-
ductor temperature remains the same, no line modifica-
tions are required.
The major advantage of this method of uprating is
clearit is very inexpensive. Since the maximum allow-
able conductor temperature remains the same (100 C),
the corresponding maximum sag is unchanged and no
The major disadvantage of this approach is also clear
from the rightmost column of Table 2.7-1. This column
shows the temperature attained by the conductor for
still air conditions, with a line load equal to the calcu-
lated rating shown in column 2. Historically, the joint
probabi li ty of maximum l oadi ng and worst-case
weather was considered a rare event. Recent field studies
indicate that, in certain areas, the probability of still air
may be in excess of 10%. Combined with the previously
noted increase in normal and emergency line loading,
the temperatures indicated in the last column of Table
2.7-1 may be a real concern, and the use of a less conser-
vative wind assumption may impact line reliability.
2.7.3 Advantages of Dynamic Rating
A Flexible Response to Uncertain Load Growth
In a regulated utility environment, circuit load growth
was reasonably predictable, and the corresponding need
for increases in circuit capacity could be predicted years
in advance. In the increasingly open access environ-
ment, circuit load growth is much less certain, and pro-
viding appropriate increases in circuit capacity is much
more difficult. Also, in our present economic state, large
capital expenditures are unattractive, especially if those
expenditures turn out to be unnecessary. Dynamic mon-
itoring leading to modest practical increases in capacity
for equally modest capital investment appears to be an
attractive uprating alternative.
Avoid Circuit Outage
With those methods of uprating existing lines that
require physical modification of the linereconductor-
ing, raising support points, retensioningthe line must
be de-energized. This may be expensive if loss of service
results in increased generation costs. Installation and
calibration of some line monitoring devices do not
require taking the line out of service for more than a few
hours. With some noncontact monitors, the line may
stay in service during installation.
Monitoring Equipment May be Moved and Reused
Monitoring equipment and communication links are, in
general, reusable. Thus they may be applied to lines on a
temporary basis, allowing postponement or avoidance
of a more traditional uprating project. If the line is even-
tually physically modified, the monitors can be used at
another location. This process is limited by the durabil-
ity of monitors and the rate of change in communica-
tions equipment.
Improved Clearance Accuracy
One of the benefits of real-time line monitoring is the
improved understanding of how existing transmission
lines behave when subjected to heavy electrical loading.
Such high loading events in a regulated environment
were rare, and errors in clearance estimation were, there-
fore, of little concern. Given the difficulty in getting new
lines approved and the increased utilization of existing
lines under both normal and emergency conditions,
accurate determination of electrical clearances along the
line is becoming essential. Real-time monitors com-
bined with direct communication links to SCADA allow
the system operator to load existing circuits with confi-
dence that minimum clearances are being met. This
allows increased utilization (higher ratings) during most
loading situations and the avoidance of dangerous clear-
ance violations during those increasingly frequent times
when the line is heavily loaded during a period of poor
rating weather conditions (low wind speed, high solar
heating, high air temperature).
2.7.4 Disadvantages of Dynamic Rating
Need for Real-time Communication to SCADA
Transmission owners have not traditionally monitored
overhead lines, especially in real-time. Whatever type of
Table 2.7-1 Effect of Assumed Wind Speed on Thermal
Rating for Drake 795 kcmil ACSR at 100C, Assuming Full
Sun and an Air Temperature of 40C
Assumed Wind
Speed for Line
Rating Calculation
Line Rating for
795 kcmil ACSR
@ 100C
when current =
rating and wind speed =
0 ft/sec (0 m/sec)
(ft/sec) (m/sec) (A) (C)
0 0 750 100
2 0.61 990 130
3 0.91 1080 145
4 1.22 1160 160
2-77
monitoring device is used, communication of the mea-
sured values back to the system operator is essential.
This requires two communication links, one from the
monitor to a nearby substation, and a second from the
substation to the operations center. These communica-
tion links must be set up and maintained if the dynamic
monitoring and rating method is to be useful.
Installation and Material Cost
The cost of any dynamic monitoring system must be
weighed against the cost of other uprating methods.
Costs include monitors, communications equipment,
software, and engineering as well as the cost of main-
taining each. In general, the cost of real-time monitor-
ing and rating should be significantly less than the cost
of more conventional uprating methods.
Operational Issues
In contrast to the other line uprating methods discussed
in this report, dynamic monitoring and rating methods
require a change in the way system operators limit cir-
cuit loading. With such methods, the system operator
sees a circuit load limit that varies significantly with
time, and at certain times can be less than the present
fixed line rating. This typically requires some mental
adjustment by operators that is partly offset by nor-
mally higher line ratings. A culture change regarding
line ratings needs to take place. This can be facilitated
by performing some upfront dynamic rating studies. It
can prove very useful to gather data offline for a period
of time, then run real-time simulations on the data. This
can help educate operators (and engineers and manag-
ers) about the technology and its usefulness. Also, this
data itself can be very useful in identifying hidden power
capacity in lines.
2.7.5 Real-time Monitors
The maximum electrical power flow down an overhead
transmission line is typically determined by the need to
limit conductor sag and thus maintain minimum ground
clearances that are specified by a maximum allowed
conductor temperature. Various monitoring methods
have been proposed and tested, all of which are typically
applied to determine the lines sags in all its spans and
the maximum current that can be carried without violat-
ing minimum electrical clearance requirements in any
span.
The following real-time monitors are either commer-
cially available or have been field-tested at a number of
locations:
Weather Stations
Weather stations generally measure wind speed and
direction, air temperature, solar intensity, and rain. If
the line current and weather conditions are known in
real-time, the conductor temperature near the weather
monitor can be calculated in real-time (EPRI 1995).
Weather stations can include standard propeller-type
anemometers, or the more sophisticated 3-D ultrasonic
units. The latter are quite expensive, but have no moving
parts and are therefore very reliable, are very accurate
even at low wind speeds, and can measure vertical air
movement. Figure 2.7-1 shows a photograph of a
weather station with both anemometer types.
Conductor Temperature Monitors
Conductor temperature monitors incorporate a clamp-
on thermocouple, attached directly to the energized
conductor and linked to a ground station by radio. The
accuracy of temperature monitors depends on how close
the measured conductor temperature at one spot is to
the average line section temperature. It has been
observed that conductor temperature can vary signifi-
cantly along its length due to large variations in wind
speed and direction.
Line Tension Monitors
A load cell can be used to determine the line tension.
The load cell is placed on the grounded side of dead-end
insulator strings. The measurement of line tension can
then be converted to the average temperature of the line
section.
A base station is mounted on the structure and con-
nected to the load cells by cable. Communication to a
base station is usually by spread spectrum radio, and the
units can be solar powered. The line is normally de-
energized when the load cells are installed.
Figure 2.7-1 A weather station with a 3-D ultrasonic
anemometer mounted next to a standard propeller-type
anemometer.
2-78
Sag Monitors
EPRI recently developed a device for monitoring con-
ductor sag in real-time. It is based on digital video tech-
nology, and is called the video sagometer (EPRI 2001).
Figure 2.7-2 provides a very simple illustration of the
basic concept and main components of the video sago-
meter system. The main components include a video
camera based on highly light-sensitive charge coupled
device (CCD) technology, a passive reflective target, a
solid-state target illuminator, a communication system, and
associated electronics.
The camera unit is typically mounted on one of the
structures of the line being monitored, but it could be
mounted on any appropriate structure in the vicinity. A
small, passive, reflective target is placed on the conduc-
tor being monitored. A low-power solid-state illumina-
tor (diode laser or LED-based device) is mounted with
the camera to illuminate the target at night or when
ambient light is not sufficient.
Image recognition algorithms residing in local firmware
determine the position of the target within the cameras
field of view. The targets ground clearance is deter-
mined from that position through a calibration proce-
dure performed during installation. The conductors sag
and/or ground clearance is determined at any point
along the span from the catenary equation. The systems
have proved to be incredibly accurate.
Readings are taken at user defined intervalstypically
about every 10 minutesand stored in an onboard data
logger. In addition to the sag/ground clearance data, the
data logger also records the date, time, temperature, cor-
relation factor, and other data described in the reference
(EPRI 2001).
Communication of the data back to a control center or
engineering office is accomplished via cell phone and/or
spread-spectrum radi o. Data can be retrieved i n
archived blocks or provided in real time, although real-
time transmission by cell phone is generally impractical.
In addition, the system can be configured to transmit
digitized images of the cameras field of view.
The systems can be powered by solar-cell/battery
arrangements, or by standard ac distribution power if
available at the site. The systems, including the targets,
can readily be installed on energized EHV transmission
lines, or readily removed and relocated.
Utilities have used these systems in a variety of ways.
Some simply monitor the ground clearance in real time
as a simplified means of rating their lines in real time.
Others use the more sophisticated approach of using the
real-time data in conjunction with EPRIs Dynamic
Thermal Circuit Rating (DTCR) software to perform real-
time rating calculations. Some have used the data in
uprating studies. Figure 2.7-3 is a photograph of an
installation of a video sagometer on a wood pole. The
communications and associated electronics are mounted
in boxes near the base of the pole.
The video sagometer underwent extensive testing at the
EPRI-Lenox facility, and now has a proven performance
record on operating transmission lines throughout
North America, and over a line voltage range of 69 kV
to 500 kV. The video sagometer offers several features:
Figure 2.7-2 A simplified illustration of the video
sagometer concept.
Figure 2.7-3 The video sagometer mounted on a wood
pole.
2-79
It can easily be installed, or moved, without taking a
line outage. In fact, of all the systems installed in
North America, most were installed while the line
was energized.
It does not have to be installed at a dead-end struc-
ture. It can be installed on any of the transmission
structures, on nearby poles, on its own pole, or
mounted anywhere in the vicinity of the line.
Because conductor sag depends on average conductor
temperature, sag measurements are equivalent to tem-
perature measurements (and tension measurements).
This is important for lines that are thermally con-
strained, and makes it possible to use heat balance
equations for rating computations.
The system has proven to be incredibly accurate, pro-
viding sag measurements to better than quarter-inch
accuracy.
The system provides a direct indication of sag, which
can be the most relevant quantity for line operation.
The system can also be placed at the most critical
span in the line.
Its operation can be simply verified at any time by
comparing the measured height of the conductor to
the sagometers output. Comparative measurements
can readily be made by any simple means (surveying
equipment, range finder, tape measure, etc.).
Because of the so-called ruling span effects, measure-
ment of sag at one span provides the sag and tempera-
ture information for the entire line section.
The systems can be powered by solar cells and bat-
tery pack, or by standard AC power if available at the
site. Both types of systems have been installed in
North America.
On-board electronics store the information for later
retrieval and/or provide the information in real-time.
Information is transmitted back to a control center
via cell phone or spread-spectrum radio. Both types
of systems have been installed in North America.
The system can also provide other information, such
as ambient temperature, battery voltage, etc. Also, a
new device has been developed that works in con-
junction with the sagometer that can monitor load at
the site. Load needs to be known in real-time in order
to perform dynamic rating calculations.
The system is able to transmit a digitized image of the
line back to a control center for further scrutiny if
needed. This feature could be used when there
appears to be anomalous line behavior, such as icing
events or other serious physical damage to the line.
2.7.6 Dynamic Rating Calculations
There is a distinction to be made between the real-time
monitoring and dynamic rating of overhead lines. Real-
time monitoring is easier, but less useful in guiding oper-
ator actions than providing dynamic ratings.
For example, the conductor temperature, sag clearance,
or tension of an overhead line can be monitored with a
temperature monitor, a video sagometer or a load cell,
and the result reported to the system operator by a vari-
ety of communication methods. Such measurements can
be very useful during periods of high electrical loading
in guiding the operator as to the real-time state of the
line. Such measurements can be used to avoid load shed-
ding or load reduction that would be required by static
rating methods. The limitations of such real-time moni-
toring involve the operators needing to know the spe-
cific limits on temperature, sag, or tension for the line
section, and the operators needing to estimate how
large the electrical load can be in order to meet physical
limits.
Dynamic thermal ratings require certain calculations
based on the real-time sag, conductor temperature, or
tension monitor data, however, since such ratings may
be directly compared to electrical load to guide operator
actions even where the operator does not have detailed
knowledge of line design limits on temperature or sag-
tension. Dynamic thermal ratings are also useful prior
to high post-contingency loadings and may therefore be
utilized to maximize load flows and minimize the likeli-
hood of clearance or over-temperature occurrences dur-
ing emergencies.
Dynamic thermal ratings are calculated on the basis of
thermodynamic heat balance in the line conductors
(Douglass and Edris 1996, 1999; EPRI 1995). If real-
time tension or sag is monitored, the tension or sag
must be converted into an equivalent conductor temper-
ature in order to serve as the basis for dynamic rating
calculations. With both tension and sag monitoring sys-
tems, the conductor temperature is not directly mea-
sured. Therefore, the temperature of the conductor must
be inferred from other measurements. This process of
relating conductor temperature to sag or tension is
called line calibration.
Figure 2.7-4 illustrates how a line calibration is done
with a tension monitor (note that a very similar graph is
used for sag measurementsi.e., sag versus temperature
is plotted). The vertical axis is the line tension. The hor-
izontal axis is the estimated conductor temperature. In
this case, the conductor temperature is assumed equal to
2-80
that reported by a net radiation sensor, which is a
short length of aluminum rod, oriented in the direction
of the line section and painted to approximate the emis-
sivity of the line. This is a reasonable assumption during
those times when the line current is low and the line
conductor may be assumed to have nearly the same tem-
perature as the painted aluminum rod.
The symbols indicate 15-minute average tension and
solar temperature data taken for the line during periods
of low current. The curve shown in Figure 2.7-1 is taken
from a normal sag-tension calculation (done with
Alcoas SAG10 program), where the final unloaded ten-
sion at 15C was set equal to about 6350 lbs (28 kN) in
order to fit the tension field data. The curve has also
been expressed as a polynomial, as shown by the fourth-
order equation shown in the figure.
Given the tension-temperature (or sag-temperature)
equation for a line section, tension (or sag) measure-
ments can be converted to equivalent conductor temper-
ature and dynamic rating calculations performed.
Weather-Based Ratings
Instruments to measure wind speed, wind direction,
solar intensity, and air temperature are placed at the
approximate height of the transmission line conductor,
preferably in the transmission right-of-way. Weather
data from airports and other commercial stations is
likely to be inappropriate for real-time monitoring of
lines. As in most monitoring methods, the line current is
obtained from conventional current transformer mea-
surements at a nearby substation.
The conductor temperature (and the sag and tension of
the line) is predicted based on weather conditions, line
current, and conductor parameters. Conductor tempera-
ture is used to determine the position of the conductor in
light of the sag-tension line design data. Alternatively,
the conductor temperature is compared to the line design
maximum allowable conductor temperature, and it is
assumed that if the design temperature is reached, then
the safety limit is exceeded and there is risk to the public.
The highest conductor temperatures are obtained for
the lowest wind speeds, and those winds that are nearly
parallel to the line direction. Therefore, the wind ane-
mometer must be of high quality, and be able to mea-
sure wind speeds below 3 ft/sec (1m/sec). The propeller
type is more accurate than the cup type, but both are
subject to start-up error after stalling at low wind speed.
The best results are often obtained from the ultrasonic
type (see Figure 2.7-1).
The calculation of line ratings by weather monitoring
does not require measurement of the line current. This
method may therefore be used to supplement the other
monitor based dynamic rating methods.
This method may not cater for variation in parameters
that could affect the conductor temperature. Variation
in the value of the parameters can be caused by variabil-
ity of the terrain or by the sheltering of a line by trees or
buildings. In addition, wind speed and direction can dif-
fer from the point of measurement, (for example, an air-
port) to the actual line. To mitigate this, there may be a
need to install a number of weather stations along the
(long) lines; associated communication problems to
transmit the readings may occur, together with uncer-
tainty of the best location of weather stations.
Conductor Temperature-Based Ratings
The sensor is usually located at one position only. It is
known that temperature varies along the span as well as
between spans. To make a judgment based on this one
reading is risky, since the temperature of the conductor
can be very different from span to span, especially if the
line changes direction or terrain (sheltered or unshel-
tered spans). The cooling is approximately 40% in a line
section parallel to the wind compared to a section per-
pendicular to the wind.
Also, the temperature measured is the conductor sur-
face temperature, not the average conductor tempera-
ture (that affects sag).
Tension-Based Dynamic Ratings
Over the last 5 years, use of line tension monitors has
become widespread within the U.S. The first commer-
Figure 2.7-4 Example of line calibration from a previous
EPRI field test at PECO Energy.
2-81
cially available device, known as the CAT-1, is installed
at over 30 utilities. There are a number of reasons for the
popularity of these devices. The tension-measuring
device is a commercial load cell, which appears to be
very reliable and exhibits little drift with varying
weather and line load conditions. The device is mounted
on the grounded side of a dead-end insulator string and
thus is not subject to high electric fields. Line tension
monitors are normally installed with the line taken out
of service.
Sag-Based Dynamic Ratings
As an example of a typical sagometer installation, a
recent field study at TVA is noted in the following (see
Figure 2.7-5). The project involved one of the video sag-
ometers unique features: the transmission of real-time
images of the span being monitored. The development
of firmware and the base-station software needed to
implement this feature has been completed.
One of TVAs primary objectives was to use these sys-
tems in real timei.e., by communicating clearance
measurements directly to the control center and deter-
mining real-time dynamic ratings for these lines. TVAs
ultimate goals are to connect the sagometers and the
weather station to its SCADA system, which would
make the data available throughout its EMS system,
and to use the clearance and weather data, coupled with
modified DTCR software, to calculate real-time ratings
for the monitored lines that would be available to system
operators.
As such, these systems were set up and installed to func-
tion in real-time, and from the day the system went into
operation, data has been collected at the base station in
real-time. Spread-spectrum radios are used instead of
cell phones to communicate between the base-station
computer and the remote video sagometer sites. The
sagometers are all within a radius of 10 miles from the
base-station computer.
A custom-made user interface to display clearance data,
along with the available clearance margin, was written
for the base-station computer (see Figure 2.7-6). The
TVAs system operators routinely access the display
screen during contingency periods and monitor the avail-
able clearance margin, which blinks if it gets below 10%.
Based on sag and weather measurements, DTCR was
executed to compare actual ratings to the static rating.
Figure 2.7-7 shows an example of the results that can be
achieved. This is a 24-hour block of rating data. The
4-hour and 15-minute dynamic rating data were deter-
mined by DTCR operating in conjunction with real-
time data from a video sagometer. As can be seen, the
dynamic ratings are significantly greater than the static
Figure 2.7-5 The solar-powered Sagometer on a
lattice structure at TVA.
Figure 2.7-6 Real-time display for TVA video
sagometers.
Figure 2.7-7 A 24-hour block of rating data.
2-82
ratings. Note that, during the morning, the dynamic rat-
ings became very high due to precipitation. Results such
as these are typical of all DTCR/sagometer applications.
2.7.7 Field Test Results
EPRI sponsored a series of field tests of dynamic rating
and monitoring techniques at several utility sites. The
outcome of these field tests were significant, and EPRI
refined both its DTCR circuit rating software and
learned certain fundamental facts about how weather
affects the rating of overhead transmission lines and
how dynamic rating methods are best applied.
1. Dynamic thermal ratings for overhead lines may be
calculated based on either real-time weather, or real-
time sag or tension data in conjunction with real-time
weather. For weather-based ratings, the wind angle
should be assumed fixed and near parallel to the line
direction to account for directional variation along
the line section.
2. In rating longer lines with multiple ruling span sec-
tions, it is likely that the line rating (dynamic or
static) decreases with line length and that dynamic
rating of lines requires multiple monitoring locations,
and the minimum number of monitors required must
be based on field measurements.
3. Sag and tension monitors work well in lines having
high current density (greater than approximately 1
amp/mm
2
) where they generally yield more accurate
ratings than single-point weather monitors. However,
in lines with low current density (less than 0.5
amps/mm
2
), weather-based dynamic ratings are more
accurate than those based on sag-tension monitors
4. Sag and tension monitoring allows one to directly
make measurements at high temperatures. Weather
monitoring does not.
For example, the data obtained in these field tests show
that there is a great deal of fluctuation in both wind
speed and direction along most line routes, particularly
during periods of low wind activity. Figure 2.7-8 shows
15-minute average wind speeds at locations only 1.5 km
apart along a line route in Philadelphia.
The field tests confirm that not only the wind speed but
also the wind direction varies along the line. This raises
questions about the usefulness and accuracy of basing
dynamic thermal line ratings on weather data from a
single location within a line section. It would seem to
imply that multiple weather monitor locations might be
required, especially for long line sections.
Comparison of Weather Monitor and Tension/Sag
Monitor-Based Dynamic Line Ratings
The main advantages of using weather-based line rat-
ings are two-fold:
The rating calculation is independent of the line
current.
The monitoring equipment is modest in cost.
Weather data may also be used to dynamically rate
nearby substation equipment.
The disadvantages are that the anemometers are quite
fragile and prone to measurement error unless calibrated
frequently and, being a measurement of weather condi-
tions at a single location, may not truly represent the
weather along the entire line section, especially the wind,
which can be extremely variable from place to place.
Field experiments conducted by Chisholm at Ontario
Hydro (EPRI 1995) indicated considerable success in
estimating average conductor temperature. The instru-
ments were placed over a five span ruling span section,
and the data was based on real-time line current and on
weather monitors placed at a distance from the line, and
it was assumed that the wind angle was fixed at an angle
of 20 to 30.
Comparisons of weather-based and tension-based line
ratings at three of the four field-test sites indicates that
there is good agreement between minimum values of
weather-based and line tension-based ratings when
using a fixed wind angle of 22 relative to the conductor
axis. This is illustrated in Figure 2.7-9.
Figure 2.7-8 Wind speed (15-min average) at two
locations 1.5 km apart along a 230-kV line in the eastern
U.S.
2-83
Based on such observations at each field test site, it
appears that weather-based ratings based on a fixed line
angle of the order of 20 are conservative under nearly
all conditions and that such weather-based ratings
can be used as a means of warning the operator during
periods of low rating conditions. When combined
with the use of less conservative fixed ratings for plan-
ning and low load operation, this weather-based
dynamic rating method would be very cost-effective.
Rating Variation in Adjacent Line Sections
Figure 2.7-10 shows the variation in tension-based line
rating with time for the 230-kV SRP line. Four ruling
span sections are monitored. I2 is E-W, and the other
three are oriented nearly N-S. Note that the E-W span
generally has the lowest rating, but that this is not true for
certain periods such as the three hours starting at 6 am.
Clearly, if the entire line were rated on the basis of a
monitor in only one section, the rating would be too
high some percentage of the time and therefore not con-
servative. Multiple monitoring locations are required to
correctly calculate the real-time line rating; however, it
appears that there is good agreement for the three line
sections oriented in the same direction (N-S).
It appears that the number of monitoring locations
(either weather or tension) required to calculate the real-
time line rating correctly must be empirically deter-
mined for each location.
Daily Line Rating Variation
Certain circuits experience a fairly regular daily load
variation, others do not. For those circuits that experi-
ence a repeatable load cycle, the daily variation in line
rating may or may not be coordinated with the load.
Consider, for example, the rating variations shown in
Figures 2.7-11 and 2.7-12.
Figure 2.7-9 Comparison of weather-based and tension-
based cumulative rating distributions.
Figure 2.7-10 Comparison of tension-based rating
estimates for four separate line sections.
Figure 2.7-11 Weather-based normal rating for SDGE 138-
kV line as a function of time of day for September 1997.
Figure 2.7-12 Weather-based normal rating for SDGE
138-kV line as a function of time of day for December
1997.
2-84
2.7.8 Summary
found in IEEE 738-1993. Given a maximum allowable
conductor temperature, the corresponding maximum
allowable current (the static thermal rating) is deter-
mined for worst-case weather conditions. In most
cases, the maximum temperature is limited in order to
avoid excessive conductor sag or, in some cases, a loss in
conductor strength.
By using real-time monitors combined with communi-
cations, these measurements can be presented to the sys-
tem operator, and/or can be used to perform dynamic
rating computations whose results can also be provided.
Results of several projects have demonstrated that it is
possible both to increase the line rating under most con-
ditions and to avoid electrical clearance violations under
severe load and weather situations.
The distinction between real-time line monitoring and
dynamic ratings is explained and the various monitoring
methods noted and analyzed.
2.8 CASE STUDIES
2.8.1 Introduction
This section includes a general discussion of the major
factors that need to be quantified in order to select an
economic and reliable uprating method for overhead
lines. No single method is most economic in all cases,
nor are all the important factors economic. Nonethe-
less, the selection of an appropriate uprating method is
never made without economic justification. Similarly,
no method can be identified as the best method since
the uprating decision depends on predictions of what
will be, and involves certain irreducible uncertainties. In
spite of these difficulties, the uprating of transmission
lines, as described in the following, should be a logical
process, flexible enough to use in most cases, and power-
ful enough to yield valuable insight into line behavior
and operation.
2.8.2 Selecting a Line Uprating Method
As described in preceding sections of this chapter, there
are many ways to increase the thermal rating of an exist-
ing line. As any experienced line designer knows, there
are a lot of different ways to accomplish the same goal,
and the cheapest way may not be the most sensible way
to provide a reliable transmission system. Engineering
judgment is often required in selecting the most appro-
priate method of uprating existing lines. Therefore, the
goal here is to identify the major factors that should be
considered in line uprating, demonstrate a general
method of viewing these factors simultaneously for a
number of design cases, and perform a reasonably
detailed application of such methods for a particular
line uprating example.
Basic Observations
Certain observations about line uprating appear to be
nearly universally accepted by utility designers:
Public safety is most crucial, and litigation is to be
avoided if at all possible. If litigation or public injury
is a possible result, the chosen uprating approach is
unacceptable.
Frequent load shedding is painful, expensive, and
makes distributed generation (DG) more attractive.
A marginal line uprating method that results in the
frequent need for operator intervention to drop cus-
tomer loads, even interruptible ones, is a poor choice
in the long run.
Avoiding the use of vibration dampers and/or armor
rod by keeping everyday conductor tension well
below NESC Code or CIGRE safe limits is an expen-
sive way to avoid conductor fatigue. However, a
vibration assessment by damper manufacturers
should precede any uprating decision that involves
the use of high everyday conductor tensions.
If the maximum design temperature of an older exist-
ing line is to be increased to more than 100
o
C, com-
pression splices should be replaced or shunted, and
loss of tensile strength in the aluminum strands must
be carefully assessed.
Sag clearance buffers are required in any transmis-
sion line because of uncertainties in weather and
design. The buffer for an existing line may be some-
what less than for the design of a new line (since
structure placement uncertainties do not exist, and
conductor elongation uncertainties are less than for a
new line), but reducing necessary sag buffers to zero
is not a good idea.
Before considering any uprating method, it must be
determined if the existing structures are in reasonably
good shape, having load capacities at least equal to
the original design assumptions. If this is not true,
then none of the uprating options discussed is appro-
priate. The transmission owner should consider
replacing the existing line with a new facility that is
both safe and reliable.
Before considering any uprating method that does
not involve reconductoring the line, the present con-
dition of the conductors must be evaluated. When the
conductor is not replaced, the increased rating will
lead to higher conductor operating temperatures, and
2-85
such methods can only be considered practical when
the existing conductors and compression splices are
in very good condition or can be made so.
Critical but Qualitative Issues
In evaluating uprating options for an overhead trans-
mission line, there are many issues to be considered.
Some issues are hard to quantify. For example, the reuse
of existing structures carries with it an uncertain, and to
a certain degree unknowable, risk that there may be
undiscovered structural flaws. This risk may be perfectly
acceptable in most situations, but not acceptable in lines
whose failure may have enormous consequences. These
factors may be hard to put a price on, but their consid-
eration is essential in deciding whether, and how, to
uprate an existing line. Some risks that should be con-
sidered are:
Corporate impact of negative publicity from injuries,
death, property damage from downed conductors, or
instances of excessive sag.
Loss of residential or commercial customers due to
service interruptions.
Intervention by regulatory bodies in response to ser-
vice interruptions.
Criticality of line to overall system reliability. To what
extent might an outage affect a broader regional ser-
vice interruption.
Certainty of projected electrical overload.
Certainty of financial return on capital investment.
All of these issues exist and are part of the transmission
owners uprating decision, but none is easy to quantify.
A consideration of these issues is more likely to produce
a sense of comfortable risk rather than a detailed
decision on uprating method.
Financial Consequence of Electrical Line Losses
A very important and peculiar issue in line uprating
concerns the matter of line losses over the life of the line.
As is demonstrated in the final detailed uprating exam-
ple in this section, when the value of line losses is con-
sidered in the uprating problem, the line designer may
select a significantly different uprating method than if
such losses are ignored.
In those cases where the rating of a moderately short
transmission line is being increased in order to avoid
load shedding during relatively infrequent post-contin-
gency loadings, it is unlikely that electrical losses should
be a factor in uprating. In those cases where the rating
of a reasonably long line is being increased in order to
allow increased daily load peaks throughout an entire
season, it is likely that electrical losses should be consid-
ered and that the designer should seek an uprating
method that reduces them.
Surely, occasional operation of lines at temperatures
approaching 200
o
C may be smart engineering, but rou-
tine operation at high temperature is not.
2.8.3 Preliminary Selection of Uprating Methods
The best solution at one utility may differ from that at
another because the line uprating decision involves a
good deal of engineering judgment based on experience.
Nonetheless, in every case, the selection of an appropri-
ate uprating method depends heavily on the physical,
electrical, and thermal characteristics of the existing
line, and on the exact nature of the need for a higher
capacity line. It is possible to identify certain existing
line parameters and system analysis results that largely
indicate the best uprating solutions.
Since there are many factors that influence the selection
of line uprating methods, it is helpful to list the most
essential ones, and to develop a table summarizing those
most likely to determine, or at least strongly influence,
the line uprating method. The resulting Uprating
Analysis Table is intended for use in the preliminary
stages of developing an appropriate uprating solution. It
is an aid to focusing the engineering inquiry on the most
productive uprating methods.
Given the large number of factors that influence the
uprating of an existing overhead line, it is crucial to
identify, and then quantify, the most important. These
factors in line uprating (with the most important shown
italicized and/or in bold) include the following:
System Analysis
The impetus for line uprating comes as a result power
system analysis. Present electrical loads are projected
into the future, and the impact of various component
outages (i.e., contingencies) on the electrical loading of
the existing line is determined. Specific probabilities are
seldom associated with post-contingency loadings, and
even the prediction of normal loads is often uncertain,
particularly with the advent of open access to com-
mercial power generators.
Nonetheless, even with such uncertainties, the system
planner must determine:
Criticality of the line to overall system reliability
(marginal or absolutely critical).
Certainty of projected electrical overload (very cer-
tain or not certain).
2-86
Frequency of high electrical loads (e.g., daily sea-
sonal peak loading or rare post-contingency emer-
gency loads).
Magnitude of the electrical overload in 2 years (e.g.,
< 10% or > 50%).
Structural Review
Transmission structures and the conductors they sup-
port are the major cost components of any overhead
line. (In comparison, the cost of insulators and hard-
ware are almost always minor.) Therefore, the design
limits and actual physical condition of the existing
structures and foundations are a major factor in select-
ing a line uprating method. It is fair to say that if the
lines structures are in poor condition, no uprating alter-
native makes much sense, reliability-wise or financially.
On the other hand, given their initial cost, if the struc-
tures are largely in their as designed condition, any
uprating method that does not require their extensive
modification is likely to be significantly cheaper than a
new line with the same capability.
Given an existing line with structures in reasonably
good condition, there is a large financial incentive to
uprate the line without needing to make major changes
in the structure geometry or load capability. As part of
any preliminary uprating evaluation, for a line with
structures in good condition, it is necessary to deter-
mine the limits of low-cost structure modifications.
For example, in uprating an existing line, how much can
the present structure attachment points be raised with-
out incurring significant cost (where significant cost is
that which exceeds 10% of the cost of new structures).
Similarly, in considering the possibility of retensioning
the line or reconductoring it, how much can the original
maximum conductor tension be increased without
incurring costs that exceed 10% of the cost of new struc-
tures. In performing such analysis, the following should
be considered:
Physical condition of the existing structures (as
designed or 10% or more in need of replacement).
Maximum increase in transverse load beyond which
the cost of tangent structure reinforcement exceeds
10% of the cost of new structures.
Maximum typical increase in tangent structure con-
ductor attachment height beyond which the cost
exceeds 10% of the cost of new structures.
Number of strain structures per mile (km) (0.05 or
1.0).
Maximum increase in maximum conductor tension
load beyond which the cost of strain (and possibly
tangent) structure reinforcement or replacement
exceeds 10% of the cost of new structures.
Evaluation of broken wire loads and other conditions
not considered in the original design.
Conductor Review
Other than transmission structures, the conductors they
support are the major cost component of any overhead
line. Therefore, the reuse of the existing conductors in a
line uprating is very economically attractive. The option
of reusing existing conductors hinges on their physical
condition, which may not be easy to determine. The
usual signs of conductor deterioration include corrosion
of the steel core wires of ACSR, corrosion and fatigue
within compression splices, and the fatigue of aluminum
wires at, or near, the mouth of support clamps due to
Aeolian vibration. None of these signs is easily deter-
mined in the field, yet there is little point in considering
the reuse of existing conductor without establishing that
it is in good condition.
Assuming that the existing conductor is in reasonably
good condition, the following factors help in selecting
an appropriate uprating method:
Physical condition of the existing conductors (as
purchased or remaining life < 10 years).
Unloaded final everyday tension of existing conduc-
tors (15% or 25% RBS).
Existing conductor type (30/7 ACSR or 37 strand
AAC).
Existing line maximum temperature (49C or 125C).
Excess clearances at existing line maximum tempera-
ture (most spans < 1ft (3 m) or most > 5 ft (1.5 m)).
Change in sag with increased temperature
(0.2 ft/10C to 2 ft/10C [0.06 m/C to 0.6 mC]).
Probability of assumed weather conditions.
% change in line rating per 10C change in line design
temperature (50% to 5%).
Preliminary Uprating Analysis Tables
To summarize the various key uprating parameters, the
following simple Uprating Analysis Table has been
developed. By defining the 11 parameters listed in the
table, one can better understand which uprating meth-
ods are likely to meet system needs at minimum cost.
Table 2.8-1 is an example.
In this example, it is clear that:
The predicted overload magnitude is probably too
large to be accommodated by the use of dynamic rat-
ing methods, but not so large as to necessarily require
reconductoring.
2-87
The predicted high electrical loading will be infre-
quent, and the resultant cost of electrical losses will
be low even if the existing conductor is reused.
Modifying the existing conductor tension or replac-
ing it with a larger conductor will require significant
expensive structure modifications, but raising the
conductor attachment points appears to be rather
easy (low cost).
The present line design temperature is moderate and
could be increased without causing annealing by
exceeding 100C.
This preliminary review seems to indicate that raising
the existing conductor attachment height, while con-
tinuing to use the existing conductor and installed ten-
sion, may be an economic uprating method.
Uprating analysis tables were developed for each of the
following uprating case studies in order to simply sum-
marize those aspects of the existing line that are most
important to the selection of uprating method.
2.8.4 Uprating Test CasesPreliminary Uprating
Study
Clearly, the uprating method applicable to a particular
line depends on a number of different parameters that
must be defined as part of the uprating process. However,
certain aspects of the line design suggest certain uprating
methods, or suggest avoiding certain approaches. This
section includes a collection of typical candidate lines
with appropriate line uprating methods identified
(including references to the section describing the
method) (see Table 2.8-2).
Table 2.8-1 Uprating Analysis Table
Table 2.8-2 List of Case Studies Considered
Case Study #
and Section
Reference Structure Conductor System
Promising Uprating
Method(s)
1 Section 5
115-kV single ckt, wood
pole H-Frame, I-string
insulators
26/7 397.5 kcmil
(203 mm
2
), clearance
limited at 75C
Normal load in summer
reaches 50% of rating,
fairly frequent emer-
gency loads reach 110%
Raise attachment points
by raising crossarm or
using floating dead-end
concept
2 Section 3 & 4
69-kV single ckt, single
wood pole with cross-
arm & post insulators,
guyed pole angle and
dead-ends
#2AWG copper with
original splices, > 3 ft,
(1 m) excess clear-
ance at 75C max
Normal load is 20%
of thermal, rare
emergency load to
120% rating
Inspect conductor and
connectors, raise max
temp of existing conduc-
tor to 90C
3 Section 5
230-kV double ckt, steel
lattice self-supp, I string
insul
18/1 477 kcmil
(243 mm
2
), clearance
limited at 75C
Normal load in summer
reaches 50% of rating,
fairly frequent emer-
gency loads reach 110%
Retension existing
ACSR conductor to allow
operation to 100C
4 Section 5
lattice self-supp, I string
insul
30/7 250 mm
2

(490 kcmil) ACSR
Normal load profile
to exceed 100% rating.
Load cycle is predict-
able.
Statistical comparison of
line rating with other
lines in region of low
wind and high air temp
allows increased rating
5 Section 7
lattice, light loading,
seacoast area
26/7 636 kcmil
(324 mm
2
) ACSR
Install tension or sag
monitors, and use
dynamic rating methods
2-88
To be useful, a detailed description of each case study is
included to make at least some initial decisions about
what approach to take, but a detailed plan profile or sag
survey data is not included. Similarly, the lines include
voltage ranges of 69 kV to 345 kVthe most common
lines that need uprating.
Case #1 115-kV Single-Circuit Wood Pole H-Frame
26/7 ACSR Conductors
This 115-kV line has a thermal rating of 114 MVA (see
Figure 2.8-1). During the preceding summer period, the
load on this line reached 110 MVA during the hottest
day, and system planners project the need for a 25%
increase in capacity (to 140 MVA). The increasing load
is likely to develop slowly over the next 10 years. The
line experiences relatively high daily loads during the
summer peak period. Post-contingency emergency loads
are not a problem.
The line consists of older wood pole H-frame structures
with 600 ft (180 m) spans. The structures are easily rein-
forced. Transverse load limits can be increased with a
minimum of additional bracing. Vertical loading can be
increased by hardware replacement, and dead-ends can
be strengthened by use of additional guying. The line is
in an NESC light loading area.
The existing conductor is 26/7, 397.5 kcmil (203 mm
2
)
Ibis ACSR strung to a final unloaded tension of 20% of
its rated strength at 32
o
F. The final everyday sag per 600
ft (180 m) span is about 9 ft (3 m) with an existing buffer
(excess electrical clearance) that is typically 3 to 6 ft (1
to 2 m) at the present line design temperature of 75
o
C.
Occasional fatigue breaks have been observed near
clamps, and about half of the original galvanizing is left
on the core wires.
The summer rating conditions (the line load peaks in
the summer) presently in use are an air temperature of
35
o
C and a crosswind speed of 3 ft/sec (0.91 m/sec).
Uprating Analysis Case #1
The uprating alternatives for test case #1 (see Table
2.8-3) include the following:
A. Revision of existing rating conditions based on the
use of monitors (dynamic ratings - Section 7) or
through a probabilistic analysis of the line rating
(probabilistic ratings - Section 5). Given the required
25% increase in line rating and the nonconservative
weather assumptions presently in use, it is unlikely
that these methods will yield an increase that large.
Also, given the high daily load cycle, electrical losses
will be significant, and neither of these methods
reduces losses.
B. While the structures appear to be in good condition,
there is evidence that the existing Ibis conductor has
sustained some damage from vibration and that its
steel core strands are certainly not in like new con-
dition. Given the relatively low cost of structure rein-
forcement, one might consider retensioning the
existing conductor (Section 5) and increasing the line
design temperature to 100
o
C. Ibis ACSR at 100
o
C
has a thermal rating of 145 MVA. Given the high
normal line loads and the marginal conductor condi-
tion, however, this may not be advisable.
C. If the capital is available, the line might be reconduc-
tored (Section 6) with a trapezoidal wire conductor
Figure 2.8-1 Case #1115-kV line.
Table 2.8-3 Uprating Analysis for Case #1
2-89
such as Dove/TW, which has 40% lower resistance
(and therefore lower losses), about the same sag as
Ibis at 80C, and a thermal rating 25% higher than
the present line. The transverse structure loads will be
10% higher with Dove/TW, but the maximum tension
loads will be nearly 40% higher, and dead-end struc-
tures will require extensive reinforcement.
D. An alternative uprating solution may involve the use
of a high-temperature, low-sag conductor such as
ACSS or ACCR, though the relatively modest
increase in line rating (25%) doesnt require it. ACSS
is particularly attractive in applications where the
existing structures cannot easily, or economically, be
reinforced. The structures in this case are easily and
inexpensively reinforced.
Case #2 69-kV Single-Circuit Wood-Pole Copper
Conductors
Typical of the oldest lines in many systems, the desired
increase in thermal line rating results from an attempt to
deal with a relatively rare single contingency that would
persist for up to 24 hours. The addition of a new 345-kV
line section will remove the contingency within 5 years.
(See Figure 2.8-2.)
Line DescriptionCase #2
69-kV system voltage, 4 suspension insulator bells.
No dampers, bolted dead-ends, no armor rod.
7 strand, #4/0 AWG Copper conductor, original
splices.
Rating conditions2 ft /sec (0.6 m/sec) perpendicu-
lar, 40C air, sun, 60C continuous/75C emergency.
Mild corrosion area, no broken strands found at
clamp locations.
Normal daily peak annual loading is only 30% of
normal continuous rating. System analysis by plan-
ning needs a 50% increase in emergency rating (post-
contingency loading).
Span lengths range from 150 to 300 ft (45 to 90 m).
Ruling span is 250 ft (76 m). Excess electrical clear-
ance at 75C ranges from 2 to 10 ft (0.6 to 3 m). Aver-
age clearance is 4 ft (1.2 m). The line length is 10 km
(6 miles) with 15 line sections going in a predomi-
nantly east-west direction.
NESC Medium loading area (0.25 in. (0.6 cm) ice
with 4 psf wind). Everyday tension at 15 C equal to
12% RBS (Rated Breaking Strength) final.
Single wood pole structures with zig-zag cross-arms
and suspension insulators.
Built in 1935, 20% damaged poles (rot) replaced in
1962. Another 15% replaced in 1988. No extensive
structural failures known. No broken conductors.
Uprating AnalysisCase #2
The thermal rating of the existing line is 47 MVA
(395 A). The system analysis indicates a need for a 50%
higher emergency rating of 70 MVA (590 A). The nor-
mal rating of the line remains at 33 MVA (280 A). (See
Table 2.8-4.)
Figure 2.8-2 Photo of Case #2 69-kV line.
2-90
The structures appear to be in reasonably good shape,
yet they are quite old, and the possibility of reconduc-
toring such old structures to meet a relatively rare emer-
gency load event does not seem economic. The existing
structures cannot be modified to raise the attachment
points, nor can they be easily reinforced. Reconductor-
ing the line would, therefore, be difficult since the con-
ductor maximum tension and diameter could not be
increased at all, and the sag could not be increased.
The predicted occurrence of high, post-contingency
electrical loading is relatively rare, being the result of an
unusual contingency. The copper phase conductors
could be operated at temperatures up to 100
o
C for short
periods of time without any significant annealing, but
the existing original splices would have to be replaced to
ensure reliable operation. This is also true of existing
hardware. The original conductor is installed at a rela-
tively low tension in short spans and exhibits no evi-
dence of fatigue damage. There is insufficient clearance
at 75
o
C to allow operation of the existing line above that
design temperature. Retensioning the existing conductor
may be a low-cost uprating method to meet clearance
requirements at a temperature above 75
o
C.
The present rating weather assumptions are reasonably
conservative. This suggests the possibility of using
dynamic uprating methods, but as discussed in Section
2.7, such methods typically produce a usable increase of
only 10 to 20% in the line rating.
Assuming that samples taken from the existing copper
conductor indicate that it retains its original tensile
strength and that a 10% reduction in tensile strength
over the remaining life of the conductor is acceptable,
the annealing curves in Section 2.2 indicate that the
existing #4/0 copper conductor can be operated at tem-
peratures in excess of 100
o
C for brief time periods with
the impact on tensile strength, as shown in Table 2.8-5.
With the (emergency) design temperature of the line
increased to 115
o
C, the emergency rating with the exist-
ing #4/0 copper conductor would be increased to 70
MVA, meeting the required emergency rating. At this
temperature, the tensile strength of the #4/0 copper con-
ductor would drop by 10% in approximately 100 hours.
If the remaining life of the line is 20 years, and contin-
gencies are limited to 24 hours when they occur every 5
years, then this is acceptable if the electrical clearance
can be maintained.
At the present emergency design temperature of 75
o
C,
the sag is 7.6 ft (2.3 m). If the unloaded sag of the con-
ductor is decreased from 5.1 ft to 3.4 ft. (1.5 m to 1 m)
by increasing the everyday tension at 60F by about 500
lbs (2232 N), then the sag at 115
o
C will be 7.4 ft (2.2 m)
and the clearances will be met. The increase in installed
tension will cause an increase in the maximum conduc-
tor tension from 1570 lbs (7009 N) to 2120 lbs (9464 N)
and will require reinforcement or replacement of dead-
ends. The transverse loads on tangent structures are
unchanged.
Case #3 230-kV, Double-Circuit Steel Lattice, 54/7
795 kcmil (405 mm
2
) Condor ACSR
Typical of the moderately aged lines in many systems,
this double-circuit 230-kV line was built in the 1960s
using steel lattice self-supporting structures. The emer-
gency thermal rating of the Condor ACSR conductor is
1170 A (465 MVA per circuit). The system analysis indi-
cates that this short line needs to carry up to 800 MVA
as a result of certain severe contingencies. Although the
contingency is likely to occur only every few years, if it
does occur at all, it is likely to persist for several weeks.
230-kV system voltage, double-circuit, 12 suspension
insulator bells, aluminum clamps.
Dampers on exposed sections, compression dead-
ends, armor rod used at all clamps.
Existing line has 54/7 strands, 795 kcmil (405 mm
2
)
ACSR conductor, the condition of full tension splices
is uncertain.
Rating conditions3ft/sec (1 m/sec) perpendicular,
30 C air, sun, 75C continuous/100C emergency.
Mild corrosion area, no broken strands found in rou-
tine climbing inspection.
Normal daily peak annual loading is 40% of normal
continuous rating. System analysis indicates that the
emergency rating needs to be 800 MVA rather than
475 MVA, and that the increased post-contingency
loading would be likely to persist for several days at
the infrequent times when it occurs.
Span lengths range from 800 to 1100 ft (240 to 330
m). Ruling span is 1000 ft (305 m). Electrical clear-
ance at 100C ranges from 1 to 3ft. (0.3 to 3 m). Aver-
age clearance at 100C is 2 ft (0.6 m). The line length
is 40 km (24 miles) with 20 line sections going in a
predominantly north-south direction.
Table 2.8-5 Loss of Tensile Strength
Temperature
Loss in tensile
strength after 100
hours
Hours for a 10% Loss
in Tensile Strength
100 3% 600
125 20% 40
2-91
NESC Heavy loading area (0.5 in. [1.3 cm] ice with 4
psf wind). Final everyday unloaded tension at 15 C
equal to 18% RBS (Rated Breaking Strength) final.
Steel lattice, self-supporting structures. Galvanizing
is in good shape. Concrete footing inspection indi-
cates they are in near-original condition.
Built in 1963, structures have been inspected by heli-
copter. No major line failures have occurred. One line
section failure in 1972 due to a crane accident. Cros-
sarm failure and conductor damaged.
Given the nonconservative nature of the rating weather
assumptions and the need for a large increase in the line
rating (465 to 800 MVA), neither the dynamic rating
methods in Section 2.7 nor the probabilistic methods of
Section 2.5 are applicable. (See Figure 2.8-3 and Table
2.8-6.)
Similarly, given the small increase in rating (7%) per
10
o
C increase in the line design temperature, even if the
line design temperature of the existing line (with Con-
dor) were increased to 200
o
C, the corresponding line
rating (690 MVA) would not be sufficient to meet the
system requirements (800 MVA).
This reduces our options to reconductoring the line (see
Section 2.6) with a conductor having less resistance than
Condor and capable of operating at 200
o
C for an
extended period of several days. In addition, the
replacement conductor will need to sag (at 200C) no
more than the original Condor ACSR did at 100
o
C, yet
the maximum conductor tension cannot exceed that of
Condor by more than 20%. As can be seen from the fol-
lowing sag-tension calculations for Condor in the origi-
nal line design (Table 2.8-7), the final sag is 37 ft (11 m)
at 100
o
C, and the maximum tension is a little over
10,000 lbs (44.6 kN).
As noted in the uprating analysis table, the clearance
limits of the existing line are tight (i.e., excess clearance
of Condor at 100
o
C is only 1 to 3 ft (0.3 to 1 m). There-
fore the replacement conductor cannot have more than
37 ft (11 m) sag at 200
o
C.
Also, the maximum tension load should not exceed that
of the original Condor conductor by more than 20% or
the costs of reinforcement will become prohibitive.
One possible solution involves reconductoring with
ACSS/TW. For example, 1033.5 kcmil (527 mm
2
),
Curlew/ACSS/TW would yield a rating of 803 MVA at a
line design temperature of 200
o
C as a result of its
greater aluminum cross section and its ability to operate
at 200
o
C for an extended period of time. The sag-tension
calculations (of Table 2.8-8) indicate that it is also capa-
ble of meeting the 37 ft (11 m) sag constraint at 200
o
C
and the 20% higher maximum tension constraint. Figure 2.8-3 Photo of structure for Case #3.
2-92
Other ACSS conductor designs may also be reviewed to
see if another design can also meet these constraints, but
with in lower capital cost.
Case #4Double-Circuit 132-kV Line
The existing 132-kV double-circuit line, according to
projected load growth, will soon reach the thermal limit
under normal operating conditions. The line is situated
in an area with mild climate and strong winds. The
phase conductor is Bear ACSR (30/7 250 mm
2
)a
British conductor similar to Hen ACSR.
The prediction of increased load is somewhat uncertain,
and in order to conserve capital, uprating methods that
require a minimum level of capital investment are
strongly preferred by management. As shown in Table
2.8-9, two essential features of the existing line that
make it a candidate for probabilistic uprating are that
the present rating calculations are based on rather con-
servative weather conditions (40C air temperature,
2 ft/sec [0.6 m/sec] wind, full sun), and the line design
temperature of the existing line is only 60
o
C. This makes
the rating of the line quite sensitive to changes in air
temperature (> 30% per 10C).
Probabilistic uprating methods require little or no capi-
tal investment since the line is neither physically modi-
fied nor monitored. The modified exceedance method
(described in Section 5) considers the electrical load pro-
file of the line. The conductor temperature (and thus the
ground clearance) is calculated over an entire season
using weather data derived through monitoring along
the line. The load profile is normalized based on the
profile at the maximum load. This is assumed to be the
likely profile at the time when the thermal rating is likely
to be met.
The present ratings for the existing line are based on the
assumption of a flat load profile using weather data
from a location some distance away from this line, in a
Table 2.8-7 Sag-tension Calculations for Case #3
Table 2.8-8 Sag-tension Calculations for Replacement Conductor of
Case #3
2-93
semi-arid region. While it isnt possible to determine the
absolute probability of ground clearance for the line, the
relative probability of ground clearance for this line
could be calculated using more appropriate weather.
When these relative probabilistic methods are applied, it
is determined that the thermal capacity of the existing
line can be increased from 529 to 863 A for normal, and
715 to 1240 A for emergency ratings. The risk is the
same as long as the line integrity from a physical view-
point is not jeopardized. It is necessary, therefore, to
assess the temperatures that joints are likely to reach.
This method of uprating must be carefully applied.
The rating applies only to a particular line. All joints
in the line must be tested via the resistance method
for integrity. All joints that are the same or higher
resistance than the conductor need to be replaced.
If the resistance method is found to be too costly,
each joint should have a wrap tie placed around it to
prevent the conductor falling in case of failure. The
joints then need to be regularly checked by infrared
camera. All joints at the same, or hotter, temperature
than the conductor need to be replaced.
Given the managerial decision to seek the lowest cost
uprating method, probabilistic methods of uprating are
very attractive. The rating of the line can be increased
sufficiently such that a major capital investment project
might be postponed for several years. It may not even be
necessary to increase the tension in the conductor or
raise towers. It must be noted, however, that such meth-
ods must be used with care. Results are not generic and
depend on the weather conditions at a specific area and
the load profile of a specific line. There are several spe-
cific cautions to be observed:
1. The weather data used in the calculation of conduc-
tor must be appropriate to line ratings. Data from an
airport weather station 20 miles (32 km) from one
end of the line may not be useful for this.
2. The conclusion that the line can be uprated because it
appears to be more conservatively rated than another
line at another location depends on the assumption
that the rating of the other line is safe.
3. The inclusion of line current variation over a typical
day means that the resulting probability distribution
of conductor temperatures may no longer be valid if
the lines daily load variation changes due to changes
in the system configuration.
4. Finally, there is no absolute certainty that the result-
ing line rating is safe. The uprating decision is based
on a comparison to another line, not on the establish-
ment of an absolute clearance assurance probability.
Case #5169-kV double-circuit, steel lattice, medium
loading area, 26/7 636 kcmil (324 mm
2
) ACSR
169-kV system voltage, double-circuit, 10 suspension
insulator bells, aluminum clamps.
Dampers on exposed sections, compression dead-
ends, armor rod used at all clamps.
Existing line has 26/7 strands, 636 kcmil (324 mm
2
)
Grosbeak ACSR conductor. The condition of full
tension splices is excellent.
Rating conditions2ft/sec (0.6 m/sec) perpendicu-
lar, 40C air, sun, 75C continuous/90C emergency.
Mild corrosion area, no broken strands found in rou-
tine climbing inspections.
Normal daily peak annual loading is 30% of its
present continuous rating. System analysis indicates
that the emergency rating needs to be increased to
310 MVA rather than the existing lines 293 MVA.
The increased post-contingency loading would be
2-94
likely to persist for no more than an hour at the infre-
quent times when it occurs.
Despite the modest predicted overloads, system oper-
ations is particularly interested in finding a way to
uprate this line without taking it out of service for
more than short time periods. Extensive reconstruc-
tion is unacceptable.
Span lengths range from 600 to 800 ft (180 to 240 m).
Ruling span is 750 ft (225 m) Electrical clearance at
90C ranges from 3 to 5 ft (1 to 1.5 m). Average clear-
ance at 90C is 4 ft (1.2 m). The line length is 15 km
(9 miles) with 10 line sections going in a predomi-
nantly east-west direction.
NESC Heavy loading area (0.5 in. (1.3 cm) ice with 4
psf wind). Final everyday unloaded tension at 15C
equal to 18% RBS (Rated Breaking Strength) final.
Steel lattice, self-supporting structures. Galvanizing
is in good shape. Concrete footing inspection indi-
cates they are in near-original condition.
Built in 1974, structures have been inspected by helicop-
ter. No major line failures have occurred.
Given the conservative weather assumptions (see Table
2.8-10) used in calculating the rating of the existing line,
the uncertainty of the predicted overload, and the mod-
est magnitude by which the post-contingency load
exceeds the present rating, the use of dynamic rating
methods appears to be worth considering.
Since system operations does not want this line taken
out of service even to uprate it, noncontact or hot stick
mounting of line monitors is attractive. Video sagome-
ters could be installed at each end of the line, each near
a substation where communications to the utility con-
trol center is simplest.
The following photograph (Figure 2.8-4) illustrates the
installation of a sag monitor on a lattice structure.
The primary advantage of this approach is that it meets
the need for a modest increase in line rating without
requiring a large capital investment. Also, if the pre-
dicted increase in post-contingency load does not occur,
the monitor and communications equipment can be
reused in another suitable installation.
The primary concerns center around the need to edu-
cate system operations personnel in dealing with line
ratings that are not constant. This can be challenging
and, if this is the first application of dynamic rating
methods in the utility, the necessary investment in engi-
neering time and operator education should not be
underestimated.
Figure 2.8-4 A Video sagometer mounted on a lattice
structure.
2-95
2.8.5 Economic Comparison of Line Uprating
Alternatives
This chapter discusses a wide variety of techniques that
allow one to increase the capacity of existing transmis-
sion lines. In the preceding part of this section, test cases
and methods are presented for performing a preliminary
uprating assessment, the goal of this being to identify
those methods most likely to apply to a particular line
uprating problem. Typically, while several methods can
be eliminated as yielding inadequate ratings increases, or
requiring excessive capital investment, multiple uprating
options are likely to survive such an initial review.
In this subsection, its assumed that a preliminary analy-
sis has been successfully completed and that the most
likely uprating methods have been identified. The goal
here is to prepare a more detailed cost/performance
comparison of likely uprating methods. A test case will
be developed to illustrate the essential features of such a
detailed analysis.
In comparing the total cost of viable uprating alterna-
tives, the major cost and savings factors for each alter-
native need to be determined. The point of the exercise
is to select the minimum cost uprating method that
meets the need for safe and reliable operation, that is
most likely to meet regulatory constraints on line modi-
fications, and that minimizes systemwide costs. In many
cases, the lowest-cost uprating solution may not be
selected because of the many noneconomic constraints
on line design.
Present Worth Calculations
Because line uprating costs can occur immediately, or
over the life of the line, the total costs of each uprating
method should be developed on a present worth basis.
Consider two uprating schemes, which require the same
total amount of capital but with different annual expen-
diture schedules, as shown in Table 2.8-11.
In each case the total cost of uprating is the same,
$15,000, if there is no difference in the value of a dollar
at different points in time. The uprating alternatives
appear to be economically equivalent. But it is clear that
the distribution of annual costsinterest on capital and
inflationhas been ignored. The concepts of present
and future worth of money are explained in many texts
and will be included in the following discussions.
Line Costs and System Savings from Line Uprating
With the exception of electrical losses, and possibly
reduced maintenance from reconductoring, all of the
benefits from line uprating accrue to the power system
at large. In contrast, with the exception of certain
dynamic rating methods, all of the costs associated with
line uprating are specific to the line being modified.
The most significant economic benefits to be expected
from uprating of existing lines include:
Reduced power generation costs by increased access
to low cost sources.
Increased revenues from increased sales of low cost
power to other utilities.
Avoidance of litigation involving clearances and envi-
ronmental effects.
Avoidance of extensive regulatory and public hear-
ings required for new lines.
Postponement of major capital investment.
Reduced maintenance.
Reduced electrical losses.
The major cost components typical of line uprating
include at least some of the following:
Replacement and/or reinforcement of tangent/sus-
pension structures.
Replacement and/or reinforcement of tension/strain
structures.
Purchase of new conductors.
Stringing, sagging and clipping of new or existing
conductors.
Replacement or addition of insulators and hardware.
Addition of wind-motion control devices.
Purchase and installation of monitors and communi-
cations.
Installation and repair of line monitoring devices and
communications.
Increased maintenance associated with higher oper-
ating temperatures.
Increased cost of operator intervention to reduce line
load.
Table 2.8-11 Comparison of Two Uprating Schemes
Year
Uprating Method A
($)
Uprating Method B
($)
0 7,000 15,000
1 2,000 0
2 2,000 0
3 2,000 0
4 2,000 0
TOTAL 15,000 15,000
2-96
Reduced maintenance access to heavily loaded lines
and equipment.
Some of these factorsstructure reinforcement, con-
ductor cost, etc.can be quite readily estimated while
othersincreased levels of maintenance, operation
complexities, etc.can be difficult to quantify. This dif-
ficulty in estimating maintenance is particularly true for
those uprating methods with which the utility has little
or no experience.
In choosing the best line uprating approach, the line
engineer typically finds himself with several different
ways to accomplish the same (or nearly the same) end
result. Given the high level of uncertainty in the power
transmission business today and the uncertainties about
the long-term cost and viability of many of the possibil-
ities, the choice of uprating method cannot be solely
economic. Nonetheless, economic analysis helps to clar-
ify the choices, and is necessary to get funds to do the
modifications.
Identifying Potential Power System Savings
Many transmission line uprating projects are not justifi-
able on a purely economic basis (i.e., the present worth
of savings over the life of the uprated line does not
exceed the present worth of construction capital and
maintenance). Considerations of safety or system reli-
ability are as likely to prompt the decision to uprate a
line, as is the opportunity to reduce electrical losses or
to operate the transmission and generation grid at a
marginally lower cost per kilowatt-hour. On tie-lines
between systems or on radial feeds from low-cost gener-
ators, however, where major savings from the purchase
of more low-cost power or increased income from the
sale of same is at stake, economic justification of uprat-
ing is more likely.
Savings in Generation Dispatch
The potential savings in electrical losses and/or
improved economic generation dispatch associated with
uprating an existing circuit by traditional meanssuch
as reinforcing the towers and reconductoringmust off-
set the very real dollars spent in the rebuilding (typically
in excess of $100,000 per mile, or $62,000 per km). Some
of the lower-cost uprating alternativessuch as the use
of dynamic thermal ratings, increased static thermal rat-
ings based on weather studies, or the selective rebuilding
of critical clearance spansare more likely to prove
economically justifiable since the capital investment is
so much lower.
Dynamic thermal uprating of all lines throughout a sys-
tem on the basis of weather data monitoring has been
economically justified in terms of reduced generation
costs in several references. The calculation of such cost
savings involves load flow studies and is peculiar to each
system. It has been noted, however, that for increases in
capacity of any single line beyond about 10%, other
lines serve to limit dispatch, and little is gained by fur-
ther increases in the capacity of that particular line.
Maximum improvement in economic dispatch is, rather,
gained by systemwide increases in line capacity.
Reference (Nabet 1986) is typical of the more general
claims at economic justification of uprating procedures.
Under a section of the paper called Benefits, G. Nabet
writes ...the use of [systemwide] ambient temperature
adjusted [dynamic thermal] ratings ....from January 1,
1979 to June 30, 1980 .... resulted in reduced off-cost
generation requirements for transmission control ...
[saving] ... 976 MWHR [worth] 1.2 million [dollars].
These savings reflect only those occasions in which off-
cost operation was invoked.
Reference (Hall and Deb 1987) presents a more detailed
attempt at the economic justification of (again dynami-
cally) uprating lines in the PG&E system. The point was
to show that by increasing the thermal ratings of all the
double-circuit lines shown in [their] Figures 8 and 9
from 800 to 1300 A, the total cost of generating power
for loads and line losses decreased by some 18%. The
authors do not attempt to justify the use of such a large
increase in line ratings with dynamic ratings, nor do
they present any cost estimates for the increased opera-
tion and maintenance expenses of the dynamic rating
system. This paper does clearly illustrate the technique
of economic justification of line uprating.
Savings in Electrical Losses
The flow of electrical current on the phase conductors
results in the loss of electrical energy due to conductor
heating. As was noted previously in this section, the
issue of whether electrical losses are included in an eco-
nomic analysis of uprating can have a major impact of
the selection of uprating method.
For example, consider a 10-mile (16-km) long, 115-kV
three-phase transmission line with Drake conductor.
Assume that the current on the phases of the line is con-
stant and equal to the static thermal rating of Drake
ACSR conductor (995 A for a maximum allowable con-
ductor temperature of 75
o
C without sun, 25
o
C ambient
and a 2 ft/sec (0.6 m/sec) cross-wind). The thermal
capacity is 198 MVA.
At 75
o
C, the resistance of 795 kcmil (405 mm
2
) Drake
ACSR is 0.1390 ohms/mile (0.09 ohms/km), and the
total of losses in the three-phase conductors is 413 kW
per mile (0.1390 * 995
2
* 3/1000 = 413). At a power fac-
2-97
tor of 0.95, carrying its full thermal capacity, the 115-kV
line is transmitting 198 MW (0.95 * 995 * 1.732 * 115)
and the electrical losses amount to 0.22% of the real
power transmitted per mile of line.
If the 115-kV line were thermally uprated to 236 MVA
(1185 A per phase) by increasing the line design temper-
ature to 100
o
C (by raising attachment points and/or
retensioning the existing conductor), and if the line cur-
rent again equaled its thermal rating, then the line losses
would be increased by 53%, from 413 kW/mile to 632
kW/mile (256 to 392 kW/km).
Assuming a wholesale electrical power cost of $0.03 per
KW-hr, the cost of electrical losses increases from
$123.39 per hour of operation at the thermal limit to
$189.60 per hour. Clearly, if operation of the original, or
uprated, 10-mile (16-km) long line is operated at or near
its thermal limit for no more than 24 hours per year, the
cost of losses is a minor consideration. If, on the other
hand, the line is operated such that its electrical load is
at or near its thermal capacity for 2 hours per day, the
cost of losses ($189.60 * 2 * 365 = $138,000 annually)
can be a major economic consideration in uprating.
Of course, the comparison of electrical losses for various
uprating alternatives is not quite this straightforward
for a real transmission line. The line loading varies with
the season, the weather, and the time of day. For many
lines, even those that are candidates for uprating, the
normal load may be well below the thermal capacity of
the line (approaching the thermal capacity of the line
only under occasional contingency loadings resulting
from emergency operation). In order to account for the
variation in line load over time, it is normally assumed
that the peak normal load is that which occurs under
normal operation of the transmission system. The
peak post-contingency load is that line load that
occurs only rarely, under emergency operation of the
transmission system. The average load is the average
line load over a certain period of normal operation.
Since the contingency peak load only rarely occurs,
the line losses that occur during such times will be
neglected.
The line load factor (LoadF) for each of those future
years is defined as the ratio of average to peak line load
over the year. In order to calculate the present worth of
line losses, however, one needs to know the loss factor
(LossF)the ratio of average annual electrical losses to
peak lossesrather than the load factor. If the average
current on a conductor over one year is 500 A, and the
peak current over the same period is 1000 A, then the
load factor is 0.50. If the current is quite constant at 500
A, except for brief excursions to 1000 A, then the loss
factor for the conductor is 0.25.
In many cases, the load factor and the loss factor are
often empirically related by a formula such as:
2.8-1
Reduced Emergency Actions
If the capacity of certain lines is increased, then the prob-
ability is reduced that the system operator will be forced
to take emergency actions (e.g., shedding load or initi-
ating quick startup generation). This increase in reli-
ability of power supply to interruptible customers and
reduced use of relatively expensive generation has a real
cost that is very much a function of the particular system.
The reduction of workload for the operator also has
genuine value to the efficient functioning of the system.
Naturally, if the operator has more work to do after the
uprating than before, then this should be evaluated as a
cost of uprating. This is particularly true for instanta-
neous dynamic thermal rating schemes.
Postponed Capital Investment
The selection of uprating method, indeed the initiation
of the uprating process itself, is the result of projected
load growth. Clearly a number of unpredictable factors
are involved in the prediction both of systemwide load,
and even more so of the load on any particular line
within the system. Every utility has seen this very clearly
over the last decade as a result of large swings in fuel
costs, generation construction costs, and conservation
efforts. Decreases in peak contingency loads on the
order of 20 to 30% have occurred with changes in oil
prices. Such major shifts in predicted load render large
capital commitments for uprating hazardous. Similarly
large changes in projected loading of certain lines can
result from wheeling decisions made by neighboring
utilities or from additions of cogeneration.
As a result of the unpredictability of future load growth,
marginal uprating methods have real economic value.
They are especially attractive if they can be applied and
then supplemented or removed at a future date depend-
ing upon whether the projected load growth does or
does not occur. During the short term, such methods
would require only the minimum capital investment.
During the long term, the capital investment in the line
would more closely match the needs of the transmission
system.
A number of marginal methods of uprating have been
discussed. The only commercially available method that
is portable is some sort of weather-based dynamic
2
LossF = 0.15 LoadF + 0.85 LoadF
2-98
thermal rating technique. The techniques of dynamic
rating and selective rebuilding of critical clearance spans
are complementary and could follow one another as
required. Conductor temperature monitors, for exam-
ple, could be removed and used on another circuit if the
line load did not require them in the future.
Estimating Line Uprating Costs
In comparing alternative methods of line uprating, one
must consider all of the cost components and the likeli-
hood that each uprating method can meet the power
system needs. Line uprating costs are very dependent on
the design details of the existing line. For example, if the
structures were designed to withstand much larger
transverse and/or longitudinal forces than are produced
by the existing conductors, reconductoring can be
accomplished without the expense of structure modifi-
cation. Similarly, if the original design allowed for
ground clearances that greatly exceed the NESC mini-
mums, it may be possible to operate the line at higher
electrical loads without any physical modification or
expense.
Structures and Foundation Modifications
Modifications to existing structures should be consid-
ered in a two-part process. First, consider only the tan-
gent structures, assuming that angle structures will have
to be either rebuilt or replaced if the conductor diameter
or tension levels are changed. Determine the cost of
modifying the tangent structures on the line as a func-
tion of the conductor diameter accounting for changes
in code or loading since the line was built. Second, if the
uprating cost for tangent structures appears to be rea-
sonable for the required increase in thermal rating or
voltage, then consider the cost of modifying angle struc-
tures and dead-ends, and do a detailed clearance study
of the line for the most economical conductor diameter.
As a rough rule of thumb, if the study of tangent struc-
tures shows that necessary structure modifications can
be accomplished for less than 50% of the cost of replac-
ing the existing structures with new, then a more
detailed analysis is justified.
Reconductoring Costs
If the existing conductors are to be replaced, then a
number of conductor options exist. A new conductor
may be bundled with the existing conductor or the exist-
ing conductor can be replaced with:
A conductor having less electrical resistance.
A special conductor capable of higher unloaded ten-
sion.
A special conductor capable of operating at higher
temperatures with reduced sag.
Bundling of new and old conductor doubles loading on
the structure. The cost of installing the second conduc-
tor in either a vertical or horizontal bundle costs only
slightly more than installing a new conductor on the
same structures. The added cost is due to the possible
need to pre-stress the new conductor and the need to
work around the existing conductor.
Reconductoring with a larger conductor costs roughly
what the stringing, sagging, and clipping of new con-
ductor usually costs. The old conductor may have signif-
icant scrap value if it is all aluminum. In any event, if it
can be reused, the cost of this operation should be
reduced accordingly.
Reconductoring with special conductors may cost some-
what more than using standard conductors. There is
often a premium of 5 to 15% associated with conductors
such as SDC or SSAC. Also the use of higher installa-
tion tensions and special handling may cause a contrac-
tor premium.
Operation and Maintenance Costs
Uprating an existing line will inevitably lead to higher
electrical loads. This can cause reduced life for conven-
tional current-carrying components and/or the need for
shorter inspection intervals. The use of real-time moni-
tors may also increase the need for maintenance and
problem-solving where none existed before. Also,
Dynamic rating monitors are installed in a hostile
electromagnetic environment. Though sometimes
installable with the line in service, their removal from
certain spans of the line can be a significant expense.
It is essential that the manufacturer provides the user
with the possibility of field correction, or at least
detection of errors, so that recalibrations can be kept
to a minimum.
Reconductoring with novel conductors may be very
attractive, but the aggressive utilization of new mate-
rials and products can add to maintenance and repair
activities.
Unless engineers have extensive experience with particu-
lar techniques of uprating, they should first move to
gain experience with a pilot project. Based on this work,
one should be certain to allow for any unforeseen prob-
lems that might develop years after the uprating occurs.
If line ratings are to be only marginally increased, the
need for operator intervention may increase. Dynamic
ratings in particular offer added complexity to the sys-
tem operator, both in assessing the adequacy of the
transmission system, and in establishing contractual
obligations for the transfer of power between systems.
2-99
The operator must help to define his needs for display of
variable thermal ratings. This need is likely to involve
additional software and display hardware in the energy
control center.
If the interval between uprating reviews of lines is to be
reduced in order to more closely match capital invest-
ment to capacity needs, the operator must be closely
integrated into the process of review and uprating deci-
sions. This implies an increase in the time that the sys-
tem operator i s to spend i n communi cati ng hi s
experience to planners and engineering personnel.
Miscellaneous Cost Issues
If a particular method of uprating is applicable to many
lines in a particular utility system, then the engineering
cost can be spread over several upratings. Otherwise, the
cost of engineering time is a real consideration in mak-
ing some of the more complex uprating schemes work.
Traditional uprating of lines by rebuilding with lines of
the next higher voltage class or reconductoring with
larger standard conductors means that the transmission
system planner need only concern himself with certain
lines at widely spaced intervals of timeat least 5 to 10
years. Marginal uprating techniques imply more fre-
quent and more sophisticated studies of line capacity,
including the possibility of small changes in capacity
that must be reviewed for sufficiency frequently.
Dynamic thermal uprating techniques offer particular
challenges to the system planner. Standard software
tools must be modified to consider the probabilistic
aspect of thermal limits, which depend on weather and
time of day. This is a very real cost, even though it is dif-
ficult to quantify.
2.8.6 Detailed Comparison of Uprating
AlternativesAn Example
Consider a 10-mile (16-km) long, 115-kV transmission
line with 336.4 kcmil (172 mm
2
) Linnet ACSR conduc-
tor installed to a final unloaded tension of 16% UTS at
60F. The wood pole H-frame structures are spaced
quite uniformly at 600 ft (180 m). According to the util-
ity operating this line, it is assumed that the conductor
has a summertime thermal capacity of 430 A (75
o
C con-
ductor, 40C air, 2 ft/sec (0.6 m/sec) cross-wind, with
sun) that corresponds to a thermal line capacity of 85
MVA. The sag-tension data for the original conductor is
shown below in Table 2.8-12.
The existing structures are in good condition, and it is
possible to reinforce strain structures to allow an
increase in the present maximum conductor tension
(6410 lbs or 28.6 kN) of up to 50% for a cost equal to
less than 10% of rebuilding all structures. The transverse
load capability of the existing tangent structures is such
that the diameter of the replacement conductor can be
up to 10% higher than that of the existing conductor
(0.72 in., or 1.8 cm) without reinforcing or replacing
tangent structures.
The suspension structure conductor attachment height
may not easily be increased, and the existing lines
ground clearances with a sag of 13.2 ft (4 m) at 75C are
barely adequate. Therefore this maximum high-temper-
ature sag may not be exceeded in any of the uprating
alternatives.
Table 2.8-12 ALCOA Sag and Tension Data
ALUMINUM COMPANY OF AMERICA SAG AND TENSION DATA
600 ft spans, 16% final, 75C max, w comp
Conductor LINNET 336.4 Kcmil 26/ 7 Stranding ACSR
Area = 0.3070 sq in. Dia = 0.720 in. Wt = 0.463 lb/F RTS = 14100 lb
Span = 600.0 ft NESC Heavy Load Zone
Creep is a Factor Rolled Rod
Temp Ice Wind K Weight Sag Tension Sag Tension
(F) (in.) (psf) (lb/F) (lb/F) (Ft) (lb) (Ft) (lb)
0. .50 4.00 .30 1.650 12.58 5915. 11.61 6409.
-20. .00 .00 .00 .463 5.85 3564. 4.13 5047.
0. .00 .00 .00 .463 6.67 3124. 4.44 4695.
60. .00 .00 .00 .463 9.55 2186. 5.75 3626.
120. .00 .00 .00 .463 12.08 1729. 7.78 2679.
167. .00 .00 .00 .463 13.23 1579. 9.78 2133.
212. .00 .00 .00 .463 14.33 1458. 11.78 1772.
2-100
The existing Linnet conductor is in reasonably good
condition with an expected life of at least 20 years. It is
capable of operation at a temperature above the present
limit of 75
o
C, but only if it can be operated at a higher
temperature without exceeding the present sag of 13.2 ft
(4 m) at 75
o
C.
Because the line is short, stability problems are not of
concern. The voltage drop per mile is considerable, how-
ever, with a current of 430 A, a conductor temperature
of 75C, and a 90% power factor is:
2.8-2
Setting a 10% voltage drop as a limit during emergency
loadings, all lines of this construction, having a thermal
limit of 85 MVA are voltage constrained at lengths
greater than 35 miles (56 km). Note that as the thermal
rating of the line is increased, the line length at which
the line is voltage drop limited decreases. Thus increas-
ing the thermal rating of a 25-mile (40-km) line of this
design to 125 MVA would make it voltage drop rather
than thermally limited. In the present example, however,
the 10-mile-long lines rating would have to be increased
to 300 MVA before it became voltage-drop limited.
The projected growth of the peak emergency line load-
ing is shown in Figure 2.8-5. Note that the thermal
capacity of the line will be exceeded by the peak contin-
gency loading in 2 to 5 years for the pessimistic (1%)
and the optimistic (3%) projections, respectively.
Preliminary Uprating Analysis
A preliminary uprating assessment of the line has been
performed, as shown in this Uprating Analysis Table
(Table 2.8-13). Certain uprating methods seem inappro-
priate. For example, since the line is presently clearance
limited with Linnet at 75
o
C and the tangent structures
will not allow an increase in conductor attachment
height, it is not possible to lift the conductor attachment
points in order to increase the line design temperature.
Alternatively, the relatively modest increase in thermal
rating and its uncertainty indicates that dynamic rating
methods may work well.
After considering the capabilities and conventions of the
transmission owner, the following three alternative
uprating methods are identified as possible:
A. Reconductor the line with a lower resistance, trape-
zoidal wire, Hawk/TW ACSR conductor, reinforcing
the strain structures. The 10% larger diameter of
Hawk/TW can be accommodated by the existing
structures.
B. Install a dynamic thermal rating system based upon
the use of conductor sag-tension monitors along the
line. After some period of time depending upon the
line load growth rate, remove the dynamic rating sys-
tem, and increase the tension of the existing Linnet
conductors to allow operation at higher temperature
for the same maximum sag.
( )
( )
100 cos sin
%
3
100 430 0.328 0.90 0.336 0.436
115
3
0.286%
LL
I Rac X
VoltDrop
kV
per mile
+

=

+

=

=
Figure 2.8-5 Example of projected growth of peak
emergency line loading.
Table 2.8-13 Uprating Analysis Table
2-101
C. Reconductor the line with Linnet ACSS without
modifying either tangent or strain structures.
Reconductoring with ACSR/TW Conductor
Reconductoring with 477 kcmil (243 mm
2
), 26/7 Hawk
ACSR/TW conductor will yield a thermal rating of 525
A (105 MVA) at 75
o
C and 700 A (139 MVA) at 100
o
C
due to the reduced resistance of Hawk/TW. The
increased thermal capacity with Hawk/TW will be ade-
quate for at least 10 years under the least-conservative
assumption of 3% annual load growth.
In order to meet the maximum sag limit of 13.2 ft (4 m)
at 75
o
C , Hawk/TW must be installed to an initial
unloaded tension of 29% UTS at 0F (-18C) and a cor-
responding maximum tension under NESC Heavy load-
ing conditions of 7750 lbs (35 kN) (21% above the
existing Linnet ACSR). If the initial unloaded tension at
0F (-18C) is increased to 33%, the sag limit is met at
100
o
C, but the maximum tension increases to 8310 lbs
(37 kN) (30% above the present line). In either case, the
maximum tension is well within the 50% increase limit
in maximum tension load.
Hawk/TW has a diameter of 0.781 in. (2 cm) (8.5%
greater than Linnet), so it is likely that the existing tan-
gent structures will not require reinforcement.
Having established the increase in thermal capacity pos-
sible by reconductoring with Hawk/TW, other substa-
tion equipment limits and replacement costs need to be
reviewed.
The total line construction cost (i.e., strain structure
modifications, replacement conductor, vibration damp-
ers, labor cost, etc.) is a weak function of maximum ten-
sion, since only strain structures need to be modified.
Its assumed that the reinforcement of strain structures
costs 5% of the structures for a new 115-kV line. It is
further assumed that this amounts to $5,000 per mile
($3k per km).
In addition to the cost of upgrading structures, the most
significant cost is that of the new conductor minus the
scrap value of the old. Typically, one may obtain con-
ductor costs and scrap value from a manufacturer. We
will assume that the new Hawk/TW ACSR conductor
costs $2.00 per ft ($6.70 per m) and that the scrap value
of the old conductor is $0.50 per ft (1.50 per m). Recon-
ductoring the line with Hawk/TW involves a total mate-
rial cost of $24,000 per mile ($14,400 per km).
Other costs include stringing, sagging, and clipping the
new conductor, new hardware, and engineering design
costs. We will assume that this cost equals that of the
Hawk/TW material.
Finally, the present worth of electrical losses over the
life of the reconductored line should be calculated. It is
assumed that the normal annual peak line load that is
presently 50 MVA will increase to 65 MVA over an esti-
mated 20-year useful life of the reconductored line. It is
also assumed that the peak contingency load is 1.6 times
the peak normal annual load of the line, and that the
loss factor is 40%.
Economic Parameters for Loss Calculation
Years of Analysis: 20 years
Interest Rate: 8%
Energy charge: $0.020/kW-hour
Energy charge Escalation Rate: 7%
Using the economic data in the preceding paragraph
and in Table 2.8-14, the present worth of electrical losses
with the existing Linnet conductor over the 20-year
period is $610k. The resistance of Hawk/TW is approxi-
mately 70% (336.4/477 = 0.71) that of Linnet. Therefore,
the savings in present worth of electrical losses for
Hawk/TW is $18k/mile ($11k/km).
During the reconductoring of the line, the system opera-
tor cannot use the circuit. Construction could take
months and higher cost generation may have to be pur-
chased during this period. The cost of this loss of the
circuit could be determined by a system load flow analy-
sis. The cost is primarily due to higher costs of genera-
tion due to non-optimum generation dispatch during
construction and increased losses on other lines whose
loads increase
Table 2.8-14 summarizes the costs and savings associ-
ated with reconductoring the line with Hawk/TW con-
ductor.
Table 2.8-14 Summary of Cost Savings Associated with
Reconductoring
Conductor Name Hawk/TW
OD (in.) 0.782
Structures $5,000/mi
Conductor $24,000/mi
Conductor labor $24,000/mi
Total construction $53,000/mi
Cost of increased losses during construc-
tion
?
Savings in PW losses over 20-year life $18,000/mi
Net PW cost of line operation over 20
years (ignore losses)
$53,000/mi
Net PW cost of line operation over 20
years (include losses)
$35,000/mi
2-102
Dynamic Uprating Method
If the existing Linnet conductor is in good condition,
the line rating can be increased through the installation
of sag-tension monitors along the line. Many papers
have dealt with these techniques.
Before doing any economic calculations, it is essential
that one determines just how much the use of a dynamic
rating method increases the thermal rating of the line.
As was discussed previously, the dynamic thermal rating
of the line is both random and chronologicalthat is,
there is a certain amount of uncertainty combined with
a certain degree of predictability based on season and
time of day. Based on reference (Hall and Deb 1987), for
a line in upstate New York, one may expect that the
dynamic thermal rating of the 115-kV line with Linnet
conductor will be:
10 to 20% above the static rating 50% of the time.
Above the static rating 90% of the time.
Below the static rating 10% of the time (usually at
night).
One must decide how to interpret these numbers in
terms of traditional planning criteria. One must recog-
nize that the operator will have to intervene occasionally
to reduce the loading of this line during times of peak
loading and minimal rating. During these times when
the load is high and the dynamic rating is low, the sys-
tem will be operated in an uneconomic mode or load
may need to be shed. Its assumed that operating per-
sonnel feel that they can intervene during 10% of the
peak loading events without incurring significant costs.
Then one may credit the dynamic rating system with
increasing the thermal rating of the line by 10%, or from
85 to 94 MVA.
For a 2% annual growth rate, the peak contingency load
will reach 94 MVA in 8 years. Therefore, one may
assume that the useful life of this dynamic rating
approach is 8 years, after which the line must be modi-
fied in some other manner to increase its thermal rating
(e.g., raise structures, reconductor, etc). The installation
of the dynamic rating system does nothing to reduce
losses since the conductor resistance is unchanged.
It may be assumed that the dynamic rating equipment
costs about $100,000 for a 10-mile (6.2 km) line. The
monitoring system is reusable on other lines after 8
years, or earlier if the load increases more rapidly than
predicted.
Line monitors are usually installed by bucket truck if
terrain permits. Installation expenses vary widely since
they depend on terrain and accessibility along the line.
This is also true for maintenance. The monitors pres-
ently available require periodic recalibration and proba-
bly should be checked annually. A guess for initial
installation cost of the line monitors might be $10,000
with an annual maintenance cost of the order of $5,000.
Prior to beginning the dynamic rating of the line, one
must allow for a complete inspection of the structures
and conductor to spot bad splices and impaired clear-
ances. An inspection of the 10-mile (6.2-km) line is
required for any of the three alternatives. The use of line
monitors offers the unique advantage of establishing an
experimental basis for high-temperature clearance.
Any line outage required to install the line monitors is
brief, typically less than 24 hours. For EPRIs video sag-
ometer, no outage is needed.
Assume that at the end of 8 years the dynamic rating
monitor system is worth 50% of its initial purchase
price. After 8 years, the dynamic rating system will be
removed and reused elsewhere in the system. The line
would then be surveyed, certain critical spans selected
for increase in clearance, the allowable conductor tem-
perature increased, and 50 MVA of power transformer
capacity added. The use of dynamic ratings would be
discontinued at this point.
Its estimated that the cost of retensioning the existing
Linnet ACSR to 20% UTS at 60F will be equal to half
that of installing new conductor or approximately
$12,000/mile ($7,400/km). The increased everyday ten-
sion will require the use of dampers costing about
$2,000/mile ($1240/km). The retensioned line will then
have a line design temperature of 100
o
C and a rating of
575 A (115 MVA). If load growth is faster than antici-
pated, then this may not be adequate, and the line will
have to be either reconductored or rebuilt.
The present worth of $14,000/mile 8 years in the future
is $10,000/mile, so the total present worth cost of this
uprating approach is $21,000/mile.
Reconductor with Linnet/ACSS
Linnet/ACSS can be applied to the line without the need
to rebuild or reinforce any of the structures. If the the
Linnet/ACSS is installed to maximum NESC Code lim-
its at 60
o
F, it reaches the sag limit of 13.2 ft (4 m) at
about 150
o
C. The line rating is, therefore, 770 A
(153 MVA).
The cost of reconductoring is limited to the conductor
and its installation. Given the premium typical of ACSS
of $2.00 per ft ($6/m) for the conductor and an equal
2-103
amount for stringing, sagging, and clipping it in place of
the original Linnet, the total cost is $63,000 per mile
($39k/km).
Dampers will also be required because of the high initial
tension levels. Therefore, the total estimated cost is
$65,000 per mile ($40k/km).
Economic Comparison of Uprating Alternatives
The total present worth of construction and losses to
meet the increased thermal rating requirements of this
10-mile line in the three different uprating methods are
summarized in Table 2.8-15.
Clearly, the use of dynamic ratings followed by a reten-
sioning of the existing conductor (if required by actual
load growth) is the most flexible approach, and requires
the least initial and total capital investment. The major
drawbacks involve the need to modify standard operat-
ing procedures to utilize real-time ratings and the mod-
est rating increase that results.
The use of ACSS requires an absolute minimum of
structure reinforcement since Linnet/ACSS has the same
diameter as the original Linnet ACSR and yields
reduced maximum tension because of its reduced modu-
lus. There is no reduction in electrical losses since the
Linnet/ACSS has nearly the same resistance as the orig-
inal Linnet. In pursuing other alternatives, it is likely
that an ACSS/TW conductor with a slightly larger
diameter than Linnet would be a better choice.
The uprating option requiring the largest capital invest-
ment is the reinforcement of strain structures and the
aggressive use of vibration dampers in order to recon-
ductor the line with Hawk/TW ACSR. This option is
unique in that it reduces electrical losses as well as
increasing the line rating.
Review of other line uprating options and refinement of
these three is clearly worthwhile. The means for identify-
ing other possible uprating options and selecting the
most appropriate has been presented in the preceding
notes.
2.8.7 Conclusions
The impetus for line uprating comes as a result power
system analysis. Present electrical loads are projected
into the future, and the impact of various component
outages (i.e., contingencies) on the electrical loading of
the existing line is determined. Specific probabilities are
seldom associated with post-contingency loadings, and
even the prediction of normal loads is often uncertain,
particularly with the advent of open access to com-
mercial power generators.
For uprating to be possible, the existing line must be in
good condition. Having established this, the identifica-
tion of possible uprating methods depends upon the
physical, electrical, and thermal characteristics of the
existing line. An Uprating Analysis Table is developed
here that simplifies the analysis of the existing line and
provides a basis for identifying promising uprating
methods in each specific line. Once the most promising
uprating methods have been identified, a detailed analy-
sis comparing the costs and capabilities of each method
is required.
The final selection of an uprating method and its suc-
cess in providing the necessary increase in line capacity
while maintaining system reliability and minimizing
capital cost involves a good deal of engineering judg-
ment, as well as the application of suitable numerical
tools.
Table 2.8-15 Present Worth of Three Uprating Options
Option A1 Reinforcing strain
structures, adding dampers,
and reconductoring with
Hawk/TW for a line design
temperature of 100
o
C.
$53,000/mile (ignoring losses)
Option A2 Same as A1 but
include loss savings
$35,000/mile
(allowing for loss savings)
Option B Apply dynamic r
ating monitors and retension
existing Linnet ACSR if pre-
dicted load growth requires it.
$20,900/mile
Option C Reconductor with
Linnet/ACSS and go to a line
design temperature of 125
o
C.
$65,000/mile
2-104
REFERENCES
Adams, H. W. 1974. Steel Supported Aluminum Con-
ductors (SSAC) for Overhead Transmission Lines.
IEEE Paper T 74 054-3. Presented at the IEEE PES
Winter Power Meeting.
Alcoa. 2003. Alcoa Sag10 Users Manual.
Aluminum Association. 1974. Stress-Strain-Creep
Curves for Aluminum Overhead Electrical Conductors.
Aluminum Association. 1989. Aluminum Electrical
Conductor Handbook. Third Edition.
Aluminum Company of America. 1961. Graphic
Method for Sag Tension Calculations for ACSR and
Other Conductors.
Aronstein, J. 1990. Conduction in Failing Aluminum
Connections. Proceedings of the Thirty-Sixth IEEE
Holm Conference on Electrical Contacts. Montreal,
Quebec. August.
Barrett, J. S., P. Ralston, and O. Nigol. 1982. Mechani-
cal Behaviour of ACSR Conductors. CIGRE Interna-
tional Conference on Large High Voltage Electric
Systems. September 1-9.
Beers, G. M., et al. 1963 Transmission Conductor Rat-
ings. AIEE Transactions. Paper 63-86.
Bennett, E. H. 1992. Designing Compression Fittings
for Long-Term Survival. Bonneville Power Engineer-
ing Symposium. April.
Black, W. Z., W. R. Byrd, R. A. Bush, and T. C. Cham-
pion III. 1983. Experimental Verification of a Real-
Time Program for the Determination of Temperature
and Sag of Overhead Lines. Paper 83 WM 144-3.
January.
Black, W. Z. and W. R. Byrd. 1983. Real Time Ampac-
ity Model for Overhead Lines. IEEE Transactions.
Vol. PAS-102. No. 7. July. pp. 2289-2293.
Black, W. Z. and R. L. Rehberg. 1985. Simplified
Model for Steady State and Real-time Ampacity of
Overhead Conductors. IEEE Transactions on Power
Apparatus and Systems. Vol. 104. October. Pp. 29-42.
Boteler, D. H. 1994. Geomagnetically Induced Cur-
rents: Present Knowledge and Future Research. IEEE
January. pp. 50-58.
Braunovic, M. 1985. Effect of Contact Aid Com-
pounds on the Performance of Bolted Aluminum-to-
Aluminum Joints Under Current Cycling Conditions.
31st Annual Holm Conference. Chicago, IL. September.
Cahill, T. 1973. Development of Low-Creep ACSR
Conductor, Wire Journal. July.
Campbell, H. E. and J. J. Burke. 1985. Power Distribu-
tion Systems Course. PTI Course. September.
Chisholm, W. A. 1986. Ampacity Field Studies on Line
with Low Operating Temperature. EPRI DTR Semi-
nar. May.
CIGRE. 1992. WG 05 Conductors. The Thermal
Behaviour of Overhead Conductors. 22-81 (WG05)
06. December.
CIGRE. 1997. WG 12-22. Thermal State of Overhead
Line Conductors. Electra. No. 121. pp. 51-67.
CIGRE. 2000. WG 22-12. Description of State of the
Art Methods to Determine Thermal Rating of Lines in
Real-Time and Their Application in Optimising Power
Flow. Paper 22-304.
Clapp, A. L. 1985. Relationships of National Electrical
Safety Code Vertical Clearances and Potentially Con-
flicting Activity. IEEE Transactions on Power Appara-
tus and Systems. Vol. PAS-104. No. 11. November. pp.
3306-3312.
Dalle, B. 1982. Size and Aging of Joints for Bare Con-
ductors of Overhead Line. Electricite de France. Decem-
ber.
Davidson, G. A., et al. 1969. Short-time Thermal Rat-
ings for Bare Overhead Conductors. IEEE Transac-
tions on Power Apparatus and Systems. Vol. PAS-88.
No.3. March.
Davis, M. W. 1979. Development of Real Time Thermal
Rating System. St. Louis, MO: Edison Electrical Insti-
tute T&D. May 19.
2-105
DeLuca, C. B. 1986. Current Cycling Connectors in
Tension. Proceedings of Seminar on Effects of Ele-
vated Temperature Operation on Overhead Conductors
and Accessories. pp. 110-119. Atlanta, Georgia. May.
Douglass, D.A. 1986. Weather-Dependent Versus
Static Thermal Line Ratings. IEEE Paper No. 86
T&D 503-7. Presented at the IEEE T&D Conference.
Anaheim, California. September.
Douglass, D. A., I. S. Grant, et al. 1986. Optimization
of Line Uprating. Presented at IEEE/PES T&D Spe-
cial Session Upgrading Transmission Lines. Ana-
heim, California. September.
Douglass, D. A. and L. S. Rathbun. 1984. AC Resis-
tance of ACSR - Magnetic and Temperature Effects.
IEEE Paper 84 SM 700-1.
Douglass, D. A. and A. Edris. 1996. Real-time Moni-
toring and Dynamic Thermal Rating of Power Trans-
mission Circuits. IEEE Transactions on Power Delivery.
Vol. 11. No. 3. July.
Douglass, D. A. and A. Edris. 1999. Field Studies of
Dynamic Thermal Rating Methods for Overhead
Lines. IEEE T&D Conference Report. New Orleans.
April 7. New Orleans, LA.
Dunlop, R. D., R. Gutman, and P. P. Marchenko. 1979.
Analytical Development of Loadability Characteristics
for EHV and UHV Transmission Lines. IEEE Trans-
actions on Power Apparatus and Systems. Volume 98.
Number 1. March/April. pp. 606-617. correction
May/June. page 699.
Dupre, H. 1951. The Problems Involved in Designing
Connectors for Aluminum Cable. AIEE 51-325.
September.
Edris, A. 2000. FACTS Technology Development: An
Update. IEEE Power Engineering Review. Vol. 20. No.
3. March.
Ehrenburg, D. O. 1935. Transmission Line Catenary
Calculations. AIEE Paper. Committee on Power
Transmission & Distribution. July.
EPRI. 1994. Handbook of Shielding Principles for Power
System Magnetic Fields: Volume 1: Introduction and
Application Volume 2: Methods and Measurements.
103630. Palo Alto, CA.
EPRI. 1995. Thermal Models for Real-Time Monitoring
of Transmission Circuits. Report No. TR-105421.
December.
EPRI. 2001. Video Sagometer Application Guide. EPRI
Report N0. 1001921. September.
EPRI. 2002. Performance of Transmission Line Compo-
nents at Increased Operating Temperatures. EPRI
Interim Report. December.
EPRI. 2005. Transmission Line Reference Book, 345 kV
and Above, EPRI, Palo Alto, CA.
Federal Power Commission. 1964. National Power Sur-
vey. Part II-Advisory Reports. U. S. Government Print-
ing Office. Washington, D. C. October.
Fink, D. G. and H. W. Beaty. 1993. Standard Handbook
for Electrical Engineers. 13th Edition. McGraw-Hill.
Foss, S. D., S. H. Lin, and R. A. Fernandez. 1983.
Dynamic Thermal Line RatingsPart 1Dynamic
Ampacity Rating Algorithm. IEEE Transactions on
Power Apparatus and Systems. Vol. PAS-102. No. 6. pp.
1858-1864. June.
Frank, W. 1959. The Critical Aspects of Steel Hard-
ware in Aluminum Connectors. AIEE Transmission
and Distribution Committee. June.
Hall, J. F. and A. K. Deb. 1987. Economic Evaluation
of Dynamic Thermal Rating by Adaptive Forecasting.
IEEE Paper 87 SM 556-4. Presented at the IEEE Sum-
mer Power Meeting. July.
Harvey, J. R. 1969. Creep of Transmission Line Con-
ductors. IEEE Transactions. Vol. PAS-88. No. 4. pp.
281-285. April.
Harvey, J. R. and R. E. Larson. 1972. Creep Equations
of Conductors for Sag-Tension Calculations. IEEE
Paper C72 190-2.
Harvey, J. R. and R. E. Larson. 1970. Use of Elevated
Temperature Creep Data in Sag-Tension Calcula-
tions. IEEE Transactions. Vol. PAS-89. No. 3. pp.
380-386. March.
Hickernell, L. F., A. A. Jones, and C. J. Snyder. 1949.
Hy-Therm Copper An Improved Overhead Line
Conductor. AIEE Transactions. Vol. 68. pp. 22-27.
2-106
Hiel, C. 2000. Development of a Composite Rein-
forced Aluminum Conductor. California Energy Com-
mission Report. November.
House, H. E., W. S. Rigdon, R. J. Grosh, and W. B. Cot-
tingham. 1963. Emissivity of Weathered Conductors
after Service in Rural and Industrial Environments.
AIEE Transactions. pp. 891-896. February.
Howitt, W. B. 1986. Elevated Temperature Performance
of Conductor Accessories. Proceedings of Seminar on
Effects of Elevated Temperature Operation on Over-
head Conductors and Accessories. pp. 120-139.
Atlanta, Georgia. May.
IEEE. 1993a. Standard 738-93. IEEE Standard
for Calculation of Bare Overhead Conductor
Temperatures.
IEEE. 1993b. IEEE Standard for Calculating the Cur-
rent-Temperature Relationship of Bare Overhead Con-
ductors. PES. IEEE Standard 738-1993.
IEEE. 1993c. Guide to the Installation of Overhead
Transmission Line Conductors. IEEE Standard 524-
1993. New York, NY.
IEEE. 1999. Limitations on Stringing and Sagging
Conductors, Paper TP64-146. Working Group of the
IEEE Towers, Poles, and Conductors Subcommittee of
the Transmission and Distribution Committee of the
IEEE Power Engineering Society.
Johnson, D. 2001. Composite ConductorsA New
Overhead Conductor (ACCR). IEEE Panel Session on
Applications and Economics of New Conductor Types.
Vancouver, B.C. July.
Kikuta, T. 2001. Low Sag Up-rating Conductor.
IEEE Panel Session on Applications and Economics of
New Conductor Types. Vancouver, B.C. July.
Koessler, R. J. and J. W. Feltes. 1993. Voltage Collapse
Investigations with Time-Domain Simulation.
IEEE/NTUA Joint International Power Conference.
Athens Power Tech Proceedings. Athens, Greece. Sep-
tember 5-8.
Lesher, R. L., J. W. Porter, and R. T. Byerly. 1994. Sun-
burstA Network of GIC Monitoring Systems. IEEE
January. pp. 128-137.
Lewis, W. A. and P. D. Tuttle. 1958. The Resistance
and Reactance of Aluminum Conductors Steel Rein-
forced. AIEE Transactions. Vol. 77. Part III.
Longo, V. J. and I. S. Grant. 1981. Economic Incen-
tives for Larger Transmission Conductors. IEEE Paper
81 WM 208-8. Presented at the IEEE Winter PES Meet-
ing, February.
Nabet, G. 1986. Thermal Ratings for Bare Overhead
ConductorsMethod Used by Pa-NJ-Maryland Inter-
connection. Proceedings of Seminars on Effects of Ele-
vated Temperature Operation on Overhead Conductors
and Accessories. Atlanta, Ga. May. (Sponsored by
Georgia Power and EPRI).
National Electric Safety Code. 1993. 1993 edition.
National Electric Safety Code. 1997. 1997 edition.
C2-1997.
Naybour, R. D. and T. Farrell. 1973. Degradation Mech-
anisms of Mechanical Connectors on Aluminum Conduc-
tors. PROC IEE. Vol. 120. No. 2. pp. 273-280.
February.
NEMA. 1973. Standard, EEOI-NEMA. Connectors for
use Between Aluminum or Aluminum-Copper Overhead
Conductors. NEMA Pub. No. CC 3-1973. August.
Nigol, O. and J. S. Barrett. 1980. Development of an
Accurate Model of ACSR Conductors for Calculating
Sags at High Temperatures. Ontario Hydro Research
Division. CEA. March.
Rawlins, C. B. 1998. Some Effects of Mill Practice on
the Stress Strain Behavior of ACSR. Presented at
IEEE Winter Meeting. Tampa, FL. February.
Reding, J. L. 1991. Investigation of Thrasher Compres-
sion Fittings on BPA's Direct Current Transmission
Line. IEEE Trans. PWRD-6. No. 4. pp. 1616-1622.
October.
Reding, J. L. 1994. A Method for Determining Proba-
bility Based Allowable Current Ratings for BPAs Trans-
mission Lines. IEEE Transactions on Power Delivery.
Vol. 9. No. 1. January.
Rural Electrification Administration. 1992. REA Bulle-
tin 1724E-200. Design Manual for High Voltage Trans-
mission Lines. 9/3/92.
2-107
Sato, K., N. Mori, et al. 1981. Development of
Extremely-Low-Sag Invar Reinforced ACSR
(XTACIR). IEEE Transactions on Power Apparatus and
Systems. Vol. PAS-100. No. 4. April.
Schmidt, N. 1997. Comparison between IEEE and
CIGRE Ampacity Standards. IEEE PE-749-PWRD-0-
06-1997. Berlin, Germany. July.
Seppa. T. O., et al. 1998. Use of On-Line Tension
Monitoring Systems for Real Time Ratings, Ice Loads
and Other Environmental Effects. CIGRE Report 102-
22. September. Paris, France.
Smith, J. 2001. (Applied Thermal Sciences). Advanced
Carbon Conductor Project Report. National Science
Foundation. August.
St. Clair, H. P. 1953. Practical Concepts in Capability
and Performance of Transmission Lines. AIEE Trans-
actions on Power Apparatus and Systems. Volume 72.
Part III. December. pages 1152-1157.
Troia, G. D. 2000. Effects of High Temperature Opera-
tion on Overhead Transmission Full-Tension Joints and
Conductors. IEEE WG 12. August.
Trash, R. et al. (ed.) 1994. Overhead Conductor Manual.
Southwire.
Thrash, F. R. 1999. ACSS/TW - An Improved Conduc-
tor for Upgrading Existing Lines or New Construc-
tion. IEEE Paper 0-7803-5515. June.
Tunstall, M. J., et al. 2000. Maximizing the Ratings of
National Grids Existing Transmission Lines Using
High Temperature, Low-sag Conductor. CIGRE
Paper 22-202. Paris, France. August.
Turner, D. K. and B. J. Belk. 1987. Elevated Conductor
Temperatures and Their Effect on Planning and
Design. Presented at the Transmission and Substation
Design and Operation Symposium. Ft. Worth, Texas.
September.
Winkelman, P. F. 1959. Sag-Tension Computations
and Field Measurements of Bonneville Power Adminis-
tration. AIEE Paper 59-900. June.
Wong, T. Y., J. A. Findlay, and A. N. McMurtie. 1982.
An On-Line Method for Transmission Ampacity Eval-
uation. IEEE Transactions on Power Apparatus and
Systems. Vol. PAS-101. No. 2. February.

3-1
CHAPTER 3 Underground Cables
3.1 INTRODUCTION
Many factors need to be considered when evaluating uprating and upgrading options for
underground transmission cables. Chapter 3 provides a general description of concepts
that a utility engineer should consider in understanding underground power cables and
selecting uprating technologies that are appropriate. While this chapter focuses on trans-
mission, many of the concepts are equally valid for distribution cables, although the cost-
benefit ratio of applying these techniques to distribution is less easy to justify.
Chapter 3 includes nine sections:
Section 3.2, Cable System Types, provides an overview on underground cable systems
and a very brief background on each of the major transmission cable types.
Section 3.3, Power Flow Limits and System Considerations, considers aspects external to
a specific cable circuit that may limit power flow regardless of the cable circuits rating.
Section 3.4, Underground Cable Ratings, provides an overview of cable system ampac-
ity, including worked examples, to understand the basic approach to calculating ratings
and the areas where uprating or upgrading could be applied.
Section 3.5, Uprating and Upgrading Constraints, lists some of the major barriers to
uprating that are inherent to each cable system type or installation location.
Section 3.6, Increasing the Ampacity of Underground Cable, is the major focus of this
chapter, describing how to increase capacity on existing circuits. Other sections of the
chapter have been provided to support this chapter.
Section 3.7, Reconductoring (Upgrading), discusses topics related to replacing cables to
increase capacity, possibly combined with other uprating considerations.
Section 3.8, Dynamic Ratings of Underground Cable Systems, includes information on
the state-of-the-art methods used for optimizing the rating on a cable system, including
topics on real-time monitoring and ratings.
Section 3.9, Case Studies for Underground Cable Circuits, describes real-world uprating
applications that have been implemented by utility-users, along with their respective
experiences.
Section 3.10, Summary of Uprating and Upgrading Approaches and Economic Fac-
tors, lists the various uprating and upgrading approaches with qualitative comparisons
of each concept.
While readers of this chapter are encouraged to have a background in underground cable
systems, the various sections provide a general overview so that those readers who are
unfamiliar with underground technologies may also come away with an understanding
and be able to utilize some of the technologies discussed on their own cable systems.
Readers are also encouraged to review appropriate industry standards and guides from
Chapter 3: Underground Cables Increased Power Flow Guidebook
3-2
the Institute of Electrical and Electronics Engineers
(IEEE), Association of Edison Illuminating Companies
(AEIC), International Electrotechnical Commission
(IEC), Insulated Cable Engineers Association (ICEA)
and others. In addition to providing valuable back-
ground on cable manufacturing, rating and installation
practices, these documents also include supporting
information about testing (type, routine and commis-
sioning) that may be done in conjunction with uprating
activities.
3.2 CABLE SYSTEM TYPES
Underground transmission is often used to transfer
power where overhead lines are impractical. Issues that
affect the selection of underground cables vary but are
generally focused around reduced rights-of-way require-
ments, aesthetics, and minimizing the environmental
impact associated with installing transmission systems.
There are three major types of cables systems:
High-pressure fluid-filled (HPFF) or gas-filled
(HPGF), pipe-type
Extruded dielectric (XD), including cross-linked
polyethylene (XLPE) and ethylene-propylene-rubber
(EPR) cable types
Self-contained liquid-filled (SCLF) or self-contained
oil-filled (SCOF)
When evaluating uprating and upgrading strategies for
underground cable systems, it is first important to con-
sider the unique characteristics of each of these cable
system types. This section describes the construction
features and operational characteristics of each cable
system.
3.2.1 High-Pressure Pipe-Type
(Fluid- and Gas-Filled)
Cable Construction
Pipe-type cables incorporate three cable phases installed
in a common steel pipe (see Figure 3.2-1). Each cable
phase consists of a stranded copper or aluminum con-
ductor, with a layer of metallic (steel or copper) binder
tapes intercalated with a carbonized black paper tape.
Larger conductors above 800 mm
2
(1500 kcmil) may be
segmented to reduce ac resistance, and hence reduce ac
losses. Over the conductor shield is a laminated Kraft
paper or, for higher voltages, a laminated paper-
polypropylene insulation. The insulation thickness is
governed by voltage. Typical AEIC insulation thick-
nesses are listed in Table 3.2-1.
The insulation wall thickness is important when evalu-
ating reconductoring options or for considering the free
area within the pipe for circulating dielectric liquid.
These concepts are discussed in Section 3.7.
Table 3.2-1 Typical Pipe Cable Insulation Thicknesses
Rated kV Phase-to-Phase Size of Conductors Insulation Thickness
(kcmil) (mm
2
)
Laminated Paper
Polypropylene
a
mils
(mm)
a. Large conductor sizes using laminated paper polypropylene insulation may require increased insulation wall
thicknesses to control the minimum electrical stress to 1750 volts/mil (68.9 kV/mm) so as not to exceed the design
limits of terminals and splices.
Paper
mils (mm)
69 167.8-4000 85 2027 n.a.
270 (6.86)
300
b
(7.62
b
)
115
350-750
800-4000
177-380
405-2027
250 (6.35)
250 (6.35)
420 (10.67)
375 (9.53)
485
b
(12.32
b
)
120
350-750
800-4000
177-380
405-2,027
n.a.
435 (11.05)
405 (10.29)
138
500-900
1000-4000
253-456
507-2027
300 (7.62)
270 (6.86)
490 (12.45)
440 (11.18)
585
b
(14.86
b
)
161
759-900
1000-4000
380-456
507-2027
n.a.
575 (14.61)
515 (13.08)
230
1000-2000
2250-4000
507-1013
1140-2027
450 (11.43)
745 (18.92)
605 (15.37)
345
1000-1250
1500-4000
507-633
760-2027 600 (15.24)
1020 (25.91)
905 (22.99)
500 2000-4000 1013-2027 745 (18.92) 1100 (27.94)
765 2000-4000 1013-2027 1200 (30.48) n.a.
b. High-pressure gas-filled (HPGF) cable insulation thicknesses.
3-3
Increased Power Flow Guidebook Chapter 3: Underground Cables
Over the insulation, it is common to see one or two met-
alized Mylar tapes applied over a carbonized black
paper tape. The Mylar tape acts as a moisture seal to
limit insulation contamination and dielectric liquid
drainage prior to installation. Metal shield tapes are
then applied over the Mylar tape. Over the shield and
moisture barrier tapes is one or two helical metal skid
wires, typically constructed of stainless steel, zinc, brass,
or bronze. The skid wires provide mechanical protection
when the three cables are pulled into the installed cable
pipe. On a few cable designs, a plastic compression
jacket is applied over the insulation shield (more often
on HPGF cables than HPFF cables) to limit the insula-
tion impregnate from draining from the insulation and
mixing with the dielectric media within the cable pipe.
Cable Pipe
The pipe is generally ASTM A-523 Schedule 20 or 40
line pipe, 6.35 mm (-in.) wall with flared ends to facili-
tate welding with chill rings. A cable trench is excavated
for the installation of the cable pipe. Typically, the trench
is usually backfilled with thermal sand or a Fluidized
Thermal Backfill (FTB) that helps ensure good heat
transfer away from the cable pipes (Figure 3.2-2).
Joints
Pipe cables may be 32 km (20 miles) long, but most
installations are only a few kilometers (miles). Installa-
tion sections are on the order of 350-1000 m (1200-
3300 ft) and require manholes and joints to connect
cable sections. Cable pipes enter both ends of the man-
hole to facilitate joining the cables. Inside manholes,
section casing lengths of 1-1.5 m (3-5 ft) generally 1.5-
3 times the cable pipe diameterare used to connect
pipe sections. Inside the casing, each cable phase is
joined together using a compression connector and
hand-applied paper or laminated-paper-polypropylene
tapes. Joints may be one of three types:
Normal Joint. The cable conductors are connected
through the casing, and hydraulic continuity is per-
mitted (see Figure 3.2-3).
Semi-Stop Joint. The cable conductors are connected
through the casing and hydraulic flow is stopped for
differential pressures below 350 kPa (50 psi). Valves
may allow complete hydraulic isolation from one side
of the joint to the other.
Figure 3.2-1 Example of high-pressure fluid-
filled (HPFF) pipe-type cable.
Figure 3.2-2 Pipe-type cable trench being backfilled with
FTB.
Figure 3.2-3 Pipe-type cable manhole with joint casing.
3-4
Full Stop Joint. The cable conductors are connected
through the casing, but there is no hydraulic continu-
ity as the full stop joint supports rated line pressure
differential.
Terminations (Potheads)
The ends of a pipe-type circuit are terminated with a
graded insulation that controls electrical stress from the
paper-insulated cable to the air-insulated terminal. A
cone of insulation is applied within a porcelain termi-
nation to provide hydraulic and electrical isolation for
the cable end. Leading up to the terminations, the three
cables within the common cable pipe are separated into
individual stainless steel pipes through a trifurcating
joint (see Figure 3.2-4). Nonmagnetic stainless steel
pipes are used between the trifurcating joint and the ter-
mination to avoid the high circulating currents and eddy
current heating that would otherwise result if conven-
tional carbon steel pipe were used. Stand-off insulators
are used at the base of the potheads to isolate the pot-
head from the support structure so that circulating cur-
rents are not induced in the riser pipes between the
trifurcator joint and pothead.
Fluid-Filled Cables
High-pressure fluid-filled (HPFF, also known as high-
pressure oil-filled) cables are installed in cable pipes
where the pipe is filled with very clean, very low mois-
ture dielectric fluid. Older HPFF cable systems (before
1970) typically used mineral oil for the pipe filling
dielectric fluid. HPFF cable systems installed after 1970
have used alkyl benzene or polybutene dielectric fluid
(polybiphenyl chlorine-based liquids were never used as
an insulating liquid in pipe cables). The dielectric fluid is
pressurized to 1400 kPa (200 psi) and is generally free to
mix with the insulation impregnant, although this
movement is limited.
Gas-Filled Cables
High-pressure gas-filled (HPGF) cables use pressurized
dry nitrogen gas inside the cable pipe. HPGF cables still
utilize dielectric-fluid impregnated into paper insulating
tapes as insulation, but the dielectric fluid is generally of
a much higher viscosity than fluid-filled cables to limit
drainage. Also, the insulation thickness on HPGF
cables is slightly greater than in HPFF cables as shown
in Table 3.2-1. Nitrogen pressure is typically on the
order of 1400 kPa (200 psig). Bottled nitrogen and a
pressure regulator located near the terminal ends are
used to maintain the pressure within the cable pipe.
Low-pressure alarms are utilized to ensure that the
cable pipes are maintained at the required pressure to
avoid damaging the pipe cable.
Other Equipment
Pumping Plants
As was mentioned above, pipe-type cables are pressur-
ized with either dry nitrogen or dielectric liquid. For the
liquid-filled cables, a pumping plant or pressuriza-
tion plant is needed to maintain and regulate the typi-
cally 1400 kPa (200 psi) pressure within the cable pipe
(Figure 3.2-5).
Cathodic Protection Equipment
The carbon steel pipe must be protected from corrosion
to avoid leaks and deterioration of the pipe. The first
level of protection is a corrosion protection layer that is
applied over the outside of the pipe. Older cable systems
used a hot applied tar coating or a somastic coating that
is similar to concrete. More recent HPFF cable systems
use pipe that is coated with a polymeric material such as
high-density polyethylene. These corrosion protection
coatings are effective in preventing corrosion if there are
no holes (holidays) or cracks in the coatings. How-
ever, some damage inevitably occurs to the pipe coating
during installation or subsequent digging after the cable
system has been placed in service. Consequently, it is
necessary to further protect the cable pipe with
impressed current cathodic protection systems or sacri-
ficial anodes.
Some HPFF cable system pipes are corrosion protected
with magnesium sacrificial anodes that are connected to
Figure 3.2-4 Above ground trifurcator (spreader head)
and pipe-type cable potheads.
3-5
the pipe at manhole locations as well as at the substa-
tions where the cable terminations are located.
Impressed current cathode protection systems must pro-
vide enough current to maintain the cable pipe at a
potential of 1.0 volt dc (or in some cases higher) with
respect to the surrounding earth. The impressed cur-
rent/pipe grounding system must also be designed to
accommodate the maximum line-to-ground fault cur-
rent while keeping the pipe potential close to ground
potential.
Several types of impressed current systems have been
used to cathodically protect and ground the cable pipe.
These are:
Resistor/Rectifier Cathodic Protection. In this type of
cathodic protection system, the ends of the pipe are
grounded through low resistance connections (several
milliohms), and a relatively high-capacity dc current
supply forces enough current through the resistor to
maintain the dc pipe potential at approximately -0.85 to
-1.0 V.
Polarization Cells with Rectifiers. In this type of cathodic
protection system, a passive device called a polarization
cell is used to ground the cable pipe at the end point
substations. A polarization cell, which is about the size
of a car battery, is characterized by a relatively high
resistance to dc voltages of several volts and a low resis-
tance to ac currents. A relatively low-capacity dc recti-
fier then supplies enough current to the pipe to maintain
the pipe potential at -0.85 to -1.0 V.
Solid-State Pipe Grounding Devices with Rectifiers. The
most recent development for pipe cathodic protection
includes power electronic devices that are capable of
conducting line-to-ground fault currents. These devices
are direct replacements for the polarization cells
described above.
3.2.2 Extruded Dielectric
Cable Construction
Extruded dielectric cables are so named because the insu-
lation is extruded onto the conductor core (Figure 3.2-6),
as compared to paper-insulated cables (HPFF or SCFF),
where the insulation is a laminar application of paper.
The cables consist of a stranded copper or aluminum
conductor. Larger conductors above 800 mm
2
(1500
kcmil) may be segmented to reduce ac impedance, and
Figure 3.2-5 Pumping plant pressure charts and control system (left) and fluid reservoir (right).
3-6
hence reduce ac losses. The typical extruded insulation
types are as follows:
Cross-linked Polyethylene (XLPE). This cable insula-
tion is the most common on modern XD cable sys-
tems with applications up to 500 kV. The insulation is
cross-linked (vulcanized), forming long polymer
chains that are joined to one another at intermediate
carbon atoms. XLPE cable manufacturing is
extremely sensitive to cleanliness and quality control
during the manufacturing process. True triple extru-
sion process, where the conductor shield, insulation,
and insulation shield are extruded together, is the
standard extrusion process. This ensures that there is
good adhesion between the insulation-shield bound-
aries and limits the likelihood that contaminants get
into the insulation. XLPE insulation is extremely sen-
sitive to the presence of moisture or water, which
could lead to water and electrical trees.
Ethylene-Propylene-Rubber (EPR). This insulation
type is often considered for distribution cables and
transmission cables up to 138 kV. The insulation is
very lossy as compared to XLPE insulation, result-
ing in high dielectric losses and charging current. For
this reason, EPR insulation is rarely considered for
higher voltage cable systems. The insulation material
is relatively forgiving about operation in the presence
of water, so that the cable may not include a hermetic
moisture barrier like the one required on XLPE
cables.
Linear Low or Medium Density Polyethylene (LLPE,
MDPE). This insulation type is less common for new
installations, although there are several installations,
predominantly in France. As compared to XLPE
insulation, LLPE and MDPE were first used at the
higher voltage levels because the extrudate could be
raised to higher temperatures without forming cross-
linking agents present. This permitted filtering the
extrudate at higher temperatures with finer-grade
mesh screens. The disadvantage is that the maximum
normal operating temperature for cables with LLPE
and MDPE is much lower, limiting the power trans-
fer for a given conductor size as compared to an oth-
erwise similar XLPE cable.
Typical insulation thicknesses for extruded insulation
are summarized in Table 3.2-2.
After the conductor and insulation shields and insula-
tion are applied, the outer layers of an extruded cable
vary depending on the application. All XLPE cables
and most other extruded transmission cables have some
type of metallic sheath consisting of a lead extrusion,
corrugated copper, aluminum or stainless steel, or cop-
per or aluminum foil laminate. There may be additional
copper or aluminum wires applied under the moisture
barrier for additional fault current capability. Over the
metallic moisture barrier is typically an extruded insu-
lating jacket of linear low, medium, or high-density
polyethylene. The jacket electrically isolates the metallic
moisture barrier to control circulating currents and
insulate to induced voltages. The jacket also provides
corrosion protection. Extruded dielectric cables do not
utilize any dielectric liquid, which has increased the use
of this cable type as it represents a low-maintenance
cable system with minimal potential for environmental
impact.
Installation Considerations
Most extruded transmission cables are manufactured as
single phases that are connected into three-phase sys-
tems (there are three-core extruded cables that may be
used for transmission voltages, but these are less com-
mon). The cable phases may be installed directly buried
in the ground or pulled into conduits. Direct buried
installations are generally less expensive and have
slightly better ampacity capability for a given conductor
size, but require that the trench for an entire installation
section be open, which is generally not possible with
urban installations. Duct bank installations are installed
in similar manners as pipe-type cables in that the con-
duit system is installed first and backfilled with concrete,
thermal sand, or Fluidized Thermal Backfill (FTB)
(Figure 3.2-7). The cables are then pulled into the con-
duits later. The cost for duct bank installations is nor-
mally greater than direct buried, both for materials and
installation, but later changes to the cable system do not
require surface excavation.
Figure 3.2-6 Example extruded dielectric (XLPE)
transmission cable.
3-7
Extruded dielectric cables may also be installed under
water. The submarine installations often utilize a
stranded armor around the outside of the cables but are
otherwise similar in construction to land cables.
Joints
As with pipe-type cables, extruded cables are typically
installed in sections of 300850 m (1000-2800 ft) that
must be joined together. For direct buried systems, the
joints may be directly buried and backfilled with
thermal sand. Duct bank installations utilize manholes
similar to pipe-type cables. Older joint technology for
XLPE cables utilized field-molded joints where
unvulcanized polyethylene tapes were applied around a
connector and cables and then heated and vulcanized
using specialized equipment. This approach required
additional time and was more sensitive to workmanship
than the present technology that utilizes pre-
fabricated joints where the cable ends are inserted into
the factory-molded joint (Figure 3.2-8).
Each cable phase is joined separately. Depending on the
sheath bonding scheme (single-point bonded, cross-
bonded, or multi-point bondedsee Section 3.6.8 for a
Table 3.2-2 Typical Extruded Cable Insulation Thicknesses
Rated kV Phase-
to-Phase
Size of Conductors Insulation Thickness
(kcmil) (mm
2
)
XLPE
mils (mm)
LLPE
mils (mm)
EPR
mils (mm)
63-70 500-2000 253-1013
650 (16.5)
a
433 (11.0)
b
a. AEIC CS7-1993 full wall insulation thicknesses.
b. Typical minimum insulation thicknesses used on commercial cable.
650 (16.5)
433 (11.0)
650 (16.5)
110-120
474
4,000
240
2000
800 (20.3)
a
620 (15.7)
b
709 (18)
630 (16)
800 (20.3)
132-138 750-3000 380-1520
850 (21.6)
a
650 (16.5)
b
850 (21.6)
150-161
592
3,947
300
2000
24.5
21.2
19.5
19.5
220-230
789
3,947
400
2,000
906 (23)
c
866 (22)
906 (23)
866 (22)
330-400 1243-2368 630-1200 1063 (27) 1063 (27)
500 2000-4000 1013-2027 1260 (32) 1260 (32)
c. Initial applications of 230-kV XLPE cable in North America used insulation thicknesses of approximately 29-mm
(1142 mils).
Figure 3.2-7 Duct bank installation for extruded dielectric
cables.
Figure 3.2-8 Factory pre-molded joint for XLPE cable
(Click-Fit joint from Pirelli).
3-8
detailed discussion), the joint may have sheath inter-
rupts built into the joint. This allows electrical isolation
of the sheath from one side of the joint to the other and
permits transpositions of sheath connections in a cross-
bonding link box or connection of the ungrounded
sheath end to a sheath voltage limiter.
The cable joints are often placed on racks on the sides of
the manhole and mechanically supported so that the
joint does not move as a result of thermal-mechanical
bending. One distinction between pipe-type cable man-
holes and manholes for other cable types is that parallel
pipe circuits may pass through the same manhole while
XLPE cables typically have individual manholes for each
circuit unless all the XLPE circuits connect to a com-
mon bus at the terminals (e.g., as shown for the manhole
in Figure 3.2-9). Pipe circuits can pass through a com-
mon manhole since utilities generally allow work on one
pipe circuit while a parallel circuit is still energized.
Extruded cables, lacking the robust steel cable pipe and
joint casing, are generally not operated this way both for
concerns about faults and the possibility of induced
voltages and currents from nearby parallel circuits.
Terminations
Terminations (terminators) provide a means of grading
the high electrical stress in the cable insulation to an air-
insulation where the cable is connected to other equip-
ment (Figure 3.2-10). Most termination designs for
extruded dielectric transmission cables utilize a small
amount (200400 liters, 50100 gallons) of silicone oil.
The metallic and semiconducting shields are removed
from the outside of the insulation, and the insulation is
sanded smooth prior to installing a stress cone and plac-
ing a polymer or porcelain termination housing over the
prepared cable end.
The terminations typically have polymer or porcelain
stand-off insulators that isolate the base plate from the
support structure. For cable systems with cross-bonded
or single-point bonded sheaths, the stand-off insulators
control circulating currents. During maintenance, the
stand-off insulators allow the cable jacket to undergo a
10-kV dc test to insure cable jacket integrity.
3.2.3 Self-Contained Liquid-Filled (SCLF)
Cable Construction
Self-contained liquid-filled (SCLF) (also known as self-
contained oil-filled [SCOF] or low-pressure oil-filled
[LPOF]) cables utilize the dielectric liquid-impregnated
laminated-paper insulation similar to pipe-type cables,
but three separate cables are installed for the three
phases. The cable is called fluid-filled because there is
a hollow fluid channel in the center of the conductor
that allows dielectric liquid to move through the cable
with thermal expansion and contraction (Figure
3.2-11). As compared to pipe-type cables, the SCLF
cable typically operates at a lower pressure of 105
525 kPa (1575 psi). Two different methods are used to
construct the stranded hollow conductors. Smaller con-
ductors typically consist of a helical steel support for the
fluid channel, over which is applied the conductor
Figure 3.2-9 Extruded dielectric cable manhole.
Figure 3.2-10 Termination for a 145-kV class
extruded dielectric cable.
3-9
strands (usually copper but may be aluminum). Larger
conductors are manufactured by using keystone
shaped copper strands to form the fluid channel, and
the helically-applied round strands are then applied over
the first layer of formed strands. Large conductors (i.e.,
1000 mm
2
or larger) are typically assembled with 4, 5, or
6 segments that are brought together around the fluid
channel. For larger conductor sizes, a binder tape (met-
allized paper tape, carbon black tape, or copper or steel
tape) is applied around the outside of the high-voltage
conductors to maintain the circular construction, par-
ticularly around segmented conductors. A carbonized
black paper tape may be intercalated with the binder
tape for a conductor shield. Many layers of insulating
tapes are then wound around the cable conductor using
a tape lapping machine in a low-humidity environ-
ment (see Figure 3.2-12). Both conventional cellulose
and laminated paper polypropylene (LPP) tapes are
used to insulate SCLF cables. The cable cores are then
impregnated with low-viscosity dielectric liquid, such as
alkyl- or dodecyl-benzene. Over the insulation, carbon
black paper tapes and metallic shield tapes are applied.
A metallic sheath is then applied around the outside of
the cable core to facilitate pressurization of the cable
system and to exclude moisture. The cable sheaths are
typically tubular lead or corrugated aluminum or cop-
per. In some cases, lead sheaths must be reinforced with
nonmagnetic metal tapes to withstand the cable fluid
pressure.
On some SCLF submarine cables, a concentric copper
conductor may be installed to reduce induced sheath
current and armor wire losses. This is common on sub-
marine cables where there may be large phase spacing,
resulting in induced currents that approach the magni-
tude of the phase currents. An insulating plastic jacket,
similar to extruded cables, is applied over the metallic
sheath.
Installation Considerations
SCLF cables are typically installed directly buried or
suspended from the walls of underground tunnels, par-
ticularly in Europe and Asia. Concrete-encased duct
bank installations have been the most common installa-
tion method used in North America. SCLF cables are
frequently used for submarine cable installations
because SCFF cables can be manufactured in long
lengths (greater than 15 km or 9 miles) without factory
or field joints.
One of the significant differences between the installa-
tion of SCLF cables and other types of transmission
cables is that dielectric fluid pressure must be maintained
at all times. The cables are shipped from the factory on
reels with integral pressure reservoirs to maintain the
fluid pressure during shipment and storage.
Joints
SCLF cable joints are, in general, more complicated
than joints for other cable types. Two different types of
joints are required for most SCLF transmission cable
circuits that have elevation differences greater than
150 m (500 ft) or that are longer than 5 km (3 miles) in
length.
The normal, or straight-through, joint is designed to
electrically connect the conductor between two adjacent
cable sections and to provide hydraulic continuity
between the two cable sections. The splice casing must
also act as a pressure vessel and in some cases electri-
Figure 3.2-11 Example self-contained
liquid-filled cable.
Figure 3.2-12 SCLF tape lapping machine for paper cable.
3-10
cally isolate the metallic cable sheaths to eliminate
induced sheath currents.
Stop joints are designed to withstand dielectric fluid
pressure between the two sides of the joint in order to
hydraulically isolate adjacent cable sections. In this case,
the joints are fitted with hydraulic connections to
nearby fluid pressurization reservoirs.
The installation of SCLF cable joints requires a high
level of skill by the cable jointer (Figure 3.2-13). These
special skills include tape lapping and conductor joining
(while maintaining a continuous flow of dielectric fluid).
Lead wiping of the joint casing to the cable sheath also
requires special skills.
Terminations
SCLF cable terminations require the installation of an
impregnated paper stress cone similar to that of HPFF
cables to maintain acceptable electrical stresses at the
point where the cable insulation shield is terminated.
The terminations typically have epoxy or porcelain
stand-off insulators that isolate the base plate from the
support structure. For cable systems with cross-bonded
or single-point bonded sheaths, the stand-off insulators
control circulating currents. During maintenance, the
stand-off insulators allow the cable jacket to undergo a
10-kV dc test to ensure cable jacket integrity.
Other Equipment
Fluid Reservoirs
SCLF cables commonly utilize fluid reservoirs on one
or both ends of the circuit and sometimes at intermedi-
ate locations along the cable route. Pumping plants or
pressurization plants similar to pipe-type systems are
also used for very large installations. The fluid reservoirs
consist of a nitrogen-pressurized bladder that expands
or compresses against a sealed dielectric liquid con-
tainer, allowing the dielectric liquid in the cables to
expand and contract with load cycling. Reservoirs for
SCLF cables are sometimes gravity fed, as shown in Fig-
ure 3.2-14.
3.2.4 Other Cable Types
There are other cable types, but most are less common
for transmission applications than pipe-type, extruded
dielectric, or self-contained.
Mass Impregnated (MI) or Paper Insulated Lead Cov-
ered (PILC) cables are sometimes used up to 69 kV for
ac systems, although they are not that common at this
Figure 3.2-13 Jointing a self-contained liquid-
filled cable.
Figure 3.2-14 SCLF fluid reservoirs.
3-11
voltage. Uprating approaches would be somewhat simi-
lar to those of extruded or self-contained cables. These
cables have paper tapes that are impregnated with a
high-viscosity dielectric fluid. They do not have external
pressurization systems, as in the case of HPFF or SCLF
cable systems. MI cables are used for ac applications but
are more common for HVDC submarine applications
where there may be a significant change in elevation
along the cable route that would otherwise be compli-
cated by hydrostatic head pressures. A moisture barrier
on the outside of the MI cable prevents moisture ingress.
Compressed Gas Insulated Transmission (CGIT) lines
have many similarities to isolated phase bus or gas insu-
lated substation (GIS) equipment. A system of SF
6
gas
and epoxy insulators are used to insulate a hollow, rigid
aluminum conductor from a tubular aluminum enclo-
sure. Factory-assembled elbows are required to accom-
modate turns in the cable route, and sections with
bellows in the enclosure are required to accommodate
expansion and contraction. A CGIT bus is generally not
buried because the aluminum enclosure is highly suscep-
tible to corrosion, although some manufacturers in the
early 21
st
century are promoting an insulated enclosure
for buried applications. The most common applications
for CGIT lines are situations where very high ampaci-
ties are required (i.e., > 2000 A), usually to connect with
overhead lines entering a station or as a high-capacity
bus within a station.
3.3 POWER FLOW LIMITS AND SYSTEM
CONSIDERATIONS
Generally, this section considers methods to improve
the thermal capacity, and therefore the rating, of under-
ground cable systems. However, transmission circuits
may be limited by factors external to the cable circuit
being considered. Specifically, system considerations
may not allow a cable circuit to be fully loaded to its
thermal capacity. This section describes these factors.
3.3.1 Thermal, Stability, and Surge Impedance
Loading Limits
Transmission circuits, in general, may be constrained
based upon one of three limitsthermal, stability, and
surge impedance loadingeach of which is summarized
in the sections that follow.
Thermal Limits
All transmission cables are limited by thermal consider-
ations. Weak link analysis applies to underground
cable ratings where the section of cable that has the
worst thermal conditions limits the overall circuit
capacity. The causes of these thermal limits will be dis-
cussed in more detail in Section 3.4.5. Sections of the
route that result in the cable having a higher operating
temperature for a given load condition will limit the
overall line capacity. These limits may result from
greater soil thermal resistivity, deeper burial depth,
higher ambient soil temperature, or possibly mutual
heating from multiple cables in the same trench or other
external heat sources. Conceivably, conditions along a
few meters of cable may limit the rating for several kilo-
meters of an underground line.
Cables also have much higher charging current per unit
length as compared to overhead lines, and the charging
current must pass through the conductor. Although the
cable charging current is 90 out of phase from the cur-
rent for real power transfer, it significantly reduces the
amount of real power that may flow through the cable
conductor for some underground lines. The cable charg-
ing current per unit length is given by Equation 3.3-1.
3.3-1
Where:
f is the power frequency.
C is the capacitance in Farads per meter.
E
0
is the line-to-ground voltage.
is the dielectric constant.
The natural log term contains the insulation and con-
ductor shield diameters.
In determining the allowable real power transfer for a
cable circuit, the square of the capacitive current (leads
the real current by 90) is subtracted from the square of
the maximum allowable current (e.g., the normal rating
or ampacity), as described in Equation 3.3-2.
3.3-2
Cable circuits are always limited by thermal constraints,
which is generally consistent with overhead lines that
are less than 80 km (50 miles) in length. Although eco-
nomic considerations constrain the lengths for under-
ground cables, the maximum length ac cable circuits
between shunt compensation reactors are technically
limited by the charging current. The charging current
increases proportionally with length and represents
actual amperes that flow through the cable.
Figure 3.3-1 shows the magnitude of the charging cur-
rent for cable circuits with the charging from one end
and from both ends of the line.
] Amperes [ 10
ln 18
2
2
9 0
0 arg

=
CONDUCTOR
INSULATION
D
D
ing Ch
E f
fCE =
I

Amperes] [
2
arg
2
Re ing Ch Maximum al
I I I =
3-12
As the length of the underground line is increased, a
point is reached where the total charging current equals
the cable ampacity. This occurs at what is called the
critical length of the cables, where no real power may
be transferred without overloading the cable circuit. The
critical length can be calculated from Equation 3.3-3.
3.3-3
Where:
I
Normal
is equal to the normal ampacity.
I
Charging
is equal to the charging current per meter.
It is obvious that the maximum feasible line length must
be significantly less than the critical length in order to
transmit reasonable amounts of real power. However,
the concept of critical line length quantifies the absolute
maximum lengths between shunt compensation that can
be achieved for different types of underground cables.
Table 3.3-1 shows critical line lengths based on typical
insulation thicknesses and parameters.
In real situations, economic considerations, rather than
the critical length considerations, limit the construction
length for most land cables. However, charging current
limitations have been a significant factor for long sub-
marine cable circuits. In these cases, owners sometimes
consider using HVDC cables where there are only resis-
tive losses to consider and the cost of the HVDC valves
and convertor stations is more easily justified relative to
the cable cost. In some other cases, shunt compensation
reactors were installed on intermediate islands.
Note that reactive compensation can mitigate cable
charging current effects on the system to which the
cables are connected, but the charging current still flows
in the cables, potentially limiting the real power transfer.
It is common practice to install shunt reactors at one or
both ends of relatively long underground transmission
lines to mitigate the effects of cable-generated charging
current on the power system. The amount of shunt com-
pensation depends upon the ability of the power system
to absorb the reactive MVARs generated by the cables
during light load conditions. The amount of leading
reactive MVARs that generation units can absorb is
generally less that the amount of lagging MVARs that
they can generate. During normal and peak load condi-
tions, the leading MVARs generated by the cable capac-
itance are typically offset by lagging power factor loads.
Consequently, high-voltage load-switching devices (i.e.,
circuit switchers) are typically used to disconnect shunt
compensation reactors during periods of relatively high
load. The reactors are then connected to the power sys-
tem during periods of light load. The reduction of trans-
mission system losses is another consideration in sizing
shunt compensation reactors for underground transmis-
Figure 3.3-1 Charging current magnitude profile.
[Meters] Length Critical
arging Ch
Normal
I
I
=
Table 3.3-1 Critical Lengths for Underground Cable Circuits
with Typical Insulation Thicknesses
a
a. Assuming 3158 kcmil (1600 mm
2
) segmental copper con-
ductor.
Voltage Kraft Paper
Laminated
Paper
Polypropy-
lene
Cross-
Linked
Polyethyl-
ene
Ethylene-
Propylene-
Rubber
Critical Lengths, miles (km)
69 kV 63 (101) n.a.
b
b. Laminated paper-polypropylene is not generally used for
voltages below 161 kV.
163 (262) 120 (193)
115 kV 48 (77) n.a.
b
130 (209) 82 (132)
138 kV 45 (72) n.a.
b
120 (193) 70 (113)
161 kV 42 (67) 52 (84) 114 (183) n.a.
c
c. The relatively high dielectric constant and dissipation
factor for EPR insulation limit the application of these
cables to 138 kV or below.
230 kV 30 (48) 36 (58) 81 (130) n.a.
c

345 kV-400
kV
20 (32) 29 (47) 53 (85) n.a.
c

500 kV n.a.
d
d. At this voltage, with conventional Kraft paper insula-
tion, the dielectric losses are so great that the ampacity is
zero.
23 (37) 47 (76) n.a.
c

3-13
sion lines. Typically, the shunt reactors are chosen so
that the lagging MVARs created by the shunt reactors
are between 60% and 100% of the leading MVARs gen-
erated by the cable capacitance. In most cases, a series of
load flow cases must be performed for light and heavy
load conditions to determine the optimum amount and
location for cable system shunt compensation.
Voltage Profile and Stability Limits
Some overhead lines may have voltage regulation prob-
lems (i.e., excessive voltage drop) when transmitting
power to lagging power factor loads. As the line length
increases, the voltage on the line tends to drop. This can
cause problems in transferring power from the sending
end to the receiving end of a line.
The relatively high capacitance of cable circuits may
have adverse effects on the voltage profile in the vicinity
of underground transmission lines. Since a capacitive
current flowing through an inductance causes a voltage
rise across the inductor, the charging current created by
the capacitance of a cable circuit can cause high system
voltage by two related phenomena. The first and most
common situation is for the cable charging current to
cause voltage rises across the inductive impedances
external to the cable circuits. This situation, which is
worst during light load conditions, is illustrated in Fig-
ure 3.3-2.
In this circuit, the reactances, X
L
and X
T
, represent
inductive impedances of transmission lines and trans-
former impedances between the generator and the cable
circuit. Since the voltages at the load must be held near
rated voltage, the system voltages rise significantly
above the rated voltage as the cable charging current
flows through the inductive impedances, X
L
and X
T
, to
the generation units. During light load conditions
(assumed in Figure 3.3-2), this voltage rise may cause
voltages above the maximum rated voltage that most
power system components are designed for (105%) with-
out shunt compensation.
Cable circuits are not generally associated with voltage
stability problems because of their capacitive nature
they are generally shorter as compared to overhead
lines, are naturally compensated, and generate voltage-
supporting vars.
Surge Impedance Loading (SIL) Limits
Surge impedance loading (SIL) limits involve a greater
than allowable phase shift in power frequency from one
end of a transmission system to the other. As a result,
the two ends of the system cannot remain synchronous,
resulting in instability and outages. This system stability
consideration is generally an issue on overhead trans-
mission lines that are 80-320 km (50-200 miles) in
length.
The positive sequence surge impedance, Z
S
, of a trans-
mission line is defined by Equation 3.3-4
3.3-4
where, L and C are the positive sequence series induc-
tance and shunt capacitance, respectively.
Cables have lower series inductances and much higher
shunt capacitances compared to overhead lines. Conse-
quentially, cables have very low surge impedance relative
to overhead lines. Typical surge impedances for over-
head and underground lines are 375 ohms and 30 ohms,
respectively.
The SIL limit, based on the line-to-line voltage (V
LL
), is
defined by Equation 3.3-5
3.3-5
SIL power transfer limits are rarely a problem for
underground transmission lines because of their low
surge impedance and relatively short lengths.
Figure 3.3-2 Voltage rise due to cable charging current.
Ohms
C
L
Z
S
=
MVA Limit SIL
2
S
LL
Z
V
=
3-14
3.3.2 Load Flow Considerations
Low cable system series impedances, and the resulting
unequal distribution of load flow between overhead and
underground lines, are important considerations when
evaluating the uprating potential on underground cable
circuits.
The reactive loading of various transmission lines in a
power system is controlled by the magnitude of the volt-
ages across the system and can be adjusted by generator
excitation and transformer taps. The flow of real power
over the lines is a function of the relative voltage angles
around the system and the interconnecting impedances.
Unfortunately, the distribution of real power flow is not
as easily controlled as reactive power flow because the
circuit impedances are fixed in most cases, and it is not
economical to control power flows by changing angles
at the generation units. Therefore, where transmission
line loadings are not approximately equal to the thermal
capacities of the circuits, the power transfer cannot be
increased once a circuit is loaded to its thermal limit
even though the other circuits may be lightly loaded.
Because underground cables have much lower series
impedances than overhead lines, a higher portion of
power will flow over the underground lines compared to
overhead lines in the same area. Figure 3.3-3 shows an
extreme situation where an underground cable is con-
nected electrically in parallel with an overhead line.
The sending and receiving end phasor voltages are the
same for both overhead and underground circuits. Then
the power flow between the two buses along the cable
and line is defined by Equations 3.3-6 and 3.3-7.
3.3-6
3.3-7
Since the series impedance of cables is typically 25-30%
of length compared to those for overhead lines, the
power flow along the cable circuit is greater, perhaps
exceeding the ampacity of the cable circuit or resulting
in underutilization of the overhead line. While the above
example is extreme, cables may be put in parallel with
overhead lines indirectly in a conventional power sys-
tem. Is some cases, relatively expensive phase shifting
transformer must be placed in series with underground
cables to more evenly distribute power flows with over-
head lines in the same area.
3.3.3 Uprating Hybrid (Underground and
Overhead) Circuits
Hybrid transmission circuits contain sections of both
overhead lines and underground cables. The reasons for
these types of installations are numerous (rights-of-way
congestion, available ROW, water crossings, tunnels, air-
ports, etc.), but the issues with uprating these types of
circuits are complicated by considering all of the equip-
ment along the circuit. Overhead lines are typically
operating at only 30-40% of their rated capacity, and in
hybrid circuits usually have a normal rating that is 40-
60% greater than the connected underground sections.
As a result, the cable section is usually studied first from
the standpoint of increasing a circuits capacity since the
circuit will be limited by the section with the lowest rat-
ing often the cable.
From the standpoint of typical design limits, an over-
head line usually can obtain 1 ampere of capacity for
each kcmil (2 amperes per square millimeter), while a
cable will generally have half that capacity. Also, over-
head lines do not suffer from mutual heating effects
among phases or circuits, but this is a significant consid-
eration where cables are installed in the same trench.
Therefore, additional overhead conductors added to
increase capacity cannot be equally matched by the
same number of underground conductors, since the bur-
ied cables will experience a net de-rating from mutual
heating.
While overhead lines typically have greater ampacity for
normal operating conditions, the large thermal time
constant of buried transmission cables typically 50-
150 hours compared with overhead lines (5-15 min-
utes) means that for short-duration emergencies, cables
typically have a higher emergency rating. Also, because
of the relatively short time constant of overhead lines,
the load cycle on the overhead line does not have a sig-
nificant impact on the normal or emergency capacity.
However, with buried cables, the long thermal time con-
stant has a big impact on ratings. This is factored into
daily ratings by considering a 24-hour loss factor (essen-
Figure 3.3-3 Simplified power system network with
parallel overhead and underground circuits.
( )
1 2
2 1
= Sin
Z
V V
P
Cable
Cable
( )
1 2
2 1
= Sin
Z
V V
P
Line OH
Line OH
3-15
tially, the daily load factor of the losses). This is dis-
cussed in greater detail in Section 3.4.
3.4 UNDERGROUND CABLE RATINGS
3.4.1 Introduction
This section provides a brief overview of the ampacity
procedures used to determine cable ratings. While there
are subtle differences in constructions among the vari-
ous cable types, this section focuses on the most com-
mon constructions to provide the reader with enough
background to understand how the uprating approaches
discussed in later sections can be applied.
Many ampacity calculations are based on a 1957 paper
by Neher and McGrath (Neher and McGrath 1957).
Later work by CIGR documented an ampacity proce-
dure into an international standard (International Elec-
trotechnical Commission 1982, 1987) that provides a
step-wise approach to calculating ampacity based upon
cable construction. The two calculation approaches give
similar results, although their treatment of daily load
cycles is different.
The paper by Neher and McGrath assumes a sinusoidal
load shape and uses a 24-hour (daily) loss factor to
account for an overall averaging effect of heat output
from the cable beyond a certain diameter (called D
X
).
Within this diameter, the temperature rise across the
thermal resistances in the cable and nearby soil is pro-
portional to the peak heat output from the cable. At dis-
tances greater than this diameter away from the cable
center, the temperature rise is proportional to the aver-
age daily heat output.
Many system planners are familiar with a load factor,
which relates the peak load to the average daily load. In
cable systems, the focus is on heat outputa function of
I
2
Rso the term daily loss factor is used, which is
essentially the load factor of the losses as defined by
Equation 3.4-1.
3.4-1
The loss factor can be estimated from the load factor
using an empirical relationship. One such relationship is
as shown in Equation 3.4-2.
3.4-2
Di stri buti on cabl es someti mes fol l ow a si mi l ar
relationship, Loss Factor = 0.2 (Load Factor) + 0.8
(Load Factor)
2
, but it is important to note that these
load factor-to-loss factor relationships are specific to a
given system and should not necessarily be applied
arbitrarily. The IEC method in its basic form assumes
that the load remains constant (e.g., a loss factor of
100%), but ratings by this method can be modified by a
companion document, IEC-60853 (International Elec-
trotechnical Commission 1989), to account for the daily
load variations, particularly when the load is not closely
sinusoidal. The IEC approach to account for load varia-
tions does not easily lend itself to hand calculations and
generally gives close agreement to the Neher and
McGrath method under typical loading patterns.
The approach described in this section is focused on the
use of the IEC methodgenerally recognized as provid-
ing better accuracy for ac losses and being easier to fol-
lowbut accounts for cyclic loading by using the loss
factor approach from the paper by Neher and
McGrath.
Chapter 3 does not specifically address emergency rat-
ings, except conceptually in later sections. For the typi-
cal user of this chapter, both normal and emergency
rating calculations can be handled by a computer pro-
gram such as UTWorkstation and DTCR that are avail-
able from EPRI, as well as other commercial sources.
For additional background on emergency rating calcu-
lations, the reader should review relevant references
listed at the end of this chapter.
3.4.2 Concept of Ampacity
An underground cable circuit rating, or ampacity, is
the solution to a basic heat transfer problem. Heat gen-
erated in the cable is removed by thermal conduction to
ambient earth and, ultimately, air. Engineers familiar
with Ohms Law know that electrical current flowing
through an resistance will produce a voltage drop (or
rise) according to Equation 3.4-3.
3.4-3
An analogous relationship may be used to describe ther-
mal conduction where heat flowing through a thermal
resistance produces a temperature drop (or rise) accord-
ing to Equation 3.4-4.
3.4-4
This basic concept is extended to model heat out of a
buried cable through the various cable layers, trench
backfill and native earth.
2
max
24
1
2
24
Factor Loss
I
I
i
i
=

=
( ) ( )
2
Factor Load 7 . 0 Factor Load 3 . 0 Factor Loss + =
AC
R Current Voltage =
Thermal
R Heat e Temperatur =
3-16
3.4.3 Losses
Losses (heat) from a cable are the source of temperature
rise above ambient earth. Heat is generated from both
dielectric heating and ac resistance heating, as described
in the following sections.
Dielectric Heating
Dielectric heating comes from charging and discharging
the insulating dielectric at 50 or 60 times per second.
The dielectric heat loss, W
Dielectric
, can be found from
Equation 3.4-5.
3.4-5
Where:
C =Capacitance, Farads/meter.
f =power frequency, Hz.
=specific inductive capacitance (dielectric
constant).
tan =insulation dissipation factor.
V
l-g
=line to ground voltage applied across the
insulation, volts.
Some typical insulation parameters are listed in Table
3.4-1.
Charging Current
Dielectric losses are generated on a per-meter basis and
affect the radial heat escaping from the cable, ultimately
affecting normal ampacity. Charging currentenergy
used to charge and discharge the insulation at power fre-
quencyis consumed when the cable is energized and
ultimately limits the allowable real power transfer (see
Section 3.3.1).
Electrical Losses
DC Resistance
Electrical losses are generated from current flowing
through the resistance of the phase conductors, concen-
tric metallic shields, sheaths or skid wires, and the pipes
(depending on cable system type). The resistance is a
function of the cross-sectional area of the material and
is defined by Equation 3.4-6.
3.4-6
Where:
= electrical resistivity of metal in ohm-m at
20C.
Area = cross-sectional area of metal in mm
2
.
The dc resistance of the conductor is usually increased
by 2.5% or so to account for the lay of the wires in the
stranded conductor.
The dc resistance can be adjusted to other temperatures,
usually the rated conductor temperature, using Equa-
tion 3.4-7.
3.4-7
Where:
= inferred temperature of zero resistance.
T = temperature of the conductor.
Some typical values of electrical resistivity are listed in
Table 3.4-2, along with their temperature correction
coefficients.
AC Skin and Proximity Effects
Losses in the conductor are affected by self and mutual
inductance; the self inductance causes the current to
concentrate near the conductor surface called con-
ductor skin effect and the magnetic field of neighbor-
[W/meter]
ln 18
10 tan 2
2
9 2
2

= =

Conductor
insulation
g l
g l Dielectric
D
D
V f
fCV W

Table 3.4-1 Typical Cable Insulation Material Parameters

a
Insulation Material
Dielectric
Constant Dissipation Factor
Range Typical Range Typical
Impregnated Paper 3.3-3.7 3.5
0.002-
0.0025
0.0023
Laminated Paper-
Polypropylene
2.7-2.9 2.7
0.0007-
0.0008
0.0007
Cross-linked Polyethylene 2.1-2.3 2.3
0.0001-
0.0003
0.0001
Ethylene-Propylene-
Rubber
2.5-4.0 3.0 0.002-0.08 0.0035
a. Values from 1992 EPRI Underground Transmission Systems
Reference Book.
r] [Ohms/mete
10
6
20

=
Area
R
dc
] Ohms/meter [
20
20
=
T
R R
dc dcT
Table 3.4-2 Electrical Resistivities and Temperature Factors
for Common Cable Metals
Material
Electrical Resistivity
Ohm-meters
Inferred Temperature
of Zero Resistance, C
Aluminum 2.8264 x 10
-8
-228.1C
Brass 6.317 x 10
-8
-912.0C
Bronze 3.5 x 10
-8
-564.0C
Copper 1.7241 x 10
-8
-234.5C
Lead 21.4 x 10
-8
-236.0C
Stainless Steel 70 x 10
-8
n.a.
a
a. There is almost no variation in the resistance of stainless
steel with respect to temperature over the typical
operating range of a cable.
Zinc 6.633 x 10
-8
-218.7C
3-17
ing conductors affects the distribution of current across
the conductor called conductor proximity effect.
When the total losses from the cable conductor are
being calculated, these two parameters must be deter-
mined. The impact of these effects on the dc resistance
called the ac-dc ratio is a function of the conductor
type and construction. Values of the conductor skin
effect correction factor, ks, and proximity effect correc-
tion factor, kp (also known as the transverse
conductivity factor) are listed in Table 3.4-3.
The ac resistance increment for conductor skin effect,
Y
CS
, can be found from Equation 3.4-8, where k
S
is the
skin effect factor based on the conductor construction.
3.4-8
The ac resistance increment for conductor proximity
effect, Y
CP
, can be found from Equation 3.4-9, where k
P
is the proximity effect factor based on the conductor
construction.
3.4-9
The conductor resistance including skin and proximity
effects can be calculated using Equations 3.4-10 and
3.4-11.
3.4-10
3.4-11
Shield, Sheath and Skidwire Resistance
Depending on the cable type, there are various metallic
layers outside of the cable insulation. A summary of
typical layers is as follows:
Pipe-Type: helically applied metallic tape(s), helically
applied skid wire(s).
SCFF: extruded lead, corrugated copper, or corru-
gated aluminum sheath, possibly with addi-
tional copper wires or tapes.
Extruded: concentric stranded copper or aluminum
wires, copper or aluminum tapes, extruded
lead, corrugated copper, corrugated alumi-
num or corrugated stainless steel.
Regardless of the cable type, the resistance of the metal-
lic layers outside of each cable phase must be evaluated.
These resistance values are then be adjusted for the tem-
perature of the respective layer. Three basic equations
may be used. For a helical metallic layer (e.g., a helically
applied tape or skid wire), Equation 3.4-12 can be used.
3.4-12
Where:
= electrical resistivity of layer in Ohm-
meters.
D
Layer
= average diameter of the helical layer.
Lay
Layer
= distance along the cable for one turn of
the tape or wire (e.g., the lay).
Area = cross-sectional area of the tape, wire or
skid wire.
For an extruded or welded metallic layer (e.g., the
sheath on an extruded or self-contained cable), the area
of the layer can be calculated by knowing the difference
Table 3.4-3 Skin and Proximity Effect Factors for Various
Conductor Types
Conductor Type ks kp
Concentric round, dry 1.0 1.0
Concentric round, in oil 1.0 0.8
Compact round, in oil 1.0 0.6
Compact segmental, dry 0.435 0.6
Compact segmental, in oil 0.435 0.37
Compact segmental, in oil 0.39
0.35 (trefoil),
0.46 (cradled)
Compact segmental (aluminum), in oil 0.35
0.29 (trefoil),
0.36 (cradled)
Hollow core, 6 segment, in oil 0.39 0.33
Hollow core, 6 segment (aluminum),
in oil
0.26
0.19 (trefoil),
0.27 (cradled)
Hollow core, 4 segment, in oil 0.435 0.37
dcT
S
S
R
fk
X
7
10 8

=

2
2
8 . 0 192
S
S
CS
X
X
Y
+
=
dcT
P
P
R
fk
X
7
10 8

+
=
27 . 0
18 . 1
312 . 0
8 . 0 192
2
2
2
2
2
phase
Conductor
phase
Conductor
phase
Conductor
P
P
CP
d
D
d
D
d
D
X
X
Y
( ) soil or air in cables for 1
CP CS dcT acc
Y Y R R + + =
( ) ( ) pipe steel a in cables for 5 . 1 1
CP CS dcT acc
Y Y R R + + =
r] [Ohms/mete 1
10
2
6
20

=

Layer
Layer
dc
Lay
D
Area
R


3-18
in diameters above and below the layer, as shown in
Equation 3.4-13.
3.4-13
For a stranded shield, the area of a single strand must be
calculated and then multiplied by the number, N, of wire
strands. The resistance can then be calculated as above
by dividing the area into the electrical resistivity of the
material (see Equation 3.4-14).
3.4-14
Once the resistances for all of the shield layers are deter-
mined, they should be taken electrically in parallel to
find the overall resistance, as shown in Equation 3.4-15.
3.4-15
Shield Loss Increments for Circulating Currents
Circulating currents exist when a cable metallic shield,
sheath, neutral or skidwire is grounded at both ends
(e.g., multi-point bonding see Section 3.6.8). The
phase currents induce a circulating current in these lay-
ers depending on the geometry and resistance of the
layer that opposed the phase current according to
Lenzs Law. Circulating currents exist in pipe-type
cables because the skidwire and metallic shield are con-
tinuously grounded to the cable pipe. Circulating cur-
rents also exist in single-core cables (e.g., extruded
dielectric or self-contained cable types) when the metal-
lic sheath is grounded at both ends or on cross-bonded
systems where the minor section lengths are not equal in
length.
Calculation of the circulating current increment requires
determining the mutual reactance between adjacent
cable phases and depends on the configuration of the
phases. Equations 3.4-16, 3.4-17, and 3.4-18 should be
applied appropriately, depending on the cable configu-
ration that most closely matches the indicated cable
positions. S, the center-to-center phase spacing, and,
D
S
, the average diameter of the shield/sheath/skidwire
layer are used in the equations.
3.4-16
3.4-17
3.4-18
The incremental increase in resistance from circulating
currents can then be found using Equation 3.4-19.
3.4-19
Shield Loss Increments for Eddy Currents
Eddy current losses occur when a continuous concentric
metallic layer exists around the cable core (e.g., a corru-
gated or extruded metal sheath or longitudinally taped
metallic shield, but not to stranded shields). Also, eddy
currents are negligible for pipe-type cables.
The mechanics for calculating the eddy current losses is
somewhat onerous but not particularly complicated. The
equations to perform these calculations were derived
empirically and are listed in Equations 3.4-20 through
3.4-24. For a more detailed explanation, see IEC-287
(International Electrotechnical Commission 1982).
3.4-20
3.4-21
3.4-22
3.4-23
( )
r] [Ohms/mete
10
] [mm
4
6
20
2
2 2
=
Area
R
D D
Area
dc
Inner Outer
] [mm
4
2
2
strand
D N
Area

=

1 2
1 1 1 1
....
S s s sn
R R R R
= + + +
7
2
4 10 ln [Ohms/meter]
for cables in equilateral (triangular) configuration
m
s
S
X f
D
=

7
2.3
for cables in cradled configuration
m
s
S
X f
D

=

7
2.52
for cables in flat/vertical configuration
m
s
S
X f
D
=

2
1
1
S
SC
acc
S
m
R
Y
R
R
X
=

+

7
'
2 10
S
f
m
R
=
2
2
0
2
1.4 0.7
3.08
1
6
2 1
0.86
2
for flat formation
S
Se
m
S
Se
D m
Y
S m
D
Y m
S
+

=

+

=

( )
2
2
0
2
0.92 1.66
2.45
1
3
2 1
1.14 0.33
2
for trefoil formation
S
Se
m
S
Se
D m
Y
S m
D
Y m
S
+

=

+

= +

( )
2
1
7
1.74
3
1
8
10
1 10 1.6
Shield
Shield
s S
S
f
t
g D
D

= +

3-19
3.4-24
where R
S
is the resistance of only the layer(s) where
eddy currents may occur and t
Shield
is the thickness of
the shield layer.
AC Resistance Including Circulating Current and Eddy
Current Increments
The conductor resistance including skin, proximity, cir-
culating current, and eddy current effects can be calcu-
lated using Equations 3.4-25 and 3.4-26.
3.4-25
3.4-26
Pipe Loss Increments for Pipe-Type Cables
Pipe cables have additional losses from eddy current and
hysteresis heating in the cable pipe as a result of the ac
cables within the pipe. The increment losses associated
with the pipe can be calculated using Equations 3.4-27
and 3.4-28.
3.4-27
3.4-28
Note that the pipe loss increment is greater for cables in
cradled configuration, so cradled configuration should
generally be assumed unless the cables are known to lay
in close trefoil configuration.
Equations 3.4-27 and 3.4-28 are for 60 Hz. If the power
system is 50 Hz, the values of Y
P
should be multiplied
by 0.76.
Ac resistance including skin, proximity, circulating cur-
rent, eddy current, and pipe loss effects is determined
using Equation 3.4-29.
3.4-29
Other Losses
Additional losses may be experienced in some cable
installations. Single-core submarine cables may have
steel or other armor wires that can have both circulating
current and hysteresis (for iron-based armor) losses.
Also, utilities use steel casings for road crossing or direc-
tionally drilled installations that can have significant
hysteresis and eddy current losses. Neutral continuity
conductors may generate heat if the phase currents are
imbalanced, but this is most often an issue on distribu-
tion systems.
3.4.4 Equivalent Thermal Circuit and Thermal
Resistances
As was discussed in Section 3.4.2, thermal conduction is
the principal means by which heat leaves a buried cable
system. The thermal equivalent to Ohms Law is applied
to this model by developing an equivalent thermal cir-
cuit that describes the various layers of cable materials
and surrounding earth which heat must pass through to
reach ambient. Thermal circuits for extruded/self-con-
tained and pipe-type cables are shown in Figures 3.4-1
and 3.4-2. The thermal resistances and heat sources are
shown, analogous to electrical resistances and current
sources in a conventional Ohms Law circuit. The tem-
perature of the conductor, T
c
, is determined by adding
up the temperature rises above the ambient earth tem-
( )
( )
4
1
0 1
12
1
12 10
Shield S
EC s Se Se
acc
t R
Y g Y Y
R

= + +

( ) 1
for cables in air or soil
acs dcT CS CP SC EC
R R Y Y Y Y = + + + +
( ) ( )
1 1.5
for cables in a steel pipe
acs dcT CS CP SC EC
R R Y Y Y Y = + + + +
6
10
0.0438 0.0226
for cables in cradled configuration
Skidwire Pipe
P
dcT
D ID
Y
R
+
=
6
10
0.115 0.01485
for cables in trefoil configuration
Skidwire Pipe
P
dcT
D ID
Y
R

=
( ) ( )
1 1.5
acp dcT CS CP SC EC P
R R Y Y Y Y Y = + + + + +
Figure 3.4-1 Equivalent thermal circuit for extruded
dielectric and self-contained fluid-filled cable types.
Figure 3.4-2 Equivalent thermal circuit for pipe-type cables.
3-20
perature, T
earth
, as heat passes through the various ther-
mal resistances.
The thermal capacitances shown in the figures account
for the fact that changing load (e.g., heat output) from
the cable does not instantaneously result in a change in
temperature. This is consistent with the Ohms Law anal-
ogy, where a change in applied voltage does not instantly
change the voltage across an electrical capacitor.
Temperature Rise
To calculate the ampacity, the conductor temperature is
determined for a given current and checked against the
maximum allowable conductor temperature based on
the insulation material. For the above thermal circuits,
the temperature rise above ambient is found using
Equation 3.4-30, where the ac losses (I
2
R) and dielectric
heat losses (W
d
) pass through the various thermal resis-
tances to give a temperature rise:
3.4-30
Typical ampacity calculations assume that the voltage
remains constant, so dielectric losses are fixed. Then,
the only unknown in Equation 3.4-30 is the current as
shown in Equation 3.4-31 (as a convention, thermal
quantities are shown with an over-line to distinguish
these values from electrical quantities):
3.4-31
Cable Thermal Resistances
The thermal resistances for each cable layer out to the
earth interface where the cable contacts the soil must
be determined to complete the thermal circuit. The ther-
mal resistivities of common cable materials are listed in
Table 3.4-4.
For layers where the inner and outer diameters are con-
centric (e.g., cable insulation, cable jackets or pipe coat-
ings, or the thermal resistance of a conduit), the thermal
resistances may be found using the Equation 3.4-32
3.4-32
where n is the number of cables within the diameter,
D
Inner
.
The thermal resistance between a cable and the inside of
a conduit or pipe can be determined using Equations
3.4-33.
3.4-33
Where:
T
mean
= mean temperature of the duct air, or nitro-
gen gas or dielectric liquid in the pipe.
D
cable
= outer diameter of the cable.
n = number of cables within the conduit or pipe.
U, V, and Y as defined in Table 3.4-5.
Earth Thermal Resistances
The cable thermal resistances describe components of
the thermal circuit out to the interface with the sur-
rounding soil. For direct buried cables, this is the pipe
coating (on pipe-type cables) or the jacket on single-
conductor cables. Single core cables installed in conduits
have an earth interface at the outside of the conduit. At
that point, the earth portion of the thermal circuit
starts.
Three thermal resistances must be considered in the
earth portion of the thermal circuit:
Thermal resistance to heat escaping from the cable or
pipe itself
( )
2
max
2
max
ac d thermal ambient
d ambient ac thermal
T I R W R T
T T T T I R R
= + +
= =
ac thermal
T
I
R R
ln [C-m/Watt]
2
Outer
Inner
D
R n
D

=

Table 3.4-4 Thermal Resistivities of Common Cable
Materials
Material
Range
(C-m/Watt)
Typical
(C-m/Watt)
Impregnated Paper 5.0-6.0 6.0
Laminated Paper-Polypropylene 5.0-6.0 6.0
Crosslinked Polyethylene 3.5-4.0 3.5
Ethylene-Propylene-Rubber 4.5-5.0 4.5
Somastic 1.0 1.0
Transite 2.0 2.0
PVC 4.0-4.5 4.0
Neoprene 3.8-5.8 4.0
Epoxy 0.7-4.45 1.0
Thermoplastic Pipe Coating 3.5-4.5 4.0
( )
[ / ]
1 0.1
mean cable
n U
R C m Watt
V Y T D
=
+ +
Table 3.4-5 Pipe and Duct Thermal Resistance Constants
Configuration U V Y
Fiber Duct or PVC in Concrete 5.2 0.91 0.010
Asbestos Cement in Concrete 5.2 1.1 0.011
Pipe-Type, HPGF 0.95 0.46 0.0021
Pipe-Type, HPFF 0.26 0.0 0.0026
Earthenware Ducts 1.87 0.28 0.0036
3-21
Thermal resistance resulting from mutual heating
among other cables or pipes
Thermal resistance correction to account for native
soil materials outside of the trench having a higher
thermal resistivity than backfill materials
The equations in this section describe each of these. The
basis for these calculations is application of superposi-
tion of the heat fields generated by each buried cable
and the assumption that the earths surface may be
treated as an isotherm. These assumptions are substan-
tiated by the Kennelly Hypothesis, and the reader is
encouraged to review relevant references (Anders 1997)
for more details.
The thermal resistivity of the soil is an important
parameter, which will be discussed in detail later. Table
3.4-6 summarizes some typical values of soils and back-
fills.
Earth Thermal Resistance
The earth thermal resistance is mainly a function of
cable burial depth. For installations where the cable is
installed in a special backfill, the thermal resistivity of
the special backfilldirectly in contact with the cable
should be used for the earth thermal resistance calcula-
tion. The fact that the native soil outside of the trench
may have a greater thermal resistivity will be corrected,
as shown in Equation 3.4-34.
3.4-34
Where:
L = centerline burial depth of the center of the
cable, pipe or conduit, mm.
D
earth
= diameter of the earth interface (cable OD,
pipe OD, or conduit OD), mm.
= the thermal resistivity of the soil in contact
with the cable, C-m/Watt.
n = number of cables within diameter, D
earth.
LF = the 24-hour loss factor for the load on the
cables, per-unit.
D
x
= diameter where average daily heat output
applies, as defined below.
Diameter, D
x
, is the diameter beyond which 24-hour
average ac heat losses apply and is a function of the soil
thermal diffusivity. An empirical relationship for finding
the soil thermal diffusivity,
soil
, as a function of the
native soil thermal resistivity, , is as shown in
Equation 3.4-35.
3.4-35
From this, the diameter, D
x
, may be found (typically
about 210 mm or 8.3 in.), using Equation 3.4-36.
3.4-36
Mutual Heating Thermal Resistance
When multiple cables are installed in the ground, heat
generated by each cable impacts the temperature of the
other cables. For the purposes of this section, all cables
are treated as though they are carrying equal loading
such that the heat output from each cable is the same.
With the assumption that the earths surface is an iso-
therm, a method of images is used to model the heat
leaving each cable and its mutual heating effects on the
other cables. Figure 3.4-3 shows examples for a two-pipe
Table 3.4-6 Typical Soil Thermal Resistivity Values
Soil Type
Thermal Resistivity
5% Moisture
(C-m/Watt)
0% Moisture
(C-m/Watt)
Fluidized Thermal Backfill 0.4 0.75
Concrete 0.6 0.8
Stone Screenings 0.4 1.0
Thermal Sand 0.5 1.0
Uniform Sand 0.7 2.0
Clay 1.0 2.5
Lake Bottom 1.0 (50% moisture) >3.0
Highly Organic Soil >3.0 >6.0
2 2
ln
[C-m/Watt]
2
2 4
ln
x
earth
earth
earth
x
D
LF
D
R n
L L D
D

+

=

+

native
( )
/hour] [mm
100
10 71 . 6
2
8 . 0
4
=
native
soil
1.02 24 [mm]
x soil
D =
Figure 3.4-3 Illustration of mutual heating effects.
3-22
cable circuit (left) and for a single extruded dielectric or
self-contained cable circuit (right).
The mutual heating effect is evaluated knowing the dis-
tance from a given cable to the image of an adjacent
cable divided by the actual distance separating the
cables, as shown in Equation 3.4-37.
3.4-37
with N being the number of conduits, cable pipes, or
cable positions with energized cables. It is important to
evaluate the mutual heating effect for the cable that will
run at the highest temperature to ensure that none of the
cables exceeds the maximum allowable temperature.
Then, the thermal resistance from mutual heating can
be evaluated using Equation 3.4-38.
3.4-38
where n is the number of energized cables at each loca-
tion. For a pipe-type cable, n equals three. For direct
buried cables or conduit installations of transmission
cables, n generally equals one. If the cables are installed
in special backfill, the thermal resistivity used to calcu-
late the mutual heating thermal resistance should be
that of the backfill.
Thermal Resistance Correction for Native Soil
Up to this point, the thermal resistances of the earth
and mutual heating have used a value of the backfill
thermal resistivity. However, in actual installations, the
thermal resistivity of the soil outside of the trench is typ-
ically greater, and this must be factored into the ampac-
ity calculations. To make this correction, an additional
thermal resistance term is needed. The approach consid-
ers the length and width of the backfill envelope, as
shown in Figure 3.4-4.
If the short, x, and long, y, dimensions of the backfill
envelope are known, it is possible to determine a cir-
cumscribing circle having a diameter, D
b
, with the same
volume of backfill material as the rectangular backfill
envelope, as shown in Equation 3.4-39.
3.4-39
Then, if the center-line depth, L
b
, of the backfill enve-
lope is known, a geometric factor for the backfill can be
calculated, as shown in Equation 3.4-40.
3.4-40
The thermal resistance correction for native soil with
different thermal resistivity,
native
, than the backfill
thermal resistivity,
backfill
, can be calculated using the
above geometric correction factor as shown in Equation
3.4-41.
3.4-41
where n is the number of cables per pipe, conduit or
location, and N is the number of pipes, conduits or loca-
tions within the backfill envelope.
The procedure described above is valid as long as the
ratio of the long dimension of special backfill to the
short dimension of the special backfill is 3 or less. A
paper by El-Kady and Horrocks (1995) describes geom-
etry factors that may be used for special backfill enve-
lopes with dimensions outside the acceptable range of
the method used above.
Temperature Rise from Dielectric Heating
The ac and thermal resistances complete the equivalent
thermal circuit so that ampacity can be calculated.
However, it is also important to consider the tempera-
ture rise caused by dielectric heating. Losses are gener-
ated throughout the insulating dielectric, but for the
purposes of ampacity calculations, the dielectric losses,
W
d
, are assumed to enter the thermal circuit half way
through the insulation. This can be seen by referring
back to the thermal circuit figures.
13 1 12
12 13 1
' ' '
...
N
N
d d d
F
d d d

=

( ) ln [ / ]
2
mutual
R n LF F C m Watt
=
( )
2
2
1 4
exp ln 1 ln
2
b
x x y
D x
y y x

= + +

Figure 3.4-4 Trench backfill width and height.
2 2
2 4
ln
b b b
b
b
L L D
G
D

+

=

[ / ]
2
native backfill
correction b
R n N LF G C m Watt

=
3-23
The thermal resistances used for dielectric temperature
rise are calculated in the same manner as for ac losses,
except that the loss factor is 100% (1.0 per unit) since
the voltage is constant. Therefore, it is necessary to con-
sider the thermal resistance values used for dielectric
temperature separately from those used for ac losses,
which are a function of the daily load variation.
The temperature rise caused by dielectric heating can be
determined as shown in Equations 3.4-42 through
3.4-44.
For extruded and self-contained cables in conduit:
3.4-42
For direct buried extruded and self-contained cables:
3.4-43
For pipe-type cables:
3.4-44
Ambient Soil Temperature
Ambient soil temperature is the temperature at the
burial depth of the cable in the absence of any non-
native heat sources. These temperatures are usually
established during a route thermal survey (described in
Section 3.6.1). Some typical values of ambient soil tem-
perature are listed in Table 3.4-7.
3.4.5 Calculating Ampacity
Once the thermal circuit is complete and the dielectric
temperature rise is known, it is possible to determine the
allowable conductor temperature rise for ac loading
(ampacity). The temperature at the conductor is high-
est, and the temperature of the insulation nearest the
conductor limits the ampacity. Some industry-accepted
maximum allowable conductor temperatures are listed
in Table 3.4-8.
The summation of electrical and thermal resistances
(units of the summation of electrical and thermal
resistances are C/ampere
2
) is determined using the
thermal circuit parameters, as shown in Equations 3.4-
45 through 3.4-47.
For extruded and self-contained cables in conduit:
3.4-45
1
2
insulation jacket
d d jacket duct duct
earth mutual correction
R R
T W R R
R R R

+

= + +

+ + +

1
2
insulation jacket
d d
R R
T W
R R R

+

=

+ + +

+ + + +
+
=

correction mutual earth ng pipe coati
pipe cable insulation
d d
R R R R
R R
W T
2
1
Table 3.4-7 Example Ambient Soil Temperatures at Typical
Installation Depths
a
a. E.g., 1.1 m (42 in.).
Location
Maximum
Summer
Maximum
Winter
Atlanta 25C 20C
Boston 22C 18C
Chicago 22C 18C
Denver 22C 17C
Honolulu 30C 15C
Johannesburg 22C 15C
London 18C 8C
Miami 30C 25C
New York 25C 18C
Palo Alto 22C 20C
Singapore 30C 25C
Table 3.4-8 Industry-Accepted Maximum Conductor
Temperatures
Insulation
Material
Maximum
Normal (Con-
tinuous)
Temperature
Maximum Emergency
Temperature
Impregnated
Paper
85C
105C peak at end: up to 100 hours
100C peak at end: 100-300 hours
Laminated
Paper
Polypropy-
lene
85C
105C peak at end: up to 100 hours
100C peak at end: 100-300 hours
Cross-Linked
Polyethylene
90C
105-130C
a
, 72 hours continuous
b
105C- peak at end of 100-300 hours
a. Depends on shield construction and agreement from
manufacturer.
b. Based on Association of Edison Illuminating Companies
CS-7 standard.
Ethylene
Propylene
Rubber
90C 130C
Linear Low
Density
Polyethylene
75C 90C
( ) AC th acc insulation
jacket jacket duct duct
acs
R R R R
R R R
R
R R R
=

+ +

+

+ + +

3-24
For direct buried extruded and self-contained cables:
3.4-46
For pipe-type cables:
3.4-47
Then, the ampacity may be calculated, as shown in
Equation 3.4-48.
3.4-48
Appendices 3.1 and 3.2 have worked examples for calcu-
lating ampacity of pipe-type and extruded dielectric
cables.
3.4.6 Effect of Various Parameters on Ampacity
This section briefly highlights the effects of various
cable and installation parameters on ampacity. The
range of variation shown is not applicable to all
installed cable systems, but the trend is generally consis-
tent for all cable installations.
Effect of Burial Depth
As the burial depth increases, the thermal resistance to
heat leaving the cable and reaching the earths surface
also increases (see Equation 3.4-34). As a result, the
ampacity declines with increasing burial depth. This is
illustrated in Figure 3.4-5.
Barring other factors, it is generally better to have a
more shallow burial depth to minimize the earth ther-
mal resistance. However, ambient soil temperatures are
generally greater near the surface (potentially reducing
ampacity) and soils tend to be drier above the water
table (potentially reducing soil moisture content and
increasing thermal resistivity). So, although a shallow
burial depth reduces thermal resistance, other installa-
tion factors must be considered when evaluating ampac-
ity and placement of cables.
Effect of Phase and Circuit Spacing
Mutual heating among cable circuits and cable phases
cause elevated temperatures that reduce the available
temperature rise for ac current. As a result, increased
phase and circuit spacing reduce mutual heating and
tend to increase ampacity. Some considerations for the
mutual heating might be the placement of cable phases
within a duct bank. For example, if a 3x3 duct bank is
being used, the outer duct positions should be filled first
before the center ducts to minimize mutual heating. Fig-
ure 3.4-6 shows the effect of circuit spacing on ampacity.
A similar trend would apply to phase spacing except in
the case of multi-point bonded cables. For multi-point
bonded cables, metallic shield/sheath circulating cur-
rents are generally higher when the phase spacing is
large, particularly for sheath constructions designed for
high fault currents. For this reason, multi-point bonded
cables (mostly at distribution voltages) are placed in a
single conduit or in close-trefoil (triangular) configura-
tion.
Effect of Native Soil and Backfill Thermal Resistivity
Except for conductor size, most of the cable construc-
tion is fixed based upon the voltage class of the cable.
The native soil outside the cable trench is also fixed (and
represents a large part of the total thermal resistance as
seen in the examples at the back of this guide) but must
be factored into the ampacity calculations. Higher soil
thermal resistivity results in a higher thermal resistance
to heat leaving the cable and lower ampacity.
( )
( )
AC th acc insulation
acs jacket earth mutual correction
R R R R
R R R R R
=
+ + + +
( )
( )
( )
correction mutual earth ng pipe coati acp
pipe cable acs
insulation acc th AC
R R R R R
R R
R R R R
+ + + +
+
=
[amperes]
max
th AC
ambient d imum
R R
T T T
I

=
Figure 3.4-5 Graph of ampacity versus burial depth.
Figure 3.4-6 Graph of ampacity versus circuit spacing.
3-25
Soils may also have great variability in soil conditions
over a few meters of cable route, contrary to overhead
lines where weather conditions may apply for a few kilo-
meters. Therefore, it is important to find the worst con-
ditions along the cable route weakest link and
base ratings on that limiting location.
From the standpoint of installations, the installer has
control over the material put back into the trench. Fig-
ure 3.4-7 shows the effect of special backfill placed in
the cable trench as a function of various native soil con-
ditions. Using a good quality thermal backfill (thermal
sand, Fluidized Thermal Backfill FTB) can improve
the allowable current carrying capacity. Section 3.6.1
describes soil testing and backfill materials.
Effect of Ambient Soil Temperature
Like native soil thermal resistivity, there is little control
over the ambient soil temperature, but it must be fac-
tored into the cable ratings. One consideration about the
ambient temperature is the burial depth. Greater tem-
perature extremes are experienced close to the surface,
so summer ambient temperatures will be more signifi-
cant at shallow depths. Also, surface coverings have an
impact on ratings. Soil ambient temperatures may be 3-
5C warmer below asphalt than other areas because of
the increased solar absorptivity of the surface. Figure
3.4-8 shows the impact of ambient temperature on
ampacity.
Effect of Conductor Size and Sheath Bonding
Conductor size directly impacts the allowable current
carrying capacity of a circuit. Larger conductor sizes
allow for more current. Also, the circulating currents
from multi-point bonded cable sheaths cause significant
losses (heat) that reduce ampacity by 20% or more
(depending on factors like sheath construction and
resistance and the phase spacing). Figure 3.4-9 shows
the impact of these parameters on ampacity.
3.4.7 Emergency Ratings
Emergency ratings reflect a temporary increase in circuit
capacity during a contingency. For cables, the emergency
ratings take advantage of a higher allowable conductor
temperature for a period of time (usually not longer than
300 hours) and the heat storage capacity long thermal
time constant of the cable and soil around the cables.
This is somewhat different from overhead transmission
lines where the thermal time constant is measured in
minutes. With buried transmission cables, the thermal
time constant is 35 to 150 hours. Figure 3.4-10 shows an
Figure 3.4-7 Graph of ampacity versus native and
special backfill thermal resistivity.
Figure 3.4-8 Graph of ampacity versus ambient soil
temperature.
Figure 3.4-9 Graph of ampacity versus conductor size as
a function of sheath bonding mode.
Figure 3.4-10 Graph of emergency ampacity versus
emergency duration.
3-26
example of emergency ampacity versus duration for a
230-kV XLPE cable. The normal ampacity at 100% load
factor is also shown on the graph.
The calculation of emergency ratings is described in a
paper by Neher (1963) and IEC-853 (International Elec-
trotechnical Commission 1989). The reader is encour-
aged to review these documents or obtain suitable
software such as EPRIs ACE (Alternative Cable Evalu-
ation) or DTCR (Dynamic Thermal Circuit Rating) to
perform these calculations.
From the standpoint of uprating, the techniques
described in this chapter apply to both normal and
emergency ratings.
3.4.8 Inferring Conductor Temperatures from
Measured Temperatures
Some cable installations have temperature monitoring
using thermocouples or distributed fiber optic tempera-
ture sensing (DFOTS) (both described in Section 3.6.4).
The temperature measurements are commonly available
in one or more locations:
Under the cable jacket (for DFOTS)
Outside of the cable jacket or cable pipe (for thermo-
couples) when direct buried
Duct air temperature (either DFOTS or thermocou-
ples)
Parallel conduit or cable pipe (either DFOTS or ther-
mocouples)
The conductor temperature of a neighboring cable can
be inferred by comparing the measured and calculated
temperatures at one of the above locations until they
agree, usually while varying the inferred native soil ther-
mal resistivity.
During in situ or laboratory soil thermal resistivity mea-
surements using a thermal property analyzer (described
in Section 3.6.2), a constant heat is injected into the soil
(or soil sample) using a thermal probe. A sensitive ther-
mistor is used to monitor changes in temperature over
time while the heat is being injected into the soil. The
slope of the change in temperature with respect to time
shows the soil thermal resistivity.
For a buried cable circuit, it is possible to infer the soil
thermal resistivity using measured cable temperatures
essentially making the cable a thermal probe. Although
the loading (heat input) is not constant, a reasonable
assessment of the soil thermal resistivity can be deter-
mined and is often a better indication of the effective
soil thermal resistivity seen by the power cable (e.g.,
including effects of moisture migration, etc.) over the in
situ measurements or soil samples collected some dis-
tance away from the cables.
By using the equivalent thermal circuit to calculate the
temperature at a location where measurements are
made, the effective soil thermal resistivity can be
adjusted until the calculated values match the measured
values. The effective soil thermal resistivity, combined
with historical load datausually 2-4 weekscan be
used to find the temperature of a conductor in the
trench. A more elegant approach is to use the dynamic
rating model described in Section 3.8 of this chapter.
3.5 UPRATING AND UPGRADING
CONSTRAINTS
Underground cable systems have unique characteristics
that must be considered when exploring uprating
options. This section describes the characteristics of each
cable system and installation conditions that may limit
uprating options and that should be considered before
exploring any of the uprating methods described later.
3.5.1 Direct Buried Cable Systems
Direct buried cable systems have several limitations for
uprating and upgrading mainly due to the fact that the
cable is relatively inaccessible, there is generally no
opportunity to provide active cooling, and the civil
works cost would eliminate most practical reconductor-
ing options. A hot spot along the cable route that is
identified and of relatively short length might be miti-
gated. Spot temperature monitoring might also be con-
sidered, but retrofitting continuous monitoring is just as
impractical as reconductoring. Soil remediation using a
lower thermal resistivity soil or, for very short distances,
applying heat pipes, could mitigate a hot spot that is
limiting a circuit.
If overburden has developed above the cables and this is
determined to be the cause of a hot spot, the overburden
could be removed to reduce the cable burial depth and
thereby improve ampacity.
3.5.2 Fluid-Filled Cable Systems
The pipe in a pipe-type cable system offers the greatest
flexibility for uprating and upgrading an underground
cable system of any cable type. While the pumping or
pressurization plant generally requires greater mainte-
nance than the other cable types, particularly extruded,
the ability to circulate the dielectric liquid (for HPGF,
fill the pipe with dielectric liquid and then circulate)
allows for both thermal smoothing of a hot spot or
forced cooling.
3-27
The pipe size may be a limiting factor because it can
constrain the voltage upgrading (requiring thicker insu-
lation wall), or ampere upgrading (requiring a larger
conductor) may be limited.
Note that mitigating a hot spot on a pipe-type or self-
contained fluid-filled cable system, allowing a greater
overall ampacity, may mean that other sections of the
circuit will be operating at a higher temperature. This
would generally result in greater fluid expulsion in pipe
type and, in particular, SCFF cables, which may require
that the pressurization plant be recalibrated, that the
fluid reservoir tanks be resized, or the nitrogen blanket
pressure above the oil adjusted to accommodate the
larger volume of dielectric liquid. Alarm settings might
also need to be adjusted.
3.5.3 Duct Bank Installations
Duct bank installations do not offer significant options
for uprating. The concrete encasement that is common
to duct banks generally has good thermal resistivity.
The impact of overburden on the circuit might be inves-
tigated and, in extreme cases, the native soil around a
section of duct bank might be replaced with lower ther-
mal resistivity material.
Often a problem associated with duct banks is the rela-
tively high congestion of cable circuits in those loca-
tions. Sometimes expensive transmission cables are
significantly derated by a low-cost and relatively very
low power transfer distribution cable. One possible
remediation method could be to remove the distribution
circuits entirely, or replace the distribution cables with a
larger conductor or multiple cables per phase to reduce
the heat output and mutual heating effects.
Duct bank installations allow for upgrading options in
that existing cables may be removed and new ones
installed relatively easily. If a particular section of a cable
route is found to limit overall ampacity and a section of
cable with a larger conductor is available, that particular
section could be replaced to mitigate the hot spot.
Ducts may also be filled with water or grouted with a
low-thermal resistivity grout to improve thermal con-
duction between the cable surface and conduit.
3.5.4 Trenchless Installations
Trenchless installations horizontal directional drilling
(HDD), pipe jacking, or microtunneling have some of
the limitations of the direct buried system such as being
essentially inaccessible. However, trenchless installa-
tions typically use inner ducts some with spare pipes
or conduits that may be used to mitigate hot spots and
improve ampacity. The inner ducts allow a cable phase
to be removed and replaced, possibly with a larger con-
ductor size. To improve ampacity, the casing in a trench-
less installation may be filled with water or a low-
thermal resistivity grout to improve heat transfer away
from the cables and conduits. Filling the annular space
between the inner ducts and the casing will allow heat to
pass through a low thermal resistivity grout (0.8 C-
m/Watt) or water (1.6 C-m/Watt) rather than air
(45 C-m/Watt).
3.5.5 Other Installation Locations
Tunnels
Tunnel installations offer some unique considerations
from the standpoint of uprating. First, the cables are
generally more accessible than for buried cable systems,
so this more easily facilitates maintenance and repairs.
However, tunnel installations do not benefit as much
from the long thermal time constant of direct buried,
pipe, or conduit installations because the cables are
essentially installed in air. Some tunnel installations
may be limited by the maximum allowable air tempera-
ture from the standpoint of work safety or heating/ven-
tilation/air conditioning (HVAC) limits. Depending on
the tunnel configuration, it may be difficult to add
forced air ventilation to the tunnel to improve capacity.
Also, extruded or self-contained cables that are racked
in troughs or by mechanical supports could be subjected
to damage from thermal-mechanical bending (TMB)
effects during increased temperature operation.
Deep Installations Including Water Crossings
Somewhat like trenchless installations described above,
plowing in cables or cable pipes in water crossings
makes installation of a specialized thermal backfill
impractical. Also, mitigating a hot spot at one of these
locations is not easily accomplished because there is lit-
tle control over the sedimentation that can build above
the cables. The best approach is to properly account for
the soil conditions both thermal resistivity and ambi-
ent temperatures in developing ratings.
Overhead (In Air) Installations
Cables installed in air, for example on bridges or risers,
do not benefit from the long earth thermal time con-
stant of buried cable systems. In most cases, the in-air
normal ampacity is greater than the buried normal
ampacity.
However, the cables may be exposed to high ambient air
temperatures and possibly solar radiation and also lack
the long earth thermal time constant of buried cables,
possibly limiting the emergency rating capacity of in-air
sections. Mitigating ampacity limits for in-air sections
may be difficult, particularly if solar radiation is an issue
since shielding the cables from solar heating can be
impractical.
3-28
3.5.6 Hot Spot Identification
Hot spot identification is sometimes difficult to do
because of restrictions on accessing the cables. Distrib-
uted temperature sensing (DTS) fiberthe state-of-the-
art for cable temperature monitoringis often difficult
to retrofit. More traditional approaches to finding hot
spots involve studying plan and profile drawings and
evaluating the impact of other known heat-producing
services in the area around cables.
If DTS can be applied, the hot spots can be readily iden-
tified, although the causes may not always be obvious.
3.5.7 Accessories
Joints
For the most part, joints do not normally limit a cable
circuits ampacity, particularly when the joints are
installed in manholes that often run cooler than the
direct buried soil sections. However, directly-buried
joints on pipe-type cables may limit ampacity because of
the added layers of hand-applied paper tape insulation
that are used to construct the joint, adding additional
thermal resistance to heat leaving the joint. Forced-
cooled pipe-type cables may result in a thermal limit at
joints. The forced-cooling fluid circulation essentially
makes the surface of the cable an isotherm. With the
added insulation applied over the joint resulting in
increased thermal resistance, the connector may be run-
ning hotter, limiting the rating.
Terminations
In general, terminations are not the limiting factor for
uprating cable systems except on forced-cooled pipe-
type cables or in areas with a very high ambient air tem-
perature. In forced-cooled pipe systems, the dielectric
liquid circulation cannot cool the termination, so the
terminations rather than the cable sections become a
limiting factor. EPRI report EL-2233 on high-capacity
terminations gives additional details about this subject
(EPRI 1982).
3.5.8 Hydraulic Circuit
In pipe-type cables, the hydraulic circuit may limit
uprating opportunities under some circumstances.
For example, if there is no return pipe or there is only
one pipe circuit, it would not be possible to circulate
dielectric liquid within the pipe. Some utilities are reluc-
tant to use fluid circulation where the hydraulic circuit
involves moving liquid down one energized cable pipe
and back another, mainly because if a failure occurred,
the healthy circuit could be contaminated from
byproducts of the fault. Although there will be less
mutual heating with the parallel circuit out of service,
the rating on the healthy circuit will probably drop
because fluid circulation cannot continue.
A small amount of fluid movement might be achieved
by oscillating the dielectric liquid within the pipe using
the 750011,000 liter (20003000 gallon) fluid reservoir
tanks on either end of the circuit if at least 3,750 liters
(1000 gallons) of additional dielectric fluid can be
accommodated in the existing reservoirs on each end.
The extent of pipe fillingdegree to which the cables fill
the free area of the pipemay also limit dielectric liquid
circulation rates because the pressure drop between
pumping plants may be too great even when using a low
viscosity dielectric liquid.
3.6 INCREASING THE AMPACITY OF
UNDERGROUND CABLES
Once there is an understanding of the possible limita-
tions associated with each cable type, it is necessary to
consider how uprating might occur on a given circuit.
This section describes various techniques that may be
applied to investigate ampacity limitations and then
ways to improve ampacity, or at least have a better
understanding of what is limiting the ampacity.
3.6.1 Route Thermal Survey
A route thermal survey has traditionally involved evalu-
ating the entire cable route in a detailed manner to
understand ampacity limitations. Many North Ameri-
can utilities adhere to Association of Edison Illuminat-
ing (AEIC) standards regarding cable design. One of the
principles of these standards is that if the soil character-
istics are not well known, the design ampacity should be
based upon a maximum operating temperature that is
10C below the allowable operating temperature (e.g.,
values in Table 3.4-8).
Regardless of following the AEIC standards or not, util-
ities have sometimes designed cable circuits without a
good knowledge of the route characteristics, particu-
larly on older circuits. In these cases, the ambient soil
temperature and soil thermal resistivity were not well
known, so assumed values were often incorporated into
rating calculations. Those following the AEIC guide-
lines obtained some additional conservatism in the rat-
ings by using the lower 10C operating temperature in
the event that the assumed parameters were inaccurate.
However, as the circuits age and load growth continues,
many utilities are revisiting the rating assumptions to
see if additional transmission capacity is available with-
out major investment in infrastructure.
Also, during the process of uprating a cable circuit, hot
spot mitigation may require removing existing trench
3-29
backfill materials and replacing it with a good quality
thermal backfill.
The following subsections discuss some of the tech-
niques employed for a route thermal survey and
describe soil and backfill characteristics that are impor-
tant to consider in evaluating methods for uprating
cable systems.
Thermal Property Analysis
In the equivalent thermal circuit, the earth thermal
resistances are the largest component, typically repre-
senting over 50% of the total thermal resistance. They
are also the least understood. As compared with over-
head lines, where weather parameters (wind speed and
direction, solar radiation, temperature) may be valid for
1-2 km of line length, soil characteristics along under-
ground cable routes can vary over a few meters. If the
cables are buried in city streets, there exists a strong pos-
sibility of encountering borrowed fill instead of native
soils. These fills may satisfy civil/construction require-
ments, but if topsoil, cinders, or organic soils are used,
the thermal performance may be very poor. For this rea-
son, it is very important to test the soils so that appro-
priate values of thermal resistivity may be used in design
calculations.
Thermal property analysis based on transient heat flow
was first suggested as early as 1888 (Wiedman 1888).
During the mid-1900s, significant research and other
work were conducted in North America (Mason and
Kurtz 1952; Blackwell 1954; Carslaw and Jaeger 1959).
This demonstrated the practical use of a thermal needle
line heat source method. The Insulated Conductors
Committee, organized in 1947, performed a special
project on soil thermal resistivity in 1951. A special sub-
committee (No. 14) headed by Professor H. F. Win-
terkorn of Princeton University continued work in this
field for 10 years and published the AIEE Committee
Report in 1960.
In the 1970s, EPRI-sponsored research resulted in the
design and development of the Thermal Property Ana-
lyzer. The basic approach was to develop a portable,
fully automated test instrument with standardized test-
ing procedure that could be employed for both field and
laboratory with results that could be extended to power
cable systems.
Thermal Resistivity
Thermal resistivity, sometimes call rho, is a property
of a material. In the contents of cable installation and
field measurements, the thermal resistivity is measured
for a soil or trench backfill. The most common
approach to thermal resistivity measurements now is the
transient thermal needle method, which is based on
the line heat source theory.
Essentially, an underground cable is a long distributed
heat source. The transient thermal needle method
takes advantage of this characteristic by using a ther-
mal probe, which contains a heating coil throughout its
length and a thermistor type temperature sensor at the
mid-point of the heater. The length-to-diameter ratio of
the probe is high enough so that end effects do not
impact the measurements. An example thermal probe is
Once the probe is installed in the soil sample or in the
native soil (field), the heater in the thermal probe is
energized with a constant power while the change in
temperature is recorded over time (usually 2030 min).
The slope of the log time-temperature curve is propor-
tional to the thermal resistivity of the soil sample. A
thermal property analyzer (TPA) was developed to
automate this process and is commonly used for both
field and laboratory measurements (Figure 3.6-2).
The transient thermal probe method (e.g., IEEE Stan-
dard 442) is a relatively quick and accurate approach to
measuring soil thermal properties, provided the theoret-
ical assumptions are understood and care is taken in the
test setup to stay within the limits of the theory. The test
assumes various conditions:
Figure 3.6-1 Thermal probe used for field thermal
resistivity measurements.
3-30
The probe is an instantaneous and constant heat
source (no thermal capacitance).
Heat flow is radial.
Conduction is the only mechanism of heat transfer.
No contact resistance exists at the soil/probe interface.
An infinite sample boundary exists.
The test sample is homogeneous and at moisture and
thermal equilibrium.
No moisture migration occurs during the test.
For these assumptions to be valid, it is important that
the probe insertion and testing be performed carefully,
usually by a qualified specialist, to ensure that the
results are valid. Contact resistance is very important
and a critical part of inserting the probe into the soil.
Also, it is important to keep the probe temperature at
reasonable values to avoid drying the soil.
A drill rig with a hollow stem auger is used to drill down
to the required depth for soil sampling and to perform
in situ thermal resistivity measurement tests. Sometimes
a backhoe or hand digging down to the required depth
is also used to access the soil where testing will be done.
In the case where the hole is advanced using a drill rig,
the thermal probe is attached to an extension rod and
then tapped into the native soil at the required depth.
The testing is then performed from the surface (see Fig-
ure 3.6-1).
In addition to field measurements, called in situ mea-
surements, soil samples are collected during soil boring
for detailed laboratory analysis and to evaluate parame-
ters such as dry-out curves (thermal resistivity as a func-
tion of moisture content under constant dry density; see
Figure 3.6-3) or thermal stability that cannot be done
effectively in the field. The samples are collected in thin-
wall Shelby tubes. If the soil is very loose or noncohesive
(granular), a split spoon sampler or large diameter Cali-
fornia sampler with liners is used to collect undisturbed
samples. If soil conditions (granular, very hard or rocky)
are such that undisturbed tube samples cannot be col-
lected, either disturbed bulk samples or auger cuttings
are taken. If bedrock is encountered, core samples of 5
8 cm (23 in.) diameter are collected to be tested in the
laboratory. Standard ASTM procedures should be
implemented for soil boring, sampling, storage, and
transportation.
Borehole logs, which characterize the soil types with
depth, are often made so that if the cable burial depth
varies, the type of native material and its thermal resis-
tivity can be known for rating purposes. A typical bore
hole log is shown in Figure 3.6-4. The borehole log
information may also be used for other geotechnical
purposes such as designing structural loads, laying out
directional drilling route, etc. This geotechnical infor-
mation is very useful for the civil contractor to deter-
mine the type of equipment required for excavation,
Figure 3.6-2 Thermal Property Analyzer (TPA) used for
field and laboratory thermal resistivity measurements.
Figure 3.6-3 Example thermal dry-out curves for various
soil types with pores between soil grain particles
saturated with water (A), and dry (B).
3-31
Figure 3.6-4 Example borehole log.
3-32
dewatering, backfilling, and other activities. The ambi-
ent temperature recorded at the start of an in situ ther-
mal resistivity test is an important value to record for
the cable designer.
In the laboratory, a soil sample is prepared to evaluate
thermal dry-out characteristic the variation in thermal
resistivity of the material as a function of moisture con-
tent. The results of these tests (dry-out curves) are pre-
sented on charts that show thermal resistivities at the in
situ moisture content (if known), at critical moisture
content, and in totally dry condition (worst case). Some
degree of drying beyond the native moisture levels should
be expected in the presence of energized power cables, so
an adjusted soil thermal resistivity that factors in the dry-
ing should be incorporated into ampacity calculations.
Once the soil thermal resistivity results are known, they
can be used in ampacity calculations. For the case of
installed cable systems, it may be necessary to do testing
outside the cable trench to get native conditions, and
within the cable backfill to characterize the special ther-
mal backfill that may have been used around the cables.
If the trench is known to have a common material
throughout, testing of the backfill material may only be
needed at a few selected locations.
Thermal Diffusivity
Although thermal diffusivity is not commonly recorded
(typical transient needle TPA equipment can measure
this parameter) for most applications, its application is in
transient calculations. In simple terms it can be treated
as the inertia in the heat and mass transfer equation.
The three termsresistivity (), diffusivity () and heat
capacity (C)are related by Equation 3.6-1.
3.6-1
Thermal Stability
Thermal stability is a system-driven parameter and is a
soil characteristic that describes how well a soil main-
tains a constant thermal resistivity when exposed to
cable heating. The main issue is to consider if the heat
leaving a cable would result in the soil being below its
critical moisture content, in which case the soil would
experience net drying and an increase in thermal resis-
tivity. Smaller diameter cables with direct contact to the
soil are more likely to result in thermal instability
because of a larger heat flux (temperature gradient) at
the cable-soil interface.
A classic example of a thermally instable material is
modeling clay. The clay can be dried at room tempera-
ture over time. If the dried sample is then placed in
water, it does not readily reabsorb water to return to a
malleable substance. Some soilsincluding soils with
high clay and silt contentshave these characteristics. A
common situation where this may be an issue for power
cables is the use of bentonite as a grout material either
in trenchless casing installations or for cable conduits;
pure bentonite has high thermal resistance and is prone
to drying. Bentonite is prone to shrinkage and cracking
(leaving voids) if drying does occur. A better solution is
to use as much sand as possible while minimizing the
bentonite content, and to seal the ends of the casing or
ducts so that the grout cannot dry.
Moisture Migration in the Soil
For any given soil or backfill, the major influence on the
thermal resistivity is the moisture content. In a dry state,
the pore spaces between soil particles are filled with air
(thermal resistivity of about 45C-m/W). As water (ther-
mal resistivity of about 1.65C-m/W) replaces air, the
soil resistivity decreases substantially by as much as 3 to
7 times, as the good heat conduction paths are expanded
(additional thermal thermal bridges). This is illus-
trated by the thermal dry-out curve (thermal resistiv-
ity vs. soil moisture content) shown in Figure 3.6-3. A
soil that is better able to retain its moisture, as well as
being able to efficiently re-wet when dried, will have bet-
ter thermal performance characteristics. The soil water
content is expressed as a percentage of the weight of
water to the dry weight of soil solids, as determined by
oven drying at 105C.
The heat generated by energized cable tends to cause soil
moisture to migrate away from the cable/backfill inter-
face. In unstable backfills or soils, this drying increases
the resistivity substantially, inducing further heating of
the cable and thus more drying of the soil. Eventually
this cycle may create a totally dry zone of the backfill
around the cable, resulting in excessive heating and
potential thermal runaway. In a stable backfill, the heat-
induced drying raises the resistivity marginally, thus
minimizing the potential for thermal runaway.
Thermal stability is best illustrated by means of thermal
dry-out curves (Figure 3.6-3). The critical moisture is
defined as the moisture content below which the rela-
tively flat nature of the thermal dry-out curve gives way
to a disproportionate increase in the thermal resistivity.
Above the critical moisture a soil will resist thermal dry-
ing (by means of capillary suction), whereas below this
value, thermal runaway is inevitable (unless soil mois-
ture is externally replenished, (i.e., rain).
Although some native soils at high moisture content
(1025%) may exhibit fairly low thermal resistivity (0.4
to 0.6 C-m/W), this value may increase a few fold when
dry. Well-graded sands and stone-dust containing 10
1 C =
3-33
15% fines (-200 Sieve size material) make good correc-
tive thermal backfills.
Cable Route Soil Test Spacing
The soil testing and sampling frequency for thermal
resistivity testing along the cable route can vary depend-
ing on the area and the length of the route. In rural
areas, where the use of fills is minimal and historical
construction has not been significant, sampling and in-
situ testing every 500 m might be done. In urban areas
or locations where fill materials have been used, sam-
pling might be done every 200500 m. Known varia-
tions in geology or other conditions might affect how
often along the route testing and sampling are done.
The goal of testing is to capture test results for any
unique soils and potential hot spots along the cable
route while categorizing where each soil type is found.
Factors Affecting Soil Thermal Resistivity
Soil Composition
The soil composition is an important characteristic
affecting soil thermal resistivity. Soils are typically a
conglomerate of various materials, and the ratio of these
materials within a soil affects the thermal resistivity.
Table 3.6-1 summarizes the dry thermal resistivity val-
ues of various components.
Because the soil components are so important in affect-
ing the thermal resistivity, a good understanding of the
geology along a cable circuit is valuable to assessing
where soil testing should be performed and how much
variation might be expected along a given cable route.
It is important to note from Table 3.6-1 and Figure
3.6-3 that dry soils have a much higher thermal resis-
tance than moist soils because the thermal resistivity of
water (1.65 C-m/Watt) is much lower than that of air
(~45 C-m/Watt). In addition to the air having higher
thermal resistivity, heat transfer takes place by radiation
instead of conduction that is much less efficient.
Soil Texture and Dry Density
The soil texture is also critically important to thermal
resistivity. The grain size distribution and grain shape
are evaluated by a sieve analysis (e.g., ASTM D422,
etc.) to determine the variation in particles both in
backfill materials and native soils. Figure 3.6-5 shows a
sieve analysis for four materials and a band of good
granular thermal backfill.
Water Content and Ground Water Level
As is seen in Figure 3.6-3, soils with higher moisture
content generally speaking have better thermal resistiv-
ity. Some soils naturally retain water better than others.
Certain soils may not retain water welle.g., they have
a high hydraulic porositybut are below the water table
so they remain saturated even though the dry density is
low and dry thermal resistivity would otherwise be high.
Dry Density
The dry density of a soil determines its ideal ability to
conduct heat away from the cables. Factors that influ-
ence the dry density are porosity, solids content, inter-
particle contacts and pore size distribution. Having a
well-graded material with a range of particle sizes
improves the dry density and minimizes pores and voids
in the material.
Other Subsurface Characteristics
Concerns for solutes and hysterisis apply only in areas
where significant fluctuation in the water table may
wash out fines from the backfill, making it thermally
poor. For most applications, this is not a concern for
cable system uprating and, in any case, would be found
during soil thermal resistivity testing.
Surface Characteristics and Vegetation
Surface conditions have an impact on soil thermal resis-
tivity. For example, soils below asphalt roadways gener-
ally will not gain or loose moisture readily under normal
conditions. However, in the presence of cables, the dry-
ing that does occur may not be mitigated by heavy rains
since the water will not be easily reabsorbed.
Surface vegetation can be significant factor affecting
soil thermal resistivity. The root systems on large trees
and plants will draw moisture out of the soil, drying it.
Also, the decaying components of plants and their root
systems will tend to increase the organic component of
soils, which tends to increase the soil thermal resistivity
(see Table 3.6-1).
Surface cover has strong influence over earth ambient
temperatures, especially at shallower depths. A differ-
ence of 45C has been measured between grasses ver-
sus asphalt cover over cables.
Table 3.6-1 Thermal Resistivities of Soil Components
Component
Dry Thermal Resistivity
C-m/Watt
Quartz 0.12
Granite 0.30
Limestone 0.40
Sandstone 0.50
Shale (sound) 0.60
Shale (highly friable) 2.00
Mica 1.70
Ice 0.45
Water 1.65
Organics (peat, etc.) ~5.00
Petroleum Oil ~8.00
Air ~45.00
3-34
Engineered Thermal Backfills
The general goal of engineered thermal backfills (ETB)
is to enhance the removal of heat away from buried
cables. In most cases, the native soil materials have a
higher thermal resistivity than good quality backfills
and, in any case, are difficult to reconstitute in the
trench with the same density as the native soil. For this
reason, special backfill materials are often designed for
use in a cable trench. These include well-graded sands,
stone screenings, and concrete or Fluidized Thermal
Backfill. In addition to having excellent thermal proper-
ties, they are engineered to meet civil requirements
(strength and ease of voids-free installation) that are
associated with the particular application. The criteria
considered for these ETB are:
Low thermal resistivity over the expected range of
operating conditions
Low critical moisture content and high thermal sta-
bility limits
No adverse affects on materials used for cable con-
duits, cable jackets, or pipe coatings
Easy to install
Inexpensive and locally available to the location
where the materials will be used
Types of engineered thermal backfills are discussed in
the following sections.
Granular Backfill Materials
These materials should be composed of hard, well-
graded, natural or crushed mineral aggregate (limestone,
granite, quartz or other similar rock). The material
should be sound (porosity less than 2%) and be free of
any organic material (peat, root matter, topsoil, vegeta-
tion) and foreign matter (wood, rubble, cinders). The
sieve analysis should match closely to that given in 3.6-5.
The maximum particle size should be no larger than 1/4-
in. sieve size with a fines content (material finer than
#200 sieve size) of 12% to 18%.
During supply and installation of this material, quality
assurance is very important. Sieve analysis on the deliv-
ered materials should be performed periodically to check
and verify its compliance with the above characteristics.
Figure 3.6-5 Grain size distribution for soil samples.
3-35
Fluidized Thermal Backfill (FTB)
One of the difficulties with any granular backfill is that
it must be installed properly, regardless of its ideal ther-
mal properties. Granular backfills should be installed in
shallow lifts 15 cm (6 in.) at a time and well compacted
to give good density. This is labor intensive, and great
care must be used when working close to directly-buried
cables or conduits so as not to damage either.
Leading up to and during installation, FTB delivered to
the project site should conform to the respective mix
design and performance specifications of low-strength
and/or high-strength FTB. This should be checked with
samples collected during the project. When installed by
pouring into the cable trench, the material should be
free flowing and without any segregation. This will help
ensure that the material completely surrounds the
cables, conduits, or pipes. The amount of water in the
FTB mix may be adjusted to increase or decrease the
flow (slump) as directed by the field engineer. If lower
slump FTB is required for a particular area, it is gener-
ally better to adjust the water content at the batch plant
rather than as the material goes into the trench. Air con-
tent (natural trapped) should not be higher than 2%.
Mixing at the batch plant and transportation to the
project site should be done in accordance with ASTM
or American Concrete Institute (ACI) specifications.
If trench shoring and sheathing is being used, these
should be removed immediately after the installation of
FTB, unless otherwise required by the field engineer. If
FTB is installed in cold conditions, care should be taken
to protect the installed FTB from freezing. This applies
to both low-strength and especially high-strength FTB.
ASTM or ACI specifications should be followed for
such installations. Sampling and testing for quality con-
trol/assurance should be performed on FTB samples
taken every 250 ft along the cable trench, or every 100
cubic yards of material installed, or as directed by the
field engineer.
Component materials from an FTB mix design are
Grouts for Cable Conduits and Trenchless Casings
For extruded or self-contained cables in ducts or the
space between inner-ducts and trenchless (directional
drilling, pipe jacking, etc.) casings, the air space between
the cable and conduit or conduit and casing is often filled
with air, which is a poor thermal conductor. Filling the
duct with a thermally conductive material improves the
cable ampacity by 510%, depending on the configura-
tion and the type of filler. The annular space is generally
small, and utilities usually want to retain the ability to
remove the cables from the conduits later in the event of
a failure or for upgrading. Therefore, it is not practical to
fill the annular space with a solid filler material (or one
Figure 3.6-6 Component materials used in a typical Fluidized Thermal Backfill.
3-36
that becomes solid over time), so a pumpable material
that will not hard set is ideal.
IEC-60287 allows that cables with grouted conduits may
be treated as direct-buried cables for the purposes of
ampacity. Various materials that have been considered
for conduit grouts are:
Bentonite and sand/bentonite slurry
Sand-cement grout
Flyash-cement grout
Grease and viscous oil, along with other compounds
Water
Factors that must be considered when selecting a grout
are the total length that must be filled and the amount of
annular space. For trenchless installations, the potential
softening of a plastic duct at elevated temperatures
including potentially the heat generated as cement-based
grouts curecould soften ducts and cause partial col-
lapse. The safe pumping pressure for the grout material
must therefore be considered when a grout is pumped on
the outside of air-filled cable conduits.
The grout material typically will have a thermal resistiv-
ity of 0.4 to 1.4 C-m/Watt, which is much lower than
air (45 C-m/Watt) at the set moisture content. A sand-
bentonite slurry backfill with a thermal resistivity of
approximately 0.7 C-m/Watt is easy to formulate and
generally easy to install. Varying the amount of sand,
bentonite and water affects the pumpability of the
grout. Bentonite tends to absorb a lot of water, so this
must be factored into the mix. Mixing the sand/bento-
nite slurry also requires special equipment (i.e., colloidal
mixer). The thermal resistivities of these components
are as follows:
Sand: 0.12-0.20 C-m/Wattoptimizes the thermal
resistivity but negatively affects pumpability.
Water: 1.65 C-m/Wattoptimizes the flowability
but negatively affects shrinkage.
Bentonite: 3.50 C-m/Wattoptimizes the pumpabil-
ity but negatively affects thermal resistivity.
These materials are combined by a soil specialist for use
by the contractor or utility during installation.
3.6.2 Review Circuit Plan and Profile
A classical approach to performing uprating on under-
ground cable circuits is to review the circuit plan and
profile drawings, preferably the as-built versions,
which may show additional details about the locations
of the buried power cables, as well as better illustrate the
locations of other underground utilities that may impact
cable ratings.
The plan drawings will show a variety of factors that
may be relevant to determining the cable ampacity and
possible locations where uprating could be considered:
Phase and circuit or pipe spacing among the cables
being studied, which would impact mutual heating
effects.
The locations of other utilities that cross the cables,
especially other transmission or distribution cable
circuits that could produce mutual heating effects.
Also, steam lines may be present.
Sections of the route that parallel other utilities,
including power cables. Parallel cables within a cer-
tain range may produce sufficient mutual heating to
cause derating. A general guideline is, if the horizon-
tal spacing is within 25% of the depth, mutual heat-
ing may be a factor (e.g., if the cables being studied
are at 4 m depth, parallel cables or other heat sources
within 1 m horizontal spacing should be examined
for mutual heating effects).
Topographical profiles may show areas where over-
burden has accumulated above the cable route.
Profile drawings mainly indicate the cable circuits depth
of cover below grade and usually the locations of other
utilities that cross the cable circuit. Areas that are
important to note on the profile drawing are:
Entry/exit to manholes since cables frequently dip to
enter a manhole
Road crossings where the cable burial depth may be
increased to accommodate the required road bedding
materials
Directional drilling locations where the burial depth
is significantly greater than conventionally-trenched
sections
3.6.3 Evaluate Daily, Seasonal, or Other Periodic
Load Patterns
Load shape is generally not that important for most
transmission equipment, particularly overhead lines
where the thermal time constant is relatively short. With
underground transmission cables, the long thermal time
constant35150 hourscan significantly impact
loading patterns for both normal and emergency rat-
ings, particularly for short-duration emergencies.
In typical normal ampacity ratings on cables, daily load
cycles are modeled by rating techniques through the
application of a load factor or loss factor. The load fac-
3-37
tor relates the average daily load to the peak load, usu-
ally following a relationship similar to the following:
Loss Factor, p.u. = 0.3 (Loss Factor, p.u.) + 0.7 (Loss
Factor, p.u.)
2
This relationship is graphed in Figure 3.6-7. As men-
tioned earlier, the loss factor (or load factor of the
losses) considers the average daily heat output relative
to the peak heat output. Consider Figure 3.6-8, which
shows several load shapes that all have the same peak
current but substantially different load and loss factors.
All of the curves in Figure 3.6-8 have the same peak cur-
rent (1000 A), but substantially different loss factors.
On a daily basis, the different loads shown will release
different amounts of energy into the surrounding soil.
This has a significant impact on conductor sizing for a
desired rating or on the available current for a given
conductor size. Note that the loss factor is also the per-
unit power delivered on a daily basis.
If the cable construction and installation conditions are
held constant and the loss factor is varied, the cable rat-
ings will vary substantially.
From the standpoint of uprating, increases in loss factor
over time mean that the ampacity will tend to decrease.
For example, on a recent uprating study for a New
England utility, the loss factor in 1959 when the circuit
was built was 57% but had grown to 83% in 2001. While
the utility was able to increase capacity on the circuit
with some extraordinary methods, the normal book rat-
ing actually decreased with respect to time because of
the increasing loss factor.
Load shape may also play an important role from the
standpoint of emergency ratings. If the daily load cycle
is such that the load during portions of the day (typi-
cally at night) is lower than at other times of the day
(typically mid-afternoon), short duration emergencies
can vary greatly. This is illustrated in Figure 3.6-9,
where the normal ampacity (1.0 per unit), A, is deter-
mined for a peak temperature of 90C, and two 4-hour
emergency ratings are determined:
Emergency Rating B: The peak temperature is 105C
with a rating of 2.6 per-unit (as compared to the nor-
mal rating). This emergency starts going into a low-
load period, so the pre-emergency load temperature
is about 73C.
Emergency Rating C: The peak temperature is also
105C with a rating of only 1.3 per-unit (as compared
to the normal rating). This emergency starts going
into a peak load period, so the pre-emergency tem-
perature is about 85C.
The above example illustrates that considering the load
shape for emergency ratings is important. Dynamic rat-
ings (see Section 3.8) is a main benefit for this type of
analysis in optimizingand generally increasingthe
current carrying capacity of an underground cable circuit.
Figure 3.6-7 Ampacity as a function of loss factor.
Figure 3.6-8 Graph of load profiles showing the same
peak current with different load and loss factors.
Figure 3.6-9 Temperature plots and ratings as a
function of rating starting time.
3-38
3.6.4 Temperature Monitoring
Using Thermocouples
Temperature measurements are an important part of
verifying assumptions when calculating ampacity and
studying ways to improve ratings. Ideally, one would
want to measure the cable conductor temperaturethe
hottest location in the cable systemto be sure that
insulation temperature limits are not exceeded. How-
ever, because the conductor is energized, this is typically
difficult to do. Instead, it is common to measure the
temperature on the outer surface of the cable either on
the pipe coating of a pipe-type cable or the jacket of the
other cable types. For conduit installations, tempera-
tures might be measured in the conduits.
To perform these measurements, thermocouples are
often used. Thermocouples are temperature sensors
based on the principle that when two dissimilar metals
are joined, a predictable voltage will be generated that
relates to the difference in temperature between the
measuring junction and the reference junction (connec-
tion to the measuring device). The types of metals that
are used depend on the application (temperature range,
location, cost, etc.). There are varieties of thermocouple
types (T, F, N, J, etc.). For cable-related measurements,
Type-T thermocouples are most often used because
they have a temperature range most closely matched to
typical cable operating temperatures. The Type-T ther-
mocouple has a blue outer jacket in the United States,
France and the United Kingdom (up until 1993) or dark
brown outer jacket in the United Kingdom (since 1993)
and Germany. Inside, the thermocouple wire consists of
a copper electrode (positive, +) and a constantan elec-
trode (negative, -). Each electrode has an insulating
coating that varies in color depending on the country of
origin (United States is blue on the positive and red on
the negative; the United Kingdom is white on the posi-
tive and blue on the negative (pre-1993) or brown on the
positive and white on the negative; France is yellow on
the positive and blue on the negative; and Germany is
red on the positive and brown on the negative). When
connecting thermocouple wire to a meter, data logger,
or other measuring device, it is important to verify that
the polarity is correct. Otherwise, the schematic will
essentially create three thermocouple junctions in series
(rather than one), which could provide misleading
results. Also, the thermocouple wire or extension grade
thermocouple wire must also be run from the measure-
ment location all the way to the test instrument.
A thermocouple junction is created by joining the cop-
per and constantan wires together as shown in Figure
3.6-10. The junction can be left bare, which minimizes
thermal capacitance and increases temperature mea-
surement response. However, depending on the environ-
ment, the junction may need to be coated or soldered to
protect the thermocouple junction from corrosion, etc.
Laboratory-grade thermocouples are typically welded
together. A thermocouple has an accuracy of typically
less than 1 C.
A key advantage to thermocouple temperature measure-
ment is that the wire itself and the equipment to mea-
sure thermocouple temperatures are both relatively
inexpensive and minimal training is required to use the
technology. Several companies including Telog Instru-
ments, Omega, and Fluke make data loggers that cost
less than US $1000 to read and possibly record thermo-
couple temperatures. Battery-powered recorders can log
data for 618 months, recording temperatures every 15
minutes for an extended period. Once suspected or
known cable circuit hot spots are identified, low-cost
thermocouples and data loggers may be placed at these
locations and checked periodically, particularly during
periods of high load.
By comparing measured temperatures with those pre-
dicted using load history and the equivalent thermal cir-
cuit from ampacity calculations, it is possible to
evaluate the assumptions used in ampacity calculations.
From an operations standpoint, monitoring the cable
temperatures gives some assurance that the cables are
not exceeding their allowable temperature during typical
load cycles.
The main disadvantage to thermocouple measurements
is that they only take a temperature measurement at one
location. It is, therefore, possible to miss hot spots if
they are not already identified. Also, the practical lead
length limit of thermocouples is about 300 m (1000 ft),
and each thermocouple requires its own pair of wires to
Figure 3.6-10 Thermocouple wires (copper and constantan
with U.S. color scheme left) and completed junction
(right).
3-39
make a measurement. Therefore, it is difficult to instru-
ment more than a few tens of locations.
Distributed Fiber Optic Temperature Sensing (DFOTS)
Distributed fiber optic temperature sensing (DFOTS)
uses a specialized optical time-domain reflectometer
(OTDR) to measure the temperature along a multimode
optical fiber. The process works by taking advantage of
temperature-dependent reflections (called backscatter
based on the Raman Effect) in the fiber. The special
OTDR instrument, such as York Sensors (Sensa) DTS-
80 or SensorTrans Model 5000 (see Figure 3.6-11),
records the magnitude of the reflection (proportional to
temperature) and the time for reflections to return after
sending an incident 1080 nm laser pulse into the fiber,
which, when combined with the fibers propagation
velocity, gives the distance to the measurement location.
By successively sending light pulses into the fiber, the
special OTDR can scan the entire fiber and obtain a
temperature trace along the fiber with a spatial resolu-
tion of approximately 1 m and a temperature accuracy
of 1C. The obvious advantage of DFOTS is that a con-
tinuous end-to-end temperature measurement is possi-
ble, allowing the ampacity study to reveal all of the hot
spots along the cable route. Later, these hot spots could
be instrumented with thermocouples for extended tem-
perature measurements at key locations.
Depending on the configuration of the fiber (number of
splices, etc.) and the temperature measurement mode
(single or double ended), a fiber length of 510 km may
be scanned. Equipment is also available that works with
single-mode fiber and can measure up to 30 km. How-
ever, this equipment suffers from both lower spatial res-
olution (410 m) and lower accuracy (23C). All fiber
test loops are limited by the losses in the system, so
fusion splicing is the preferred method for joining fibers.
Fiber used for DTS measurements is typically installed
in a parallel conduit or directly buried alongside an
existing cable or pipe. Retrofitting a fiber on a direct-
buried system is impractical unless there is a conduit
(communications or power) within a meter or so of the
energized cables in which the fiber may be installed. An
example of fiber that might be installed directly buried
or in a conduit is shown in Figure 3.6-12.
The fiber optic cable typically consists of four to six 50 x
125 m fibers, each with a 900 m tight buffer (only 1 or
2 fibers are needed, but some may be damaged during
installation so spares are desirable), Kevlar strength
members to improve pulling strength (usually only
3000N, 675 lb), and a fire-retardant PVC jacket.
Some XLPE cable manufacturers are embedding optical
fibers under the jacket of the cable to facilitate tempera-
ture measurements (Figure 3.6-13). Since this is physi-
cally closer to the conductorultimately where we
would like to know the temperaturethis has some
advantages.
A disadvantage of DTS equipment is the cost of the
electronics to measurement the fiber temperature, which
may be upwards of US $60,000. Also, the equipment is
Figure 3.6-11 Distributed temperature sensing equipment.
Figure 3.6-12 Optical fiber cable used for DTS
measurements.
Figure 3.6-13 XLPE cable with integrated optical fiber
under the jacket.
3-40
not well suited for operation in the field on an extended
period of time, which makes spot measurements or
extended measurements in manholes or stations diffi-
cult. Most of the equipment has an operating range of
1035C, which is fairly limited in particularly warm or
cool climates.
3.6.5 Ampacity Audit
An ampacity audit involves investigating ampacity
for a cable circuit by applying the various techniques
described in this chapter. The basic concepts include
performing a soil thermal survey to determine soil char-
acteristics and ambient temperatures, evaluating load
history, and then calculating ampacity. If AEIC guide-
lines are being followed on a circuit that previously had
used assumed soil parameters, the 10C increase in con-
ductor temperature by itself generally allows a 20%
increase in ampacity.
The ampacity audit is geared towards verifying the
ampacity by whatever means are available and assessing
which locations along the route limit the overall circuit
ampacity. This might possibly include obtaining a
DFOTS temperature trace for the route to find hot
spots, or looking at route plan and profiles to find limit-
ing installation conditions. These hot spots would
then be investigated further to see how they might be
mitigated.
3.6.6 Remediation of Hot Spots
Remediation of hot spots is sometimes possible if the
scope of the hot spot is limited.
If the hot spot is the result of overburden, or increased
burial depth, it might be possible to remove some of the
overburden above the cables. This reduces the thermal
resistance to heat leaving the cable and may improve
ampacity.
Poor soil thermal resistivity can often lead to hot spots,
particularly if a low quality thermal backfill was used
or not backfill at all. A hot spot may be eliminated or
partially mitigated by excavating around the cables and
installing a good quality thermal backfill. This will
improve the heat transfer characteristics away from the
cable, lowering the operating temperature for a given
load condition.
Heat Pipes
In extreme cases, usually where one circuit experiences a
hot spot from mutual heating of another circuit, the
installation of heat pipes can help. The heat pipe is a
passive device that takes advantage of the heat of vapor-
ization to remove heat from a location. A heat pipe is
constructed using an alcohol-water or ammonia-water
mixture in a partially filled copper tube. A partial vac-
uum is drawn on the tube to adjust the vapor pressure to
the operating temperature range for the particular
application. The heat pipe is then installed at an angle
with the low point installed near the heat source (cable,
steam main, etc.). Heat from the source is absorbed by
the liquid alcohol-water or ammonia-water solution,
causing a phase change to vapor, which rises, carrying
the heat away. The gaseous vapor then condenses back
to a liquid away from the hot spot and then drains back
to the hot spot location. This continuous process
removes heat from the location. The number and instal-
lation geometry of the heat pipes are typically designed
by a specialist.
An example of a heat pipe installation to mitigate the
effects of crossing cable circuits is shown in Figure
3.6-14.
3.6.7 Active Uprating
The following methods are mostly applicable to pipe-
type cables, although there are applications that could
extend to extruded or self-contained cables.
Figure 3.6-14 Heat pipes being installed to mitigate a hot
spot where a steam main crosses a pipe-type cable.
3-41
Fluid Filling
Gas-filled pipe-type cables may be uprated slightly by
replacing the dry nitrogen gas with a dielectric liquid
such as polybutene or alkylbenzene. Since a liquid is a
more efficient heat transfer media than a gas, fluid-fill-
ing alone provides a small ampacity improvement (~2-
3%). However, this allows for additional uprating meth-
ods to be applied.
Fluid Circulation
Fluid circulation is a relevant uprating technique when a
short section, relative to the overall circuit length, is lim-
iting the pipe-cable rating. By circulating the dielectric
liquid within the cable pipe, the heat generated in the
hot section will be transferred to other portions of the
route, mitigating the hot temperatures at that location.
One requirement for this to be implemented is to have a
fluid return pipe or a parallel cable pipe that will permit
a continuous circulation path as shown in Figure 3.6-15.
Flow rates may be up to 800 liter/min (200 gpm), but
slow circulation with only 20 liter/min (5 gpm) may be
used for small hot spots or where the fluid viscosity lim-
its the flow rate.
If no fluid return pipe or parallel cable pipe is present,
fluid oscillation may be used. In this configuration, fluid
is moved through the pipe at 440 liter/min (110 gpm)
and cycled between the 4,00011,000 liter (10003000
gallon) fluid reservoirs at either end of the pipe circuit.
Major considerations for fluid circulation are the free
area in the pipe and the viscosity of the dielectric liquid
used in the pipe. As a result, the pressure rise when
pumping dielectric fluid through a pipe could be too
excessive for practical uses. If the flow rate is limited to a
value below what is necessary to mitigate a hot spot, cir-
culation may not be possible to mitigate a hot spot; the
utility might consider changing to a lower viscosity
dielectric liquid or re-examining pressure rise limitations.
The basic principle of fluid circulation is based on work
done by CIGR and discussed in Electra (CIGR
1979). The approach is to evaluate sections of the fluid
circulation route that have basically the same character-
istics and then use boundary conditions to match the
flow rate from one section to the next. Dielectric liquid
(or water in parallel circulation/cooling tubes) is rela-
tively noncompressible, although the density varies with
temperature around the circulation loops. To consider
this, the dielectric fluid characteristicsdensity and spe-
cific heatare adjusted for each section that is being
modeled. As fluid circulates through the pipes, the tem-
perature of the fluid leaving one section is assumed to be
the temperature entering the next section, satisfying the
boundary conditions.
To model each section, the un-cooled (temperature
that would result absent of any cooling or circulation
movement in the pipes) temperature is calculated based
on circuit loading, the cable construction, and installa-
tion conditions at each section. Then, heat absorbed or
removed would cause increases or decreases in the
dielectric fluid temperature as it moves through the
pipes. The temperature change with respect to distance
is of the form shown in Equation 3.6-2.
3.6-2
In the equation, x is the distance, K is a value propor-
tional to the maximum change in temperature possible
for a section of infinitively long length, A is a value
relating the mutual heating affects among the pipes
(either cabled or fluid return), and P is a value charac-
terizing the rate of temperature change as air moves
along the pipe section. There is one exponential term for
each pipe in which fluid is circulating.
Figure 3.6-15 Example pipe cable dielectric fluid circulation loop with heat exchangers.
( ) exp( )
UN COOLED
T x T K A P x
=
3-42
The un-cooled temperature is the steady-state temper-
ature of the dielectric liquid inside the cable pipe that
would result if the loading remained fixed on the ener-
gized cables and no fluid circulation was in place. As the
flow rate is reduced, the value of K approaches zero. As
the flow rate increases, the value of K approaches the
difference between the inlet temperature of the dielectric
liquid and the uncooled temperature for a given section.
Changes in dielectric liquid temperature as it passes
through the cable pipes provide an indication of the heat
being removed in each section based on the mass flow
through the section as defined by Equation 3.6-3.
3.6-3
Where:
is the mass flow rate in kg/sec.
is the density in kg/m
3
.
A is the free area within the pipe in m
2
.
is the velocity in m/sec.
The heat removed, Watts, in a given section can then be
found from Equation 3.6-4.
3.6-4
Where:
C
P
is the specific heat in kJ/kg-C.
T
OUT
and T
IN
represent the outlet and inlet temper-
atures, respectively, of the dielectric liquid in a
given section.
By knowing the net heat absorbed or lost in a given sec-
tion and the length of that section, it is possible to eval-
uate the net heat removed (or gained) in a given section.
For a forced-cooled system (described next), the net
heat gained by the system will assist with sizing the
forced-cooling plant and heat exchangers.
Fluid circulation could be applied to extruded dielectric
or self-contained liquid-filled cables by installing parallel
water cooling pipes next to the cables and then circulat-
ing water through those pipes. Although technically fea-
sible, this is not often done. In addition, some utilities
such as National Grid in the United Kingdom circulate
the dielectric liquid in the fluid channel of self-contained
liquid-filled cables. Again, this is relatively rare.
Forced Cooling (Water or Oil)
Forced cooling is an extension of fluid circulation. The
main difference is that, rather than just moving heat
around from hot spots to cold spots in the cable
route, the dielectric liquid (or water in the case of water
cooling of extruded or self-contained cables) is diverted
from the cable pipe, passed through a heat exchanger to
remove heat, and then reintroduced to the cable pipe.
This has the potential of increasing the ampacity by 50
70%, although the cost and maintenance of these active
systems can be high.
Considerations for Active Uprating
With fluid circulation or forced cooling in pipe-type
cables, there are some cautions associated with using
these uprating methods. Pressures along the hydraulic
loop may become excessive as a result of hydrostatic
head pressure, fluid flow restrictions near joints, and
cross-over plumbing between feeders and fluid return
pipes. The high pressures could cause the termination
housing to fracture, potentially resulting in a cable fail-
ure, dielectric fluid leak, and fire.
Pressure drop along long circulation loops must be con-
sidered. The degree of snaking of the cable phases
within the pipe can affect the fluid flow and pressure
drop, potentially limiting the flow rate. The pressure
drop as a function of length can found by evaluating the
Darcy-Weisbach equation, as shown in Equation 3.6-5.
3.6-5
Where:
f is the friction factor, empirically determined
based upon the Reynolds Number cable-to-
pipe inner diameter ratio.
is the density, kg/m
3
.
V is the flow velocity, m/sec.
D
h
is the hydrostatic diameter, m.
The pumping plant, in particular the fluid circulation
pump, must be able to accommodate the circulation
pressure. The density and viscosity of the dielectric liq-
uid will impact the allowable pressure drop, in addition
to the free area within the pipe and the degree of cable
snaking. On long circulation loops, multiple loops may
be needed with intermediate fluid circulation stations to
limit the pressure drop. Complex control systems, par-
ticularly in the event of a cable failure, must also be
developed to manage the various cooling loops and to
stop fluid circulation in the event of a fault.
Fluid circulation is often considered for the buried pipe
sections. However, when forced cooling is used, the riser
sectionslengths of pipe between the trifurcator and
terminationmay become limiting and could require
specialized plumbing to allow circulation in these areas.
Diffusion chambers may also be necessary to avoid
damaging the outer layers of insulation. A factor to
consider for uprating older pipe circuits, where the riser
=
A m
m
( )
P OUT IN
q m C T T
=
2
2
h
dP V
f
dL D

=
3-43
section may be limiting, is that installation of a diffusion
chamber on a riser that is not so equipped can be diffi-
cult. This is because working close in on the riser pipe
would be difficult with the cable already in place, and
removing and installing a new termination to put in the
diffusion chamber may be impractical.
3.6.8 Shield/Sheath Bonding Scheme
As discussed to some extent in Section 3.4, there are
three methods for grounding the shield/sheath on single-
conductor (extruded dielectric, self-contained fluid-
filled) cable systems: multipoint bonding, single-point
bonding and cross-bonding. Multipoint bonding
involves tying the shield/sheath connections together
and to local ground at both terminals and usually at
intermediate manholes, resulting in a path for induced
circulating currents but with minimal induced voltages.
Single-point bonding involves grounding the
shield/sheaths at only one location along a given sec-
tion, preventing circulating currents but leaving the
other end un-grounded where a standing voltage will
appear. Cross-bonding involves dividing the cable sec-
tions into groups of three minor sections that are close
to the same length and transposing sheath connections
at the one-third and two-third locations, thereby elimi-
nating net circulating currents and minimizing induced
voltages.
Generally, single-conductor transmission cables are
designed with cross-bonding or single-point bonding to
minimize shield/sheath circulating currents in the pres-
ence of relatively high phase currents. The one exception
to this common practice is submarine cable installa-
tions, where multipoint bonding of the sheath is almost
mandatory because of the long installation lengths and
typically wide phase spacing. Contrary to transmission
practice, most distribution circuits are multipoint
bonded where the utility transformer and customer ser-
vice panel are both grounded for safety and so the neu-
tral can carry imbalance currents. The circulating
currents in multipoint bonded systems generate addi-
tional I
2
R losses (heat) in the shield/sheath that impacts
ampacity.
Multipoint bonding systems generally have about 20-
30% lower ampacity than single-point bonded or cross-
bonded systems constructed with similar cable sizes. If
the ampacity audit reveals that a circuit has lower
ampacity than desired and happens to be multipoint
bonded, the shield/sheath connections might be recon-
figured for sectionalized single-point bonding or cross-
bonding to eliminate the circulating current and gain a
significant improvement in ampacity. The reconfigura-
tion may require changing out some or all joints since
the joints must have shield interrupts to provide for sin-
gle-point bonding or to facilitate transposing the sheath
connections for cross-bonding. If a system that was pre-
viously multipoint bonded is being reconfigured for sin-
gle-point bonding, a ground continuity conductor
should be installed to provide a low impedance path for
fault current; the shield breaks will otherwise block the
flow of fault current. Also, single-phase or three-phase
link boxes or cross-bonding boxes will be needed.
If uprating using a reconfigured sheath bonding scheme
is being considered, and utility practice is typically with
multipoint bonding systems, care should be used to
clearly mark all manholes where standing voltages may
appear. A shield that is not grounded locally, but is con-
nected to ground at the adjacent manhole, may experi-
ence a significant voltage rise with respect to local
ground during nearby system fault conditions from a
combination of induced voltages and system potential
shifts. This is not a phenomenon peculiar to single-point
grounded arrangements, as any shielded cable is suscep-
tible when the shield is connected to a remote ground
yet remains ungrounded locally. The remedy for this sit-
uation is to provide secure, temporary shield grounding
as appropriate. This has always been a recommended
practice. This is particularly important when working
on a de-energized single-point bonded circuit that paral-
lels an energized circuit, since the parallel circuit can
induce a voltage.
Also, single-point bonded or cross-bonded cable sys-
tems require periodic maintenance to check the jacket
integrity and ensure that there are no unexpected cur-
rent circulation paths. Fault location efforts may also be
complicated by single-point bonded or cross-bonded
sheaths, possibly requiring that the bonding connections
be reconfigured during cable fault location. Jacket fault
location in duct bank installations is difficult unless the
ducts are under the water table. Although not techni-
cally necessary, many cross-bonded cable systems are
installed with a parallel ground continuity conductor.
Induced currents in this parallel conductor can generate
enough I ^ 2 x R losses to produce mutual heating that
can affect ampacity of the phase conductors.
3.7 RECONDUCTORING (UPGRADING)
3.7.1 Introduction
In contrast to uprating, which is generally defined as
improving the capacity of existing equipment, upgrad-
ing considers using available infrastructure to economi-
cally put in new cables, replacing existing conductors.
This section discusses some of these issues.
3-44
A basic assumption for this section is that the cable sys-
tem is either pipe-type or a duct installation. Reconduc-
toring a direct buried cable is impractical and really
should be considered a new installation, along with the
installation of a parallel circuit when there is no existing
infrastructure (pipe or conduits). Neither of these cases
is covered by this chapter.
One possible scenario might be reconductoring one of
two parallel cable circuits. The condition of the older
circuit should be evaluated to see if loss of life or higher
operating temperatures might affect reliable perfor-
mance. The upgrading topics in this guidebook would
apply to the scenario, although the details of evaluating
the impact on the older circuit or other connected
equipment are beyond the scope of this chapter.
3.7.2 Larger Conductor Sizes
The main issue with considering a larger conductor size
is whether the cable will fit within the same conduit or
pipe. Paper-insulated insulation thicknesses are fairly
standardized based upon voltage class, but the designer
might consider using a larger conductor size combined
with a switch from conventional Kraft paper insulation
to laminated paper polypropylene paper (PPP) insula-
tion, which has a higher dielectric strength and, there-
fore, a lower insulation wall thickness. With pipe-type
cables, the issue is to maintain sufficient clearance in the
cable pipe. The traditional guideline for new pipe instal-
lations is to have at least 12.5 mm (0.5 in.) of clearance
in the pipe, as determined by Equation 3.7-1. However,
with older pipes, where upgrading would more often be
consideredeither pipes that have cables to be replaced
or old but unused empty pipes that might have new
cables installeda minimum clearance of 25 mm (1 in.)
is often considered prudent to allow for the possibility
that overburden or settling may have increased the oval-
ity of the pipe that possibly could affect a successful
installation.
3.7-1
Where:
D is the inner diameter of the cable pipe.
d is the diameter of the cable over the insula-
tion, plus 1.5 times the height of a skidwire, all
values in inches.
Figure 3.7-1 shows the clearance in a pipe-type cable.
For extruded or self-contained cables where normally a
single phase would be installed in each conduit, the
clearance is simply the difference between the inner
diameter of the conduit and the outer diameter of the
cable. Again, it is typical to have 12.525 mm (0.5
1.0 in.) of clearance in the conduits. If cables are not
already in the conduits, a mandrel through the conduits
should be used to check the maximum size cable that
can pass through the pipe or conduit.
For extruded cables in particular, the construction of
the metallic moisture barrier and metallic shield could
be adjusted to allow for a larger cable. As an example,
Commonwealth Edison (now ComEd, An Exelon Com-
pany) removed 138-kV paper-insulated cables from
ducts to be replaced by cross-linked polyethylene cables.
If the typical 3.2 mm (0.125 in.) lead sheath moisture
barrier and full-wall 138-kV insulation 21.6 mm (0.85
in.) were used, it would have meant the standard XLPE
cable design could not fit in the existing conduits.
Instead, the installation included a reduced insulation
wall (16.5 mm, 0.65 in.), a copper laminate tape with
copper shield wires, and a reduced jacket wall (as com-
pared to typical industry practice for that size cable),
greatly reducing the outer cable diameter and allowing
the cable system to fit in the conduits.
The relative effect of reconductoring on ampacity is dis-
cussed in Section 3.4.
3.7.3 Cupric Oxide Strand Coating
Some cable manufacturers have investigated using
cupric oxide coated strands to make conductors. By
coating the strands with cupric oxide, each strand has a
slightly increased resistance (in the radial direction) to
each adjacent strand. This improves the skin and prox-
imity effect factors of the conductorreportedly down
to 0.3 for both valuesover conventional copper
stranded insulations. This reduction in ac incremental
losses significantly improves the ac to dc resistance ratio
and ampacity. To illustrate, the pipe-type cable example
( )
2
1
1.366 1
2 2
D d
C d D d
D d

= +

Figure 3.7-1 Clearance in a pipe-type cable.
3-45
in Appendix 3.1 would show an increase in ampacity by
almost 3% if a cupric oxide strand coated conductor
were used.
3.7.4 Voltage Upgrading
Voltage uprating, possibly combined with an increase in
conductor size, would significantly improve the power
transfer limits. As an example, an 8-in. cable pipe that is
typically used for 138-kV cables with 11.2 mm (0.44 in.)
of Kraft paper insulation might be reconductored with
11.4 mm (0.45 in.) of PPP insulation, allowing an
increase to 230-kV cable with virtually no change in the
outer cable diameter and a 67% increase in power trans-
fer. Additional improvements in capacity might be pos-
sible if a larger conductor size can also be used. Section
3.2 lists typical insulation thicknesses for the various
cable constructions and insulation materials.
The main issue with voltage upgrading is that, in addi-
tion to possibly significant costs in new cables, several
pieces of substation equipment must also be replaced to
accommodate the new voltage level. This is sometimes
not so significant if the higher voltage level being con-
sidered for voltage upgrading already exists within both
terminals. Otherwise, the cost and effort generally make
voltage upgrading infeasible.
3.7.5 Superconducting Cables
At the time this chapter is being prepared, supercon-
ducting cables are largely in the research stage. Limited
sections of cable have been installed in controlled set-
tings (e.g., parallel to a 100% redundant overhead line,
in a laboratory setting, or in a nonessential capacity
underground). Several high-temperature superconduct-
ing (HTS) cable projects have been demonstrated largely
with U.S. Department of Energy funding, using either
warm or cold dielectric technology (Figure 3.7-2).
As the names imply, the warm dielectric cable uses
insulation that is at or above room temperature to sup-
port the energized line-to-ground voltage, while cold
dielectric cable utilizes cryogenic (liquid nitrogen
~80K) insulating medium. Examples of recent research
in superconducting cables are summarized below:
1996-1999
Pirelli/EPRI: 50 m of 115-kV cable with 2000 A.
Sumitomo/TEPCO: 30 m of 66-kV cable with 1000 A.
2000-2002
Pirelli/Detroit Edison: 130 m of 24-kV with 2500 A
(system could not be energized because of a vacuum
leak in the cryostat).
Sumitomo: 100 m of 66-kV cable with 1000 A.
Southwire: 30 m of 12.5-kV cable with 2600 A (oper-
ated in parallel to an overhead line on Southwire's
property).
NKT: 30 m of 30-kV cable with 3000 A (utility sub-
station).
Condumex: 5 m of 2000 A (Condumex test facility).
Nexans / American Superconductor / LIPA: 610 m
circuit 138-kV cable with 2510 A is being developed
for a new installation on Long Island.
Figure 3.7-2 Examples of superconducting cables
(courtesy of American Superconductor Inc.).
3-46
While favorable results have been observed, the available
technology is not yet practical for typical transmission
voltages, and the operating lengths have not been
increased sufficiently to allow transmission at what
have typically been considered medium-voltage levels
(e.g., transferring bulk power at high-current/low-volt-
age). At this point, superconducting cables do not yet
offer a commercially viable means for uprating existing
transmission circuits, though this is likely to change
within the next decade.
3.8 DYNAMIC RATINGS OF UNDERGROUND
CABLE SYSTEMS
3.8.1 Background
As discussed in Section 3.3.1, the thermal time con-
stants of underground transmission lines are signifi-
cantly longer than those of overhead lines and power
transformers. This is a result of the thermal inertia (or
heat capacitance) of the earth that surrounds the cables.
As a result of this thermal inertia, the dynamic rating of
underground transmission lines significantly exceeds
their steady-state ratings, provided that there are signifi-
cant variations in the line loading. Conversely, the
dynamic rating of underground transmission lines is not
much higher than their steady-state ratings if the lines
are consistently loaded near their steady-state ratings.
The term dynamic rating means the present rating of
a line, taking into account its load history and real-time
measurement of parameters (mainly ambient earth tem-
perature).
In actuality, the normal ampacity of a circuit does not
change, except for changing ambient temperature, since
it is generally based on an assumed daily loss factor and
installation conditions. Therefore, a dynamic normal
rating remains relatively constant unless some decision
about future loading is evaluated. Predicting future
loading patterns is difficult since unplanned system
changes may affect loading patterns, but the use of pre-
dicted load patterns, usually based on utility SCADA
systems, does allow for some future load estimations
(usually limited to 24 hours).
Tracking the cable conductor temperature is the main
goal of dynamic ratings, since this is ultimately what
limits the power transfer on the circuit. For paper-insu-
lated cables, tracking the conductor temperature also
permits an assessment of insulation aging and ulti-
mately the life of the cable system. EPRI funded a
detailed investigation of paper-insulated pipe-type
cables in the 1990s. Transformer ratings have been based
on a variety of factors including the insulation aging
and loss-of-life criteria. However, application of insula-
tion aging and acceptable loss of life has not often
been considered for underground cable ampacity.
The main benefit to dynamic ratings is the ability to
track the cables temperatures with time and changing
load conditions and then base emergency ratings on the
actual, rather than assumed, pre-emergency tempera-
tures. The benefits of this type of evaluation were illus-
trated in Section 3.6.3 and Figure 3.6-9.
3.8.2 EPRI Dynamic Ratings on Cables
Development of the EPRI Dynamic Thermal Circuit
Rating (DTCR) system was started in 1991 with
advanced models for overhead transmission lines, power
transformers, and underground cables. The cable model
used in DTCR was initially based largely on the under-
ground cable ampacity program Alternative Cable
Evaluation (ACE) in the EPRI Underground Trans-
mission Workstation (UTW) that was started in 1990.
ACE was an off-line ampacity program that would
determine normal and emergency ratings based on user-
specified input data. The ratings generated by ACE were
similar to what many utility engineers refer to as book
ratings in that they were static ratings based on
assumptions (usually worst-case) and then tabulated for
reference by engineers and operators.
As the utility industry changed dramatically during the
1990s, toward achieving higher profits, there was
increased pressure to get more capacity from existing
equipment while minimizing new construction whenever
possible. To this end, DTCR was developed to take real-
time data from off the shelf monitoring hardware,
and determine optimal ratings (not worst-case) for the
conditions at the time the ratings were performed.
Although this required some philosophical changes at
utilities to consider circuit ratings as moving, changing
entities rather than fixed parameters, the overall benefit
was to demonstrate a greater capacity in transmission
assets, including underground transmission cables.
The general theory behind DTCR is that, probabilisti-
cally, there are relatively rare circumstances where the
worst-case rating conditions occur at the same time that
the greatest possible circuit loading is required or
desired. As a result, by evaluating the ratings on a real-
time basis using actual, rather than worst-case condi-
tions, the real-time rating is much higher and the allow-
able power transfer is greater. This is illustrated
graphically in Figure 3.8-1.
In Figure 3.8-1, there is a relatively small region (rating
regime) where the dynamic rating distribution overlaps
the loading, indicating that at most times, the dynamic
3-47
rating is greater than the book ampacity. DTCR
allows the circuit to operate closer to the limit by per-
forming a real-time rating evaluation using measured
(rather than conservative or worst-case estimates)
parameters.
Cable Models
One approach to dynamic ratings on cablesthat used
in the Underground Cable Module in DTCRis based
on the paper by Neher and McGrath (1957), and on the
International Electrotechnical Commission (IEC) Pub-
lications 287 and 853 (International Electrotechnical
Commission 1982, 1989). These same calculation meth-
ods are used in off-line rating tools such as EPRIs
UTWorkstation ACE program.
The numerical technique used in DTCR to track con-
ductor temperatures is an additive wave method,
whereby the temperature response to a constant heat
input is tracked into the future. For cables, DTCR looks
at small intervals (< 0.5 hours), over which the load
can be treated as constant. The temperature response to
the load with respect to time is a function of the heat
input to the cable (either positive for increasing load or
negative for decreasing load). Each time the load
changes, a new temperature response wave i s
launched. As the load changes from interval to interval,
the total temperature response at a given time is the
summation of all the previous temperature response
functions from each change in heat from the cable (as a
function of load and the change in resistance with tem-
perature). This basic concept is described in a paper by
Neher (Neher 1963) and illustrated in Figure 3.8-3 with
an arbitrary load.
The numerical calculations are based on equations from
IEC-287 and IEC-853 and make use of attainment fac-
tors for changes in heat output from the cables. The
cable temperature changes with respect to time and
changing load based on the thermal response of the
cable and environment. The cable or pipe temperature
response is defined by Equation 3.8-1.
3.8-1
Where:
W
i
is the heat output from the cable.
R
A
, R
B
is the thermal resistance at steady state.
a, b are time constants representing how the
temperature changes with respect to time
for a given heat input.
For the temperature response of the environment, there
are two models in DTCR: one that assumes all cir-
cuits/cables carry identical currents, and a second where
two circuits may carry unequal loading.
Equal Loading
For the case of two three-phase circuits of extruded or
self-contained cable, all six (6) cables are assumed to
carry the same load even if the circuits are electrically
disconnected. The temperature rise above ambient from
ac loading is described by Equation 3.8-2, utilizing
exponential integrals, for the cable that is identified to
be the hottest cable in the trench.
3.8-2
Figure 3.8-1 Probabilistic view of dynamic ratings and actual circuit loads.
( ) ( ) ( ) 1 exp( ) 1 exp( )
Cable i A B
t W R at R bt

= +

2 2
2 2 6
2
16
( )
4
'
4 4
earth
s
HottestCable i
j j
j
D L
Ei Ei
t t
t W
r r
Ei Ei
t t

=

+

=

+ +


3-48
Where:
W
i
is the heat generated by each cable.
Ei is an exponential integral.
D
earth
is the earth diameter for the cable or conduit.
L is the burial depth for the hottest cable
among the group.
r
j
is the actual distance between the hottest
cable and adjacent cables.
r
j
is the distance between the hottest cable and
the image of adjacent cables.
The watts generated by each cable, Wi, are identical for
all cables in the group.
Unequal Loading
For the case of two circuits with unequal loading, the
heat output from the second circuit is defined separately
from the first circuit. The mutual heating effects with
respect to time are then evaluated as shown in Equation
3.8-3.
3.8-3
Where:
W
i1
, W
i2
are, respectively, the heat generated by
each cable in circuit 1 and circuit 2.
Cable Dynamic Rating Model
The combined effects of the two temperature response
functions gives the complete temperature response of
the cable conductor to a given cable heat output. When
there is changing load, typical of most transmission cir-
cuits, the accumulated temperature responses for each
change in load will give the conductor temperature with
respect to time. The general procedure and flow of the
DTCR cable model are illustrated in Figure 3.8-2.
Graphically, the changing load is illustrated in Figure
3.8-3, where the top graph shows an arbitrary load pat-
tern, with loads that are both increasing and decreasing
over time. The bottom graph shows the individual tem-
perature response waves to each change in load (fine
l i nes) and the summati on of al l the temperature
response waves (dark line). Also shown in the graph is
the addition of the dielectric temperature rise, which
remains constant with respect to time as long as the line-
to-ground voltage remains constant (this is an assump-
tion). As compared to other modules (overhead lines
and power transformers) within DTCR, the cable model
requires extensive computations since the long earth
thermal time constant requires looking back hun-
dreds of hours in the loading history. As a result, ratings
are normally not performed more often than once every
15 minutes or so. Fortunately, the long thermal time
2 2
Pr 1 1
2 2 3
2
2 2 6
2
4
16
( )
4
'
4 4
'
4 4 4
earth
s
imaryCable i
j j
j
j j
s
i
j
D L
Ei Ei
t t
t W
r r
Ei Ei
t t
r r
W Ei Ei
t t

=
=

+

=

+ +

+ +

Figure 3.8-2 General flow of information in DTCR cable model. (STE means Short Time Emergency
rating, and LTE means Long Time Emergency rating.)
3-49
constant also means that there are limited temperature
changes over that time.
Because of the considerable heat storage capacity of
underground cables, the operating power levels prior to
a contingency loading have a large effect on the actual
thermal behavior during emergency loading events.
EPRIs DTCR technology allows continuous monitor-
ing and establishment of such power equipments actual
pre-contingency thermal state. As a result, load shed-
ding during emergencies can often be avoided and capi-
tal investment in new equipment postponed.
3.8.3 Benefit of Dynamic Ratings
The thermal rating of underground cables is tradition-
ally calculated using worst-case seasonal loads and soil
temperatures. Underground cable thermal parameters
are based on manufacturers data, installation assump-
tions, and industry standards. Rating calculations are
typically performed off-line using worst-case assump-
tions to derive seasonal limits (maximum soil tempera-
ture, worst-case loss factors).
With dynamic ratings, actual soil temperatures and load
data are used in place of worst-case approximations,
allowing higher operating limits under most conditions
and more accurate thermal modeling under all condi-
tions.
Other than cabl e parameters and configuration,
dynamic thermal rating calculations for underground
cable require soil characteristics and temperature.
The equipment parameters can be verified by compar-
ing calculated to measured equipment temperatures,
which in the case of high-voltage underground cable is
the earth interface temperature, or a more general com-
parison between a measured temperature and calculated
temperature in a location near the power cables. The
measured temperature from a thermocouple or DFOTS
can be compared to the dynamic rating system-calcu-
lated temperature, as described in Section 3.4.7.
DTCR Circuit Ratings
The DTCR software allows the user to calculate real-
time equipment temperatures, multiple thermal ratings,
and remaining-time (or Time to Temperature Over-
load, TTO) during emergency loadings, given real-time
load and ambient ground temperature for underground
cable circuits. Circuit ratings, rather than equip-
ment ratings, are possible by modeling all of the power
equipment on a given circuit (e.g., series-connected
transformers, overhead lines, cables, etc.), or in the case
of cables, each unique installation section along a given
route, and then letting DTCR calculate ratings for all
equipment and locations on a circuit. Then, DTCR
selects the lowest rating for each category (normal, LTE,
STE) and reports these values as circuit ratings.
Utility Implementation of Dynamic Ratings
Some dynamic rating systems interface directly with
monitoring hardware. This has some disadvantages:
The hardware is probably redundant to monitoring
already done by the utility.
The hardware is sometimes very specialized, meaning
that the original vendor must be recalled for mainte-
nance and repairs, usually at a high cost.
Telemetry for any monitoring outside of a substation is
often complex, expensive, and prone to interruption.
The computer system used to perform rating calcula-
tions, by virtue of the direct connection to the moni-
Figure 3.8-3 Additive Wave model for
temperature tracking in underground cables.
3-50
toring hardware, must be located in the field where
weather and security are difficult to manage.
Any changes to the hardware usually require a
new custom-designed interface to collect the moni-
tored data.
There is an issue of getting calculation results from
the remote location back to the utilitys operations
center, where it can be used by engineers and opera-
tors to optimize circuit capability.
For these reasons, dynamic rating systems may be inter-
faced to the utility SCADA system so that the SCADA
system can handle all of the utility-specific telemetry
issues, and the dynamic rating system can concentrate
on just the rating and temperature calculations. Some
favorable characteristics are as follows:
The preferred mode of operation is one wherein
remote monitors provide data to the SCADA/EMS
database using utility-specific communication links,
and the dynamic rating system obtains its real-time
data from that database rather than directly from the
remote monitors.
The utility SCADA/EMS system must transfer real-
time monitor data from the database to a simple real-
time input data file on the dynamic rating system
computer through use of Network Access calls or
other operational programs.
The real-time input data on the dynamic rating sys-
tem is read from the ASCII file created by a SCADA
Network Access call.
A specific SCADA input file containing real-time
temperatures and ratings is continually updated by
the dynamic rating system and made available to the
utility SCADA/EMS system. This allows dissemina-
tion of the calculation results to anywhere within the
utility.
All calculations can now be performed in an engineering
office or operations center environment (rather than in a
substation) without the need for special monitoring
device communication links. Utility operations pro-
grammers need only develop simple SCADA Network
Access calls to write ASCII input data files to the
dynamic rating computer, rather than spending a lot of
time writing equipment-specific interface programs. Fig-
ure 3.8-4 shows the general layout of EPRIs dynamic
rating system (DTCR), which utilizes this approach.
Input
Data
File*
Input
Data
File*
Input
Data
File*
Input
Data
File*
Real-Time
Historical
Data File
(mmddyy.###)
Real-Time
Output File
(ratings,
temps., etc.)
Real-Time
SCADA Output
File (ratings,
temps, etc.)*
File
Handler
File
DTCR
Calculation
Algorithms
User Interface
DTCR Software Product
Telephone
Communications
Program
Direct Connect
Communications
Program
Radio Link
Communications
Program
Modem
Cell
or
Land Line
Wire or
Fiberoptic
(RS232)
PC in Engineering
Office or Operations
Center
wireless
SCADA/EMS
Database
Network Access
Call
Network Access
Call
Monitor
Monitor
Modem
Monitor
Output to Engineering
or Operations Center
DTCR Functional Diagram
* File with most recent real-
time data appended.

D
i
r
e
c
t

M
o
n
i
t
o
r

I
n
p
u
t
Figure 3.8-4 Schematic overview of DTCRs location within the utility architecture.
3-51
3.8.4 Required Monitoring
Underground cable dynamic rating systems require vari-
ous monitored parameters. Some of the parameters that
could be evaluated are summarized as follows:
Ambient Soil Temperature
The ambient temperature is a fundamental parameter to
be monitored since it directly affects the allowable tem-
perature rise from ac loading. Soil temperature monitor-
ing should be located where it is representative of the
ambient conditions imposed on the cable system. Usu-
ally, this means being at least 10 m (30 ft) away from an
energized cable circuit at typical installation depths. A
thermocouple treea series of thermocouples installed
at various depths at the same locationis ideally suited
for this type of monitoring, since it gives a range of tem-
peratures that may be used for the variation in burial
depth encountered along the circuit.
Load
The load is also an important parameter to know, even if
doing quasi-dynamic ratings (described in Section 3.8.4).
The load serves two purposes. First, it allows the conduc-
tor temperature to be calculated as a function of load.
Second, it allows a correlation between any measured
temperatures and the circuit load (see Section 3.4.7). This
is important for verifying ampacity capability.
Soil Thermal Resistivity
Real-time monitoring of soil thermal resistivity is not
common, mostly because the soil thermal resistivity
does not change rapidly. Some utilities permanently
install a thermal probe so that additional thermal resis-
tivity measurements can be made to account for sea-
sonal variations or weather effects.
As discussed in Section 3.6, thermal resistivity is a very
important parameter, since earth components of ther-
mal resistance represent more than half of the total ther-
mal resistance to heat leaving the cable.
Pipe-Type Cable Monitoring
Some parameters are uniquely important to pipe-type
cable dynamic ratings, particularly those that utilize
fluid circulation or forced cooling. Inlet and outlet pres-
sure to a cooling loop, fluid flow rates through the cool-
ing loop, inlet dielectric liquid temperature, and cooling
system outlet temperature all may be monitored to eval-
uate the performance of a pipe-type cable system with
active forced cooling. The capability of the cooling plant
to remove heat from the dielectric liquid is very impor-
tant, since this ultimately dictates how much forced
cooling can be applied to the pipe-type system.
3.8.5 Quasi-Dynamic (Real-Time) Ratings
Quasi-dynamic ratings utilize many of the principles
of real-time ratings except they may not be done on a
continuous basis. For example, a utility may have a
cable circuit that is only heavily loaded two months of
the year. On that basis, the cost to implement a dynamic
rating system may not be justified, but the ratings dur-
ing that two-month window are still critical.
Quasi-dynamic ratings might be applied by monitoring
load and temperatures for a period of time and then cal-
culating what the conductor temperature might be as a
result of that load. From this, the temperature of the
cable conductor at rated temperature can be extrapo-
lated for rating purposes. For example, the earth inter-
face temperature of a cable system may be monitored
with thermocouples for several months until a period of
high loading occurs. At that time, the utility may down-
load measured temperatures from a thermocouple data
logger and compare measured to calculated tempera-
tures to evaluate the assumptions used for rating mod-
els, both real-time and off-line ratings.
Quasi dynamic ratings may also apply to reviewing his-
torical load profiles using dynamic rating algorithms.
The load history on a circuit for a year or twoperhaps
covering high-load periods during summer months
could be run through a dynamic rating tool to evaluate
peak loading periods and study loss-of-life criteria (on
paper cables).
3.9 CASE STUDIES FOR UNDERGROUND
CABLE CIRCUITS
Section 3.9 describes uprating projects recently con-
ducted at utilities. It is hoped that these case studies help
to illustrate the general application of uprating tech-
niques described in this chapter.
3.9.1 CenterPoint Energy
Description of Circuit and Summary of Rating Constraints
and Utility Goals
A 138-kV HPFF underground transmission line was
constructed in 1969 from CenterPoint Energys Polk
substation (located at the intersection of Polk and La
Branch Streets in Houston, Texas) to CenterPoints
Garrott Substation (located at the intersection of Gar-
rott Street and Blodgett). The total length of this
138-kV underground transmission line is approximately
2.37 miles (12,500 ft). A 2500-kcmil, compact segmental
copper, 138-kV HPFF cable with 505 mils of insulating
tapes was used to construct the Polk Garrott transmis-
sion line.
3-52
In March 2001, a loop feed to a new CenterPoint sub-
station, Midtown Substation, was constructed by tap-
ping into the existing Polk to Garrott underground
transmission line. This loop feed to the Midtown Sub-
station (located on La Branch Street between Taum and
Drew Streets) segregated the Polk to Garrott under-
ground transmission line into two parts with the lengths
of 0.96 and 1.41 miles. CenterPoint wanted to evaluate
the power transfer capabilities of the Polk-Midtown-
Garrott and Polk-Downtown 138-kV underground
transmission lines in light of these changes and to opti-
mize the current-carrying capacity of the circuits. The
ampacity audit was based on an evaluation of recent
and historical data including:
Long-term load current, ambient soil temperatures,
and cable pipe temperatures during summer operat-
ing conditions
Distributed fiber optic temperature sensing (DFOTS)
measurements performed on the circuits in February
2002 for hot spot identification
As-built plan and profile drawings for the two cable
circuits
Cable manufacturing data
CenterPoint Energy also wanted to evaluate the condi-
tion of the thirty-two-year-old Polk-Garrott HPFF
cables in coordination with the ampacity analysis. Con-
sequently, two investigations were performed for this
purpose. First, Detroit Edison (DECo) performed dis-
solved gas analysis (DGA) and laboratory testing of
cable paper tape samples obtained during construction
of the loop feed to the new Midtown Substation. Power
Delivery Consultants (PDC) also performed cable dissi-
pation factor measurements at rated voltage using
EPRI-developed instrumentation. A previous EPRI
project (Transmission Cable Life Evaluation and Man-
agement) indicated that cable tape physical property
measurements, DGA, and dissipation factor measure-
ments are the best diagnostic tests to determine cable
loss-of-life.
The primary focus of the DTCR project was to examine
the ratings on the Polk-Midtown-Garrott circuit. This
circuit consists of two segments: 5,060 ft from Polk to
Midtown Substation, and 7,440 ft from Midtown Sub-
station to Garrott Substation. DTS measurements
showed that the hotspot for the Polk-Midtown-Garrott
underground line is a crossing with the Polk-Downtown
138-kV underground transmission line at the intersec-
tion of Polk and La Branch (just outside of the Polk
Substation). The cable used for the Polk to Downtown
138-kV line is identical to the Polk to Garrott line. The
depth of cover over the Polk-Midtown-Garrott line is
approximately 11 ft-2 in. at the hot spot location, and
the vertical clearance to the Polk-Downtown line (above
it) is approximately 3 ft. CenterPoint placed a thermo-
couple on the Polk to Garrott cable pipe near the inter-
section. Initially, the thermocouple temperature was
monitored with a data logger, but this thermocouple is
now connected directly to CenterPoint's SCADA sys-
tem. A thermocouple was also placed on the Polk-
Downtown circuit at the location of the crossing and
connected to SCADA.
In 2000, CenterPoint Energy (Houston Lighting &
Power) began this investigation (completed in December
2002 EPRI Report 1007539) to increase the circuit
capacity on the high-pressure fluid-filled (HPFF) pipe-
type cable connecting the Polk and Garrott Substations.
Various studies were performed to evaluate uprating
possibilities for this circuit, including the application of
distributed fiber optic temperature sensing (DFOTS).
Results of DFOTS revealed that a hot spot existed where
another pipe-type cable (CenterPoints Polk-Downtown
circuit) crossed over the Polk-Garrott circuit. Although
an overall increase in ampacity was found for the general
cable circuit, a net decrease in ampacity resulted from
the modeling of the mutual heating of the two cable cir-
cuits where they cross. This was anticipated prior to
beginning the project, so DTCR was implemented on
the Polk-Garrott circuit in an effort to optimize available
circuit capacity.
The principal goal of the project was to investigate an
optimized circuit rating in light of the interference tem-
perature effects detected by DFOTS and experienced by
the crossing pipes. A secondary objective was to demon-
strate that, under normal loading patterns, the maxi-
mum normal temperature (85C) of the conductor
would infrequently be exceeded.
Results of Uprating and Benefits to Utility
The following conclusions may be reached from review-
ing the application of DTCR at CenterPoint:
The DTCR modifications and subsequent data anal-
ysis showed that CenterPoint Energys Polk-Mid-
town-Garrott and Polk-Downtown pipe-cable
circuits can benefit from dynamic ratings. DTCRs
predicted load pattern assumes a typical 24-hour loss
factor cycle consistent with static ratings. However,
the emergency loading is a function of pre-load con-
ditions, and DTCR accurately calculates the conduc-
tor pre-load temperature based on historical loading
patterns. Also, DTCR uses the present loading to
predict time to temperature overload (TTO).
While a detailed ampacity study indicated there was
effectively a reduction in the established book rating
of 3.4%, applying DTCR allowed for a net increase in
the normal rating of 20.7%, based on considering a
3-53
dynamic rating using summer 2000 load data for a
quasi-real-time dynamic rating analysis.
In summary, the real-time dynamic ratings of the
Polk-Midtown-Garrott and Polk-Downtown under-
ground transmission lines are expected to be signifi-
cantly higher than the static ratings that were
calculated under a separate project, assuming there is
similar load pattern variability to that observed dur-
ing the summer 2000.
The following conclusions are a result of the cable con-
dition assessment testing performed on the Polk-Gar-
rott 138 kV underground transmission line.
DGA testing of pipe fluid samples and laboratory
testing of a cable sample indicated that Polk-Garrott
138-kV cable offers an exceedingly long life, which is
characteristic of HPFF cables.
Results of the dissipation factor measurements con-
firmed that the Polk-Garrott 138-kV HPFF cables do
not show any signs of insulation deterioration after
more than 30 years of operation.
3.9.2 United Illuminating Company
Description of Circuit and Summary of Rating Constraints
and Utility Goals
In 1989, United Illuminating Company (UI) performed
an engineering evaluation on the ampacity of the exist-
ing 1.4-mile-long UI 115-kV high-pressure gas-filled
cable Circuits 1710 and 1730, which connect UIs
Pequonnock Substation to the Seaview Tap in Bridge-
port, Connecticut, where the lines transition to over-
head conductors. A 1600-ft section under Bridgeport
Harbor, where the cables were buried approximately
25 ft under high-resistivity sediments, appeared to limit
the overall circuit rating. UIs construction records indi-
cated that the cable configuration under the harbor con-
sists of three pipes in a 5-ft trench, with a spare (empty)
pipe centered between the two cabled pipes.
Results of the 1989 thermal tests showed that soil resis-
tivity ranged widely, from 90 to 250 C-cm/Watt. This
range of values produced a degree of uncertainty in the
ratings. In addition, the degree of siltation and the
actual pipe positions since the cable pipes were installed
in 1961 was unknown. Because of the uncertainty of the
pipes locations, 1989 soil tests were done at least 50 ft
away from the expected pipe position to avoid possibly
damaging the pipes with the soil-coring equipment. The
uncertainty of some parameters from the 1989 study,
the limiting of the entire circuits ampacity by the har-
bor section, combined with UIs interest in increasing
the total power transfer on the circuit, precipitated UI
in undertaking a more thorough ampacity evaluation of
the harbor section. An additional goal was to consider
means for increasing ampacity on the circuit.
Power Delivery Consultants, Inc. (PDC) was contracted
in 2001 to perform a very detailed evaluation of the
Bridgeport Harbor portion of UIs 1710 and 1730 lines
using sophisticated modeling and state-of-the-art tech-
nology to gather information about the installation and
environment. The evaluation included several technolo-
gies:
Gyroscopic testing on the empty cable pipe to
develop accurate cable pipe plan and profile informa-
tion for the harbor crossing.
Hydroscopic surveying of the harbor bottom to eval-
uate the degree of siltation over the cable pipes since
they were installed in 1961 and, ultimately, to deter-
mine the cable depth of cover.
Distributed temperature sensing (DTS) using EPRIs
DTS equipment and fiber optic cable installed in the
spare cable pipe
Continuous thermocouple temperature monitoring
using installed thermocouples and data loggers
Updated soil sample testing to characterize the soils
at the depths of interest
Forced air ventilation to characterize possible uprat-
ing by removing heat from the cabled pipes
An EPRI report (1007534) documents the results of
these evaluations and a detailed ampacity study to
determine the actual capacity of UIs 1710 and 1730
lines under both normal and emergency ampacity con-
ditions, and describes possible approaches for increasing
the ampacity on the lines.
In 2002, UI implemented two of the recommendations
of the ampacity study:
Applied forced-air cooling on the parallel cable pipe.
Implemented DTCR to monitor conductor tempera-
ture and evaluate real-time temperatures on the cir-
cuit as the result of circuit loading.
Results of Uprating and Benefits to Utility
The initial ampacity study resulted in several recom-
mendations to mitigate the rating limits on UIs cable
circuits. These are summarized in Table 3.9-1. Two of
the options, forced air cooling on a parallel empty cable
pipe and dynamic ratings, were later implemented.
The forced-air cooling equipment is shown in Figure
3.9-1. The actual cost to install the forced-air cooling
equipment was substantially higher than the initial esti-
3-54
mate, largely due to some complications associated with
the civil works of the installation.
The following conclusions and recommendations were
reached based upon implementing DTCR on United
Illuminating Companys 1710/1730 circuits:
DTCR shows that the maximum normal operating
temperature of paper-insulated pipe-type cables (e.g.,
85C) is rarely exceeded, even though there are fre-
quent occasions when the actual circuit loading
exceeds the normal book ratings.
The combined effects of circuit loading on 1710 and
mutual heating from circuit loading on 1730 some-
times result in the conductor temperature on the 1730
circuit exceeding its maximum normal temperature.
As was indicated in an earlier study, this results from
the typically increased load levels on the 1710 and
1730 circuits, the increased daily loss factors, and the
fact that loads on the 1730 circuit are increasingly
approaching the loads on the 1710 lines.
UI may want to consider performing a long-term
loss-of-life evaluation on the 1710 and 1730 circuits
to see how the calculated operating temperatures on
the circuits have influenced accumulated loss-of-life.
Despite the occasional incursions above rated tem-
perature, the evaluation would likely show that dur-
ing a typical calendar year, the below 85C operating
temperatures for most of that time indicate that less
than a calendar year of life has been consumed. This
would provide increased confidence to UI that this
occasional high-temperature operation should not
adversely affect the future operation of the circuit. If
pursued, this additional work would apply paper-
insulated pipe cable aging characteristics developed
during EPRI work on accelerated aging at Waltz
Mill.
From the standpoint of evaluating normal ratings,
the book ratings were previously found to be 743 A
(on the 1710 circuit). Real-time ratings show a range
of improvements (1025%) depending on the condi-
tions (Table 3.9-2).
If the 1710 and 1730 circuits are ever decommis-
sioned, an evaluation of the paper insulation aging
should be performed to determine if the typically
higher operating temperatures on the 1710 cables
show increased aging over the 1730 cables, since both
circuits are made of similar vintage cable and have
had similar in-service lives.
Table 3.9-1 Summary of United Illuminating Pipe Cable Uprating Methods
Forced Air
Cooling
Recondi-
tioning
Fluid
Filling Circulation
Forced
Cooling
Reconduc-
toring XLPE
Water
Cooling
Dynamic
Rating
Location Harbor only Harbor only Circuit Circuit Circuit Circuit Circuit Harbor only Circuit
Maximum
estimated
increase
4.0%
a
a. Based on testing in September-October 2001; small additional increase possible.
5.6%
b
b. Land rating becomes limiting.
<3.0% 7.8% 2550% 18.0%
c
c. Not recommended for pipe-type retrofit with conventional technology.
5.6%
d
d. Not recommended; air-cooling could achieve same result.
e
e. Rating improvements by dynamic ratings vary depending on circumstances.
Additional
mainte-
nance
Minimal None Moderate Moderate High None None Moderate None
Environ-
mental Con-
cerns
Low None High High High None None Moderate None
Estimated
Cost
$100k $1.2M $1.5M $200k
f
f. Increase in cost minimal after fluid filling.
$1.5M
g
g. Assumes fluid filling already done.
$3M
c
$500k
d
$200k
Figure 3.9-1 Blower assembly for forced-air cooling on
pipe-type cable circuits.
3-55
Further work on DTCR may be warranted to pro-
vide for a cyclic rating factor on emergency ratings of
longer than 24 hours to fully take advantage of the
dynamic normal rating concept.
3.10 SUMMARY OF UPRATING AND
UPGRADING APPROACHES AND
ECONOMIC FACTORS
Tables 3.10-1 and 3.10-2 summarize the major improve-
ments in ampacity that may be realized by the various
uprating and upgrading techniques described in this
chapter. The specific costs for each depend largely on
the unique aspects of each installation or project, but
qualitative costs, along with many other factors, are
listed in the tables for reference.
The case studies in Section 3.9 give an indication of spe-
cific costs for one of the uprating projects.
Table 3.9-2 Comparison of Normal Book Rating to Real-
Time and Dynamic Ratings
Rating Type Minimum Maximum
Normal Book (Previous Study) 743A
Real-Time Normal 621A 817A
Dynamic Normal (100% Loss Factor) 0A (722A) 877A
Dynamic Normal (with Cyclic Rating
Factor)
0A (774A) 929A
Table 3.10-1 Summary of Uprating and Upgrading Techniques Applicable to All Cable Types
Evaluation Criteria
Measure Soil
Rho, Heat Pipes,
Mitigate Hot
Spots
Real Time Monitoring
Replace
Conductors
Voltage
Upgrading
Thermo-
couples
(Temp.)
Dist. Fiber
Optic (Temp.)
Dynamic
Ratings
Rating Incremental Increase
a
a. Incremental increase in ampacity over previous options.
20% N.A.
b
b. Monitoring the cable circuit alone, does not provide for rating improvements. Real-time monitoring, with the monitored
parameters fed into rating calculations provides for optimum ampacities.
10-20% 3-20% 66-80%
Reliability High High High High High High
Maintenance Low Low Low Low Low
Increased, but
Low
Losses None None None Low Low
Increased, but
Low
Lead
Time
Installation 2 mos. 3 mos.
c
c. If not done during the original installation of the cable system.
3 mos. 2 mos. 14 mos. 24 mos.
Operating None Real Time Real Time Real Time None None
Cost
Installation
d
d. Including material costs.
Low Low-Med Low-Med Low High High
Operating None Low Low Low Med.-High Med.-High
e
e. Generally, higher voltage equipment requires more extensive and expensive maintenance.
Table 3.10-2 Summary of Uprating Techniques Predominantly for Pipe-Type Cables
Evaluation Criteria Fluid Filling Circulation Forced Cooling
Slow Rapid
Passive Heat
Exchanger
Forced-Air Heat
Exchanger Refrigerated Cooling
Rating Incremental
Increase
2% 21% 8% 16% 31% 16%
Reliability High Medium Medium Medium Low Low
Maintenance Medium Medium Medium Medium High High
Losses None Low Medium Medium High High
Lead
Time
Installation 10 mos. 2 mos. 2 mos. 2 mos. 3 mos. 6 mos.
Operating N.A. Low
a
a. Typically less than 24 hours.
Low
a
Low
a
Low
a
Low
a
Cost
Installation High Low Low Medium High High
Operating Low Low Medium Medium Medium High
3-56
REFERENCES
The following references were used for this chapter and
may be useful to the reader for additional background
on the topics discussed.
Anders, G. J. 1997. Rating of Electric Power Cables,
IEEE Press/McGraw Hill.
Bascom, III, E. C., D. A. Douglass, G. C. Thomann,
and T. Aabo. 1996. Hybrid Transmission: Aggressive
Use of Underground Cable Sections with Overhead
Lines. CIGR 21/22-10.
Bascom, III, E. C. and J. A. Williams. 2002. Taking
Your Cables Temperature. Transactions of the T&D
World Expo. 7-9 May. Indianapolis, Indiana.
Bascom, III, E. C. and J. A. Williams. 2002. Ampacity
Evaluation and Distributed Fiber Optic Testing on Pipe-
Type Cables Under Bridgeport Harbor. Electric Power
Research Institute Publication 1007534. December.
Bascom, III, E. C. and J. H. Cooper. 2002. Condition
and Power Transfer Assessment of CenterPoint Energys
Polk-Garrott Pipe-Type Cable Circuit. Electric Power
Research Institute Publication 1007539. December.
Bascom, III, E.C. 2003. Underground Cable Uprating
and Upgrading Tutorial. Transactions of IEEE PES
Transmission & Distribution Conference. Paper 03TD0362
(Panel Session). Dallas, Texas. 7-12 September.
Bascom, III, E. C., J. A. Williams, M. A. Pasha, S. M.
Rahman, and W. Zenger. 2003. Ampacity Evaluation
of High-Pressure Gas-Filled (HPGF) Pipe-Type Cables
Under Bridgeport Harbor. Transactions of IEEE PES
Transmission & Distribution Conference. Paper
03TD0093. Dallas, Texas. 7-12 September.
CIGR 1979. The Calculation of Continuous Ratings
of Forced-Cooled Cable. Working Group 21.08. Study
Committee 21. Electra. No. 66. pp. 59-84. October.
CIGR. 1987. The Calculation of Continuous Rating
for Forced-Cooled High-Pressure Oil-Filled Pipe-Type
Cables. Working Group 21.08. Study Committee 21.
Electra. No. 113. pp. 97-120.
El-Kady, M. A. and D. J. Horrocks. 1995. Extended
Values of Geometric Factor of External Thermal Resis-
tance of Cables in Duct Banks. IEEE Transactions on
Power Apparatus and Systems. Vol. PAS-104.
EPRI. 1982. High Ampacity Terminations. EL-2233.
January.
EPRI. 1985. Volume 1: Calculating AC/DC Resistance
Ratios for High-Pressure Oil-Filled Cable Designs -
Designer's Guide. EL-3977.
EPRI. 1985. Volume 2: Calculating AC/DC Resistance
Ratios for High-Pressure Oil-Filled Cable Designs -
Details of Mathematical Derivations. EL-3977.
EPRI. 1992. Underground Transmission Systems Refer-
ence Book. TR-101670.
EPRI. 1997. Thermal Properties Manual for Under-
ground Power Transmission. TR108919. November.
EPRI. 1998. Transmission Cable Life Evaluation and
Management. TR-111712. September.
EPRI. 2002. Increased Power Flow Guidebook Over-
head Transmission Lines. 1001817. December.
Holman, J. P. 1997. Heat Transfer. 8
th
Edition.
McGraw-Hill. New York.
IEEE. 1988. IEEE Guide for Application of Sheath-
Bonding Methods for Single- Conductor Cables and the
Calculation of Induced Voltages and Currents in Cable
Sheaths. 575-1988.
Iizuka, K. 1974. Power Cable Technology Hand Book.
DenkiShoin, Tokyo. pp. 80.
International Electrotechnical Commission. 1982. IEC-
287. Calculation of the Continuous Current Rating of
Cables (100% Load Factor). International Electrotech-
nical Commission.
853-2. Calculation of the Cyclic and Emergency Cur-
rent Rating of Cables. International Electrotechnical
Commission. 1st Edition.
Neher, J. H. and M. H. McGrath. 1957. The Calcula-
tion of the Temperature Rise and Load Capability of
Cable Systems. Paper 57-660. AIEE Insulated Conduc-
tors Committee. June.
Neher, J. H. 1963. The Transient Temperature Rise of
Buried Cable Systems. Paper 63-917. IEEE Insulated
Conductors Committee. June.
3-57
Parmar, D. and J. Steinmanis. 2003. Underground
Cables Need a Proper Burial. Transmission & Distribu-
tion World. April. pp. 44-51.
Purnhagen, D. W. 1984. Designers Handbook for
Forced-Cooled High-Pressure Oil-Filled Pipe-Type
Cable Systems. Electric Power Research Institute. EL-
3624. Project 7801-5. July.
Stevenson, Jr., W. D. 1982. Elements of Power System
Analysis, 4
th
Edition. McGraw-Hill. New York.
Williams, J. A., T. R. Grave, and E. Kallaur. 1986.
Uprating of High Pressure Gas-Filled Feeders by
Fluid Filling and Rapid Circulation. IEEE Confer-
ence. Anaheim, CA. September 15-19.
Williams, J. A., E. C. Bascom III, B. Horgan, and T.
Aabo. 1991. Field Test Program and Results to Verify
HPFF Cable Rating. IEEE Transactions.
Williams, J. A. and J. H. Cooper. 1998. Distributed
Fiber Optic Temperature Monitoring and Ampacity
Analysis for XLPE Transmission Cables. Electric
Power Research Institute. TR-110630. June.
Williams, J. A. 2000. Application of Fiber-Optic Tem-
perature Monitoring to Solid Dielectric Cable. Electric
Power Research Institute. Publication 1000469. Novem-
ber.

3-58
APPENDIX 3.1 PIPE-TYPE AMPACITY
EXAMPLE
This appendix contains a sample calculation of pipe-
type cable ampacity using the procedures outlined in
this chapter and detailed in the references.
The cable configuration is as shown in Figure A3.1-1.
Figure A3.1-1 Cable configuration for pipe-type ampacity example.
3-59

tis 0.005 :=
Thickness of insulation shield tapes, inches
shield
0.7 :=
ohm-meters Electrical resistivity of stainless steel shield tapes
Skid Wire Size/Type 0.1in. x 0.2in., 2 x 3in. lay
skidwire
0.7 :=
ohm-meters Electrical resistivity of stainless steel
skidwire
f 60 := Hz
LF 0.62 =
Daily (24-hour) loss factor, per unit (entered below)
n 3 :=
Number of cables within pipe or conduit
N 2 :=
Pipe Data:
Pipe is HPFF
OD
pipe
10.75 :=
Pipe outside diameter, inches
ID
pipe
10.25 :=
Pipe inner diameter, inches
t
coating
0.07 :=
Pipe coating (Somastic) thickness, inches
coating
3.5 :=
Thermal resistivity of the pipe coating, C-m/w
Pipe-Type Cabl e Ampaci ty Worked Exampl e
Cables are 345kV HPFF cables with 2500kcmil segmental copper conductors, 905 mils kraft
paper insulation, 0.1x0.2, 2x3in. lay stainless steel skid wires in an 10-inch cable pipe, 2 circuits,
0.62 loss factor.
Cable Data:
A 2500 :=
Conductor area, CI
conductor
0.017241 :=
ohm-meters Electrical resistivity, copper conductor
ks 0.39 :=
Conductor skin effect factor, in oil
kp 0.46 :=
Conductor proximitty effect factor, in oil, cradled
Dc 1.824 :=
Diameter of the segmental conductor, inches
Tc 85 :=
Maximum normal conductor operating temperature, C
tcs 0.005 :=
Thickness of conductor semiconducting shield, inches
ti 0.905 :=
Insulation wall thickness, inches Values for
Paper
Insul ati on
SIC 3.5 :=
Dielectric constant of the insulation
tan 0.0023 :=
Dissipation factor of the insulation, numeric
insulation
6.00 :=
Thermal resistivity of the insulation, C-m/Watt
3-60

Center_Depth
cbf
39 :=
Value for center-line depth of cbf is arbitrary
Cal cul ate Cabl e Geometry:
D
cs
Dc 2tcs + :=
Diameter over the conductor shield, inches
D
insulation
D
cs
2ti + :=
Diameter over the insulation, inches
D
is
D
insulation
2tis + :=
Diameter over the insulation shield, inches
D
skidwire
D
is
1.5 0.1 + :=
Diameter over 0.1" skid wires, inches
D
earth
OD
pipe
2t
coating
+ :=
Clearance
ID
pipe
2
1.366D
skidwire
ID
pipe
D
skidwire
( )
2
1
D
skidwire
ID
pipe
D
skidwire
0.5
+ :=
Clearance in pipes, inches
D
cs
1.834 =
D
insulation
3.644 =
D
is
3.654 =
D
skidwire
3.804 =
D
earth
10.89 =
Clearance 1.992 =
Inst al l at ion Dat a:
x1 15 := x2 15 :=
Horizontal location of pipe center, inches
burial1 42 := burial2 42 :=
Burial depth to pipe center, inches
Ta 25 :=
Ambient earth temperature, C
f 60 :=
Power frequency, Hz
E 345000 1.05 :=
System maximum line to line voltage, volts
native
0.9 :=
Thermal resistivity of the native earth, C-m/w
backfill
0.5 :=
Thermal resistivity of the duct concrete, C-m/w
Width
cbf
53 := Height
cbf
29 :=
Width and height of backfill, inches
3-61

x1 x125.4 := x2 x225.4 := x1 381 =
Width
cbf
Width
cbf
25.4 := Width
cbf
1.346 10
3
= mm
Height
cbf
Height
cbf
25.4 := Height
cbf
736.6 = mm
Center_Depth
cbf
Center_Depth
cbf
25.4 := Center_Depth
cbf
990.6 = mm
Calculate the dielectr ic losses:
W
d
2fSIC
E
3
2
tan10
9
18ln
D
insulation
D
cond_shield
:= W
d
10.741 =
W/m/phase
Calculate Conductor Resistance:
R
dc20

conductor
Area
conductor
:= R
dcT
R
dc20
Tc 234.5 ( )
20 234.5 ( )
:= R
dc20
1.361 10
5
=
R
dcT
R
dcT
1.025 :=
Assume 2.5% Stranding of Conductor Ohms/meter
Xs
8f ks ( ) 10
7
R
dcT
:= Xp
8f kp ( ) 10
7
R
dcT
:=
Y
cs
Xs
2
192 0.8Xs
2
+
:= S D
skidwire
:= Y
cs
0.056 =
Y
cp
Xp
2
192 0.8Xp
2
+
D
conductor
S
2
0.312
D
conductor
S
2
1.18
Xp
2
192 0.8Xp
2
+
0.27 +
+
...
:=
Y
cp
0.061 =
R
acc
R
dcT
1 1.5Y
cs
Y
cp
+ ( ) +
:=
Metri c Conver si on of Variables
Area
conductor
A
1.9735
:= Area
conductor
1.267 10
3
= mm
2
D
conductor
Dc25.4 := D
conductor
46.33 = mm
D
cond_shield
D
cs
25.4 := D
cond_shield
46.584 = mm
D
insulation
D
insulation
25.4 := D
insulation
92.558 = mm
D
insl_shield
D
is
25.4 := D
insl_shield
92.812 = mm
D
skidwire
D
skidwire
25.4 := D
skidwire
96.622 = mm
OD
pipe
OD
pipe
25.4 := OD
pipe
273.05 = mm
ID
pipe
ID
pipe
25.4 := ID
pipe
260.35 = mm
D
earth
D
earth
25.4 := D
earth
276.606 = mm
burial1 burial125.4 := burial2 burial225.4 := burial1 1.067 10
3
= mm
3-62

Ohms/meter
Calcul ate Mutual Reactance (assume cr adled configuration)
X
m
2 2f ( ) 10
7
ln
2.3D
insl_shield
D
skidwire
:=
X
m
5.977 10
5
=
Ohms
Y
sc
R
s
R
acc
1
1
R
s
X
m
2
+
:=
Y
sc
1.314 10
3
=
R
acs
R
dcT
1 1.5 Y
cs
Y
cp
+ Y
sc
+ ( ) +
:=
Calcul ate the pipe loss increments to AC resistance:
Y
p
0.0438D
skidwire
0.0226ID
pipe
+
R
dcT
10
6
:= Y
p
0.578 =
R
acp
R
dcT
1 1.5 Y
cs
Y
cp
+ Y
sc
+ ( ) + Y
p
+
:=
R
acc
2.06 10
5
= R
acs
2.063 10
5
= R
acp
3.075 10
5
=
Ohms/meter
Qs
R
acs
R
acc
:= Qp
R
acp
R
acc
:= Qs 1.002 = Qp 1.493 =
AD/DC resistance ratios
Calcul ate the shiel d and ski dwi re loss i ncrements:
Area of shield tape is width of tape times thickness. Resistance of shield tape is the area times the helical
length times the resistivity divided by the area. Shield tape thickness is 0.005in stainless steel, with 1/8"
lapped, with typical tape width of 7/8". Assume there are 2 tapes.
width
shield_tape
7
8
25.4 := thickness
shield_tape
0.005 25.4 := lap
shield
1
8
25.4 :=
area
shield_tape
thickness
shield_tape
width
shield_tape
:= lay
shield
width
shield_tape
lap
shield
:=
R
shield

shield
area
shield_tape
1
D
insl_shield
lay
shield
2
+ := R
shield
3.804 =
Ohms/meter
Ski d wi re resistance
Area of elipse is pi * major_radius * minor_radius. Area of skid wire is half this. Skid wire is 0.1 x 0.2,
3-inch lay, 2 wires. Material is stainless steel.
minor_radius 0.1 25.4 := major_radius 0.2
25.4
2
:= area
skidwire

2
minor_radiusmajor_radius :=
lay
skidwire
3 25.4 :=
Length of a helix is (1 +(pi*D/Lay)^2)^.5
R
skidwire

skidwire
area
skidwire
1
D
skidwire
lay
skidwire
2
+ := R
skidwire
0.284 =
Ohms/meter
R
s
1
1
R
skidwire
1
R
skidwire
+
1
R
shield
+
1
R
shield
+
:= R
s
0.132 =
3-63

D
b
985.973 =
L
b
Center_Depth
cbf
:= L
b
990.6 =
G
b
ln
2L
b
4L
b
2
D
b
2
+
D
b
:= G
b
1.322 =
Cal culate diameter Dx
soil
6.71 10
4
native
100 ( )
0.8
:=
Thermal diffusivity of native soil
mm
2
hour

soil
1.834 10
3
=
D
x
1.02
soil
24 :=
Diameter beyond which 24 hour average losses apply, mm
D
x
213.978 =
Earth thermal
resistance for AC
losses at loss
factor
R
earth

backfill
2
n ln
D
x
D
earth
LFln
2burial1 4burial1
2
D
earth
2
+
D
x
:=
R
earth
0.381 =
Earth thermal
resistance for
dielectric losses at
100% loss factor
R
earth'

backfill
2
n ln
D
x
D
earth
ln
2burial1 4burial1
2
D
earth
2
+
D
x
:=
R
earth'
0.652 =
Cal culate Cable Thermal Resi stances
R
i

insulation
2
ln
D
insl_shield
D
conductor
:= R
i
0.663 =
U 0.26 := V 0 := Y 0.0026 :=
Constants for HPGF
Tm
oil
66.827 =
Estimate of mean temperature in duct
R
oil
nU
1 0.1 V YTm
oil
+ ( ) D
insl_shield
2.15 +
:= R
oil
0.175 =
R
pipe_coating

coating
2
nln
D
earth
OD
pipe
:= R
pipe_coating
0.022 =
Cal culate Ext ernal Thermal Resi stances
Cal culate geometric correction factor for backfil l envelope:
x Height
cbf
:=
x=short dimension of backfill
y Width
cbf
:=
y=long dimension of backfill
D
b
e
1
2
x
y
x
y
ln 1
y
2
x
2
+
ln x ( ) +
:=
3-64

T
d
W
d
R
i
2
R
oil
+ R
pipe_coating
+ R
earth'
+ R
mutual'
+ R
correction'
+
:= T
d
20.896 =
T Tc T
d
Ta := T 39.104 =
Adjust " i terated" values
bel ow unt il t hey equal
" calculated" .
I
rated
T
RacRth
:= I
rated
940.2 =
I
total
2I
rated
:= I
total
1880.4 =
Tc
calc
I
rated
2
R
acc
R
i
R
acs
R
oil
+ R
acp
R
pipe_coating
R
earth
+ R
mutual
+ R
correction
+ ( ) +
T
d
Ta + +
... :=
Tc
calc
85 =
Tshield
calc
Tc
calc
I
rated
2
R
acc
R
i
W
d
R
i
2
:= Tshield
calc
69.357 = T
shield
69.357
Toil
calc
Tshield
calc
I
rated
2
R
acs
W
d
+
R
oil
2
:= Toil
calc
66.827 = Tm
oil
66.827
T
earth_interface
Toil
calc
I
rated
2
R
acs
W
d
+
R
oil
2
I
rated
2
R
acp
W
d
+
R
pipe_coating
:=
T
earth_interface
63.477 =
Mutual thermal resistance between two pipes (j ust for normal ampacities)
d'
12
x1 x2 ( )
2
burial1 burial2 + ( )
2
+
.5
:=
Distance between one cable pipe and image
of other cable pipe
d
12
x1 x2 ( )
2
burial1 burial2 ( )
2
+
.5
:=
Distance between one cable and the other
F
12
d'
12
d
12
:= R
mutual

backfill
2
nLFln F
12
( ) := R
mutual
0.161 =
R
mutual'

backfill
2
nln F
12
( ) := R
mutual'
0.26 =
Thermal
resistance
correction for
native soil for AC
losses
R
correction
LF

native

backfill
( )
2
nNG
b
:=
R
correction
0.313 =
Thermal
resistance
correction for
native soil for
dielectric losses
R
correction'

native

backfill
( )
2
nNG
b
:=
R
correction'
0.505 =
Calcul ate Normal Ampacity on primary cabl e
RacRth R
acc
R
i
R
acs
R
oil
+ R
acp
R
pipe_coating
R
earth
+ R
mutual
+ R
correction
+ ( ) + :=
LF 0.62
3-65
APPENDIX 3.2 EXTRUDED AMPACITY
EXAMPLE
This appendix contains a sample calculation for a cross-
linked polyethylene (extruded) cable ampacity using the
procedures outlined in this chapter and detailed in the
references.
The cable configuration is as shown in Figure A3.2-1.
Figure A3.2-1 Cable configuration for extruded ampacity example
3-66

n 1 :=
System line to line voltage, volts
E 230000 :=
24 hour loss factor
LF 0.60 :=
Power frequency, Hz
f 60 :=
Earth ambient temperature summer, C
Ta 20 :=
Distance between adjacent circuits, inches
d
circuits
20 :=
Distance between adjacent phases, inches
d
phases
10 :=
Depth to center of top conduit, inches
Burial 42 :=
Install ati on Data:
Duct inside diameter, inches
Width and height of backfill, inches
Height
cbf
33 := Width
cbf
33 :=
Thermal resistivity of the duct concrete, C-m/w
backfill
0.50 :=
Thermal resistivity of the native earth, C-m/w
native
1.00 :=
Thermal resistivity of the duct, C-m/w
duct
6.00 :=
Thermal resistivity of the jacket, C-m/w
jacket
3.50 :=
Thermal resistivity of the insulation, C-m/w
insulation
3.50 :=
Ther mal Resi stivities:
Number of occupied ducts
N 6 :=
Number of conductors per duct
Thickness of conductor semiconducting shield, inches
tcs 0.0787 :=
Maximum allowable normal conductor operating temperature, C
Tc 90 :=
Diameter of the segmental conductor, inches
Dc 1.416 :=
Conductor proximitty effect factor, dry segmental
kp 0.6 :=
Conductor skin effect factor, dry segmental
ks 0.435 :=
ohm-meters
conductor
0.017241 :=
Conductor area, CI
A 1750 :=
Cabl e Data:
The ampacity calculation is for a 230kV 1750 kcmil segmented copper conductor circuit, with
866mils XLPE insulation, and lead metallic sheath, 2 circuits installed in a vertical duct bank and
0.60 loss factor. The sheath is cross bonded to minize circulating current.
Extruded Dielect ric Cabl e (XLPE) Ampaci ty Worked Exampl e
ID
duct
6.065 = ID
duct
OD
duct
20.28 :=
PVC Duct outside diameter, inches
OD
duct
6.625 :=
Duct Data:
Dissipation factor of the insulation, numeric
tan 0.001 :=
Dielectric constant of the insulation
SIC 2.3 :=
J acket thickness, inches
tj 0.125 :=
Thickness of lead moisture barrier / metallic sheath
tms 0.125 :=
Thickness of insulation semiconducting shield, inches
tis 0.0984 :=
Insulation wall thickness, inches
ti 0.866 :=
3-67

OD
duct
168.275 = OD
duct
OD
duct
25.4 :=
mm
D
jacket
101.656 = D
jacket
D
jacket
25.4 :=
mm tms 3.175 = tms tms 25.4 :=
mm D
met_shield
95.306 = D
met_shield
D
ms
25.4 :=
mm D
mean_shield
92.131 = D
mean_shield
D
mms
25.4 :=
mm D
insl_shield
88.956 = D
insl_shield
D
is
25.4 :=
W/m
W
d
1.144 = W
d
2 f SIC
E
3
2
tan 10
9
18ln
D
insulation
D
cond_shield
:=
Cal cul at e the di el ectri c l osses:
mm d
circuits
508.000 = d
circuits
d
circuits
25.4 :=
mm d
phases
254.000 = d
phases
d
phases
25.4 :=
mm Burial 1.067 10
3
= Burial Burial 25.4 :=
mm ID
duct
154.051 = ID
duct
ID
duct
25.4 :=
mm
D
jacket
D
ms
2tj + :=
D
mms
3.627 =
Mean metallic shield diameter, inches
D
mms
D
is
D
ms
+
2
:=
D
ms
3.752 =
Diameter over the metallic shield/moisture barrier, inches
D
ms
D
is
2tms + :=
D
is
3.502 =
Diameter over the insulation shield, inches
D
is
D
insulation
2tis + :=
D
insulation
3.305 =
Diameter over the insulation, inches
D
insulation
D
cs
2ti + :=
D
cs
1.573 =
Diameter over the conductor shield, inches
D
cs
Dc 2tcs + :=
Cal cul at e Cabl e Geomet ry:
mm D
insulation
83.957 = D
insulation
D
insulation
25.4 :=
mm D
cond_shield
39.964 = D
cond_shield
D
cs
25.4 :=
mm D
conductor
35.966 = D
conductor
Dc25.4 :=
mm
2
Area
conductor
886.749 = Area
conductor
A
1.9735
:=
Metri c Conversi on of Variabl es
Clearance 2.063 =
Clearance in conduit, inches
Clearance ID
duct
D
jacket
:=
D
jacket
4.002 =
Diameter over jacket, inches
3-68

R
acc
R
dc90
1 Y
cs
+ Y
cp
+ ( ) :=
R
acc
2.638 10
5
=
ohms/meter
Cal cul at e t he shi el d l osses:
Assume that the conductors will lie in a flat (vertical) configuration for purposes of calculating shield
losses. The metallic sheath consists of lead, and the bonding scheme is cross-bonded, so there will
be no circulating current and only eddy current losses.
Sheath material is lead

lead
21.410
8
:=
Area
sheath

D
met_shield
2
D
insl_shield
2
4
:= Area
sheath
918.965 = mm
2
Area
sheath
Area
sheath
10
6
:= m
2
R
sheath

lead
Area
sheath
:= R
sheath
2.329 10
4
=
Tm
shield
73.856 =
Estimated sheath temperature:
R
sheath
R
sheath
Tm
shield
236.0 ( )
20 236.0 ( )
:= R
sheath
2.819 10
4
=
Cal cul at e Conduct or Resi stance:
R
dc20

conductor
Area
conductor
:= R
dc20
1.944 10
5
=
ohms/meter
R
dc90
R
dc20
234.5 Tc +
234.5 20 +
:= R
dc90
2.479 10
5
=
ohms/meter
Assume 2.5% Stranding of Conductor
R
dc90
R
dc90
1.025 := R
dc90
2.541 10
5
=
ohms/meter
Xs
8 f ks ( ) 10
7
R
dc90
:=
Xs 2.581 =
Y
cs
Xs
2
192 0.8Xs
2
+
:= Y
cs
0.034 =
Xp
8 f kp ( ) 10
7
R
dc90
:= Xp 3.561 =
Y
cp
Xp
2
192 0.8Xp
2
+
D
conductor
d
phases
2
0.312
D
conductor
d
phases
1.18
Xp
2
192 0.8Xp
2
+
0.27 +
+
...
:= Y
cp
4.468 10
3
=
3-69

R
duct
0.084 = R
duct

duct
2
ln
OD
duct
ID
duct
:=
R
cable_to_duct
0.304 = R
cable_to_duct
5.2
1 0.1 0.91 0.01Tm
duct
+ ( ) D
jacket
( ) +
:=
Estimate of mean temperature in duct
Tm
duct
67.629 =
R
j
0.036 = R
j

jacket
2
ln
D
jacket
D
met_shield
:=
R
i
0.504 = R
i

insulation
2
ln
D
insl_shield
D
conductor
:=
Cal cul ate Cabl e Thermal Resi stances
R
acs
2.734 10
5
= R
acs
R
dc90
1 Y
cs
+ Y
cp
+ Y
sc
+ Y
se
+ ( ) :=
Single point bonded so no circulating currents
Y
sc
0 :=
Y
se
0.038 = Y
se
R
sheath
R
acc
g
s
Y
Se0
1 Y
Se1
+ ( )

1
tms ( )
4
1210
12
:=
g
s
1.008 = g
s
1
tms
D
mean_shield
1.74
1
D
mean_shield
10
3
1.6
+ :=
1
47.050 =
1
8
2
f
lead
10
7
:=
Y
Se1
3.852 10
4
= Y
Se1
0.86 m ( )
3.08
D
mean_shield
2d
phases
1.4 m 0.7 +
:=
Y
Se0
3.468 10
3
= Y
Se0
6
m
2
1 m
2
+
D
mean_shield
( )
2d
phases
:=
m 0.134 = m
2 f 10
7
R
sheath
:=
Eddy cur rent l oss increment
3-70

Diameter beyond which 24 hour average losses apply, mm
D
x
205.148 =
The self and mutual earth thermal resistances each consist of two terms. The first term assumes
that the concrete resistivity applies everywhere, the second term corrects for the excess of the
native earth resistivity over the concrete resistivity.
Middle phase has highest total thermal
resistance
L Burial d
phases
+ := L 1.321 10
3
= mm
R
earth

backfill
2
ln
D
x
OD
duct
LF ln
2L 4L
2
OD
duct
2
+
D
x
:= R
earth
0.171 =
Earth thermal resistance for AC losses
R
earth'

backfill
2
ln
D
x
OD
duct
ln
2L 4L
2
OD
duct
2
+
D
x
:= R
earth'
0.274 =
Earth thermal resistance for dielectric
losses
R
correction
LF

native

backfill
( )
2
n G
b
:= R
correction
0.082 =
Correction for native backfill thermal resistance
for AC losses
R
correction'

native

backfill
( )
2
n G
b
:= R
correction'
0.136 =
Correction for native backfill thermal resistance
for dielectric losses
Calculate Exter nal Thermal Resi stances
Cal cul ate geometr ic correcti on factor for backfil l envelope:
x Width
cbf
:= x x25.4 := x 838.200 = y Height
cbf
:= y y 25.4 := y 838.200 =
D
b
e
1
2
x
y
x
y
ln 1
y
2
x
2
+
ln x ( ) +
:= D
b
921.455 =
L
b
Burial d
phases
+ := L
b
1.321 10
3
=
G
b
ln
2L
b
4L
b
2
D
b
2
+
D
b
:= G
b
1.714 =
Cal cul ate di ameter Dx
n
6.7110
4
native
100 ( )
0.8
:=
Thermal diffusivity of native soil
mm
2
hour

n
1.685 10
3
=
D
x
1.02
n
24 :=
3-71

I
rated
T
cond
RacRthermal
:=
Ampacity per phase group, amperes:
I
rated
1101.4 =
I
total
2I
rated
:=
Total ampacity, amperes
I
total
2202.8 =
Check and adj ust the est imated shi el d and duc t ai r temper atur es:
Wc I
rated
2
R
acc
:=
Loss at the conductor, w/m
Wc 32.003 =
T's Tc Wc R
i
:=
Check temperature of shield which was estimated at
T's 73.856 =
Tm
shield
73.856
Ws I
rated
2
R
acs
:= Ws 33.169 =
T'm T's Ws R
j
R
cable_to_duct
2
+
:=
Check mean temperature between cable surface and
inside of duct which was estimated at:
T'm 67.629 =
Tm
duct
67.629
Mutual Heati ng
F1
2Burial d
phases
+
d
phases
2Burial 3d
phases
+
d
phases
:= F2
1
2 Burial d
phases
+ ( )
2
d
circuits
2
+
.5
d
circuits
:=
F2
2
2Burial d
phases
+ ( )
2
d
circuits
2
+
.5
d
circuits
2
d
phases
2
+
.5
2Burial 3d
phases
+ ( )
2
d
circuits
2
+
.5
d
circuits
2
d
phases
2
+
.5
:=
F F1F2
1
F2
2
:= F 1.262 10
4
=
Mutual thermal resistance for AC losses:
R
mutual

backfill
2
LF ln F ( ) N 1 ( ) n LF

native

backfill
( )
2
G
b
+ := R
mutual
0.860 =
Mutual thermal resistance for dielectric losses:
R
mutual'

backfill
2
ln F ( ) N 1 ( ) n

native

backfill
( )
2
G
b
+ := R
mutual'
1.434 =
Cal cul at e temperature r is e fr om di elect ri c heati ng
T
d
W
d
1
2
R
i
R
j
+ R
cable_to_duct
+ R
duct
+ R
earth'
+ R
mutual'
+ R
correction'
+
:= T
d
2.884 =
Cal cul at e the Ampaci ty
RacRthermal R
acc
R
i
R
acs
R
j
+
...
R
acs
R
cable_to_duct
+
...
R
acs
R
duct
+
...
R
acs
R
earth
+
...
R
acs
R
correction
+
...
R
acs
R
mutual
+
...
:= RacRthermal 5.533 10
5
=
C / A^2
T
cond
Tc T
d
Ta := T
cond
67.116 =

4-1
CHAPTER 4 Power Transformers
4.1 INTRODUCTION
Power transformers are a crucial link in the electric power system. Power transformer fail-
ures can result in extremely high financial losses, both in lost revenue and the capital
expenditure of unit itself. Therefore, it is essential that power transformers be operated in
a safe and prudent manner. In addition, given the high equipment cost, power transform-
ers must be operated in a manner to give a reasonable service life in order to realize a pos-
itive return on investment.
Power transformers represent a significant portion of capital investment costs. Under
existing conditions in the industry, utility budgets are reduced and networks are being
forced to support greater power transfer over existing transmission circuits than ever
before. As such, there is increased interest in safely utilizing all available capacity of
power transformers.
The thermal design of power transformers is based on an assumed daily average ambient
temperature of 30C, with a maximum of 40C. It is also based on continuous loading at
the rated maximum current. These assumptions are essential to the specification and
design of power transformers; however, they are probably never realized in service. In
most cases, these assumptions result in very conservative thermal ratings for liquid-filled
power transformers. Safe utilization of this latent capacity is the subject of this chapter.
In general, transformer load capacity is limited by equipment (winding and oil) tempera-
tures. Industry standards (IEEE C57.12.00 in the U.S.) specify a maximum average wind-
ing rise that defines the rated load. In other words, when operating at rated nameplate
current, the average winding rise shall not exceed the given value. For newer transformers
with thermally upgraded insulation, manufactured starting in the late 1960s, this maxi-
mum average winding rise is 65C. Combined with the 30C ambient and a 15C gradient
between the winding hottest spot temperature and the average winding temperature, the
peak temperature in the winding is 110C. This temperature represents a somewhat arbi-
trary benchmark temperature that gives a reasonable life expectancy based upon the col-
lective experience of the industry. Earlier transformers built before the introduction of
thermally upgraded insulation were based upon a rated average winding rise of 55C, or a
hot spot of 95C.
Chapter 4: Power Transformers Increased Power Flow Guidebook
4-2
Section 4.2, Transformer Design, describes the gen-
eral construction of power transformers, types of
cooling, losses generated by transformers, and fac-
tory testing.
Section 4.3. Risks of Increased Loading, outlines
short- and long-term risks related to the loading of
transformers.
Section 4.4, Thermal Modeling, provides an overview
of heat transfer mechanisms and describes the four
most prevalent thermal models.
Section 4.5, Thermal Ratings, discusses factors
behind thermal ratings, including ambient air tem-
perature, load, and maintenance considerations.
Section 4.6, Winding Temperature Measurement,
describes methods of measuring and monitoring
winding temperature.
Section 4.7, Modest Increases in Capacity from Exist-
ing Transformers, outlines methods of increasing
cooling.
Section 4.8, Examples, provides two examples of
increasing capacity.
4.2 TRANSFORMER DESIGN
4.2.1 General Construction
Power transformers are commonly categorized into two
main construction configurations: core form and shell
form. These two categories are clearly defined in the
case of three-phase power transformers. The definition
becomes less distinct in the case of single-phase trans-
formers, especially in the smaller sizes. In this range, the
distinction is of minor importance, however.
Core-Form Construction
A core-form power transformer is constructed with
windings that are in the general form of concentric cyl-
inders (Figures 4.2-1 and 4.2-2). The magnetic core con-
sists of stacks of laminated grain-oriented silicone steel
sheets. The width of these sheets is varied so that, as
stacked, the cross section of each leg conforms as nearly
as practical to the circular form of the innermost wind-
ing. With very few exceptions, the winding assemblies
and core legs are assembled with their axis in a vertical
or upright plane.
The magnetic circuit is completed with yoke assemblies
that bridge the three legs at the top and bottom. On
some of the larger transformers, part of the required
cross section of the yokes might be provided by outer
legs at each end. These are referred to as fifth legs and
allow a reduction in the vertical dimensions of both the
top and bottom yokes. This reduction is often an impor-
tant factor in meeting shipping clearances since core
form power transformers are shipped with the coils and
core legs in an upright or vertical position.
The legs and yoke assemblies are joined with mitered
interfaces that have alternating steps that overlap to
provide good magnetic coupling. Modern practices are
to use laminations of 0.009-in. thickness with at least
five overlapping steps. This pattern is repeated across
the entire build of the core assembly. The legs and yokes
are maintained as an independent assembly by banding
the legs with a nonconducting material such as fiber-
glass and in some cases, metallic bands that must be
provided with an insulating gap to prevent circulating
currents to flow around the band.
Ducts for the flow of cooling oil are provided between
concentric winding sections and sometimes within wind-
ing sections if required by the thermal characteristics of
the design (Figures 4.2-3 and 4.2-4).
Cooling ducts may also be required in the core legs
and/or yokes (Figures 4.2-5 and 4.2-6). These ducts are
formed by providing spacers between laminations at dis-
crete intervals determined by the dimensions of the core.
These ducts between laminations are usually mirrored in
the top and bottom yokes.
Figure 4.2-1 Typical core-form construction (front view). Figure 4.2-2 Typical core-type internal assembly.
4-3
Increased Power Flow Guidebook Chapter 4: Power Transformers
In a core-form transformer, the winding design permits
cooling oil to flow vertically through the concentric
windings and the core leg ducts (Figure 4.2-7).
Shell-Form Construction
In shell-form power transformers, the windings consist
of phase assemblies produced by winding individual sec-
tions in a flat plane on a winding machine (Figures 4.2-8
and 4.2-9). These windings are initially assembled flat
with insulation and cooling ducts between sections.
Complete phase assemblies are then clamped and ori-
ented in a vertical direction so that the plane of the indi-
vidual sections are upright. The core laminations are
inserted through and around the windings in the final
Figure 4.2-3 Core-type winding assembly with
concentric cooling ducts.
Figure 4.2-4 Cooling duct being assembled on winding.
Figure 4.2-5 Inner winding assembled on core
(note oil ducts in core and below windings).
Figure 4.2-6 Cross section of core form construction.
Figure 4.2-7 Oil flow in core-type winding.
4-4
assembly process. The core and coil assembly is not
designed to be lifted as a unit in most cases.
In shell-form construction, the oil flows vertically
through the ducts between individual winding sections
(Figure 4.2-10). The spacers in these ducts are posi-
tioned to direct the oil across the width of each winding
section.
Directed Flow Designs
When forced oil cooling is employed, the transformer
internal assembly may be designed with oil manifolds
that direct the incoming cool oil to the lower part of the
core and windings. In a nondirected flow transformer,
the incoming cool oil is pumped into the bulk oil of the
lower tank and moves through the core and windings in
much the same manner as normal convection currents
but with greater velocity (Figures 4.2-11 and 4.2-12).
A directed flow design is normally used with larger sizes
of transformers equipped with heat exchangers rather
than radiators (Figures 4.2-13 and 4.2-14).
Figure 4.2-8 Typical shell-form construction (plan view).
Figure 4.2-9 Shell-form transformer.
Figure 4.2-10 Shell-form phase assembly under
construction (cooling ducts are assembled between
sections).
Figure 4.2-11 Core-form transformer designed for directed
flow.
4-5
4.2.2 Types of Cooling
Oil-immersed power transformers are cooled by four
different cooling configurations, each with unique heat
transfer. They are as follows:
ONAN (OA) Also referred to as self-cooled, no
pumps are used to circulate oil, and no fans are used
to increase airflow over the radiators. Oil circulates
upward through the windings and down through the
radiators by natural thermosiphon flow.
ONAF (FA) No pumps are used to circulate the oil.
Fans are used to force air over the radiators to
increase heat transfer from the bulk oil to the sur-
rounding air. As with ONAN, oil circulates upward
through the windings and down through the radia-
tors by natural thermosiphon flow.
OFAF (Non-directed FOA) Pumps are used to cir-
culate the oil. Fans are used to force air over the radi-
ators. The forced circulation of the oil increases the
convective heat transfer from the windings to the oil.
The forced air increases the convective heat transfer
from the oil to the air. With OFAF, there are no ducts
to direct the oil over the winding. In general, the bulk
of the forced oil flow passes upward between the
winding and the tank, bypassing the windings. Natu-
ral convection is still the predominant mode of heat
transfer from the windings to the adjacent oil.
ODAF (Directed FOA) Pumps are used to circulate
the oil. Fans are used to force air over the radiators or
heat exchangers. The forced circulation of the oil
increases the convective heat transfer from the wind-
ings to the oil. The forced air increases the convective
heat transfer from the oil to the air. With ODAF,
ducts are added to direct the oil over the winding.
This forces a significant portion of the forced oil to
flow upward through the vertical winding ducts.
Washers may also be employed in a zig-zag fashion to
force the oil to flow back and forth between disk sec-
tions of a disk winding, further increasing the convec-
tive heat transfer from the winding to the oil.
4.2.3 Losses
As with all electrical apparatus, transformers generate
losses. These losses are manifested in the form of heat.
This heat increases the temperature of the transformer
components, and therefore limits the load capacity.
Therefore, it is essential to discuss losses when discuss-
ing loading. The losses in a transformer are due to
several different mechanisms, but are generally lumped
into two categories: load losses and no-load losses
(Figure 4.2-15). Load losses are losses that vary with
load current, but not with excitation. No-load losses are
losses that do not vary with load current, but rather
vary with excitation or voltage.
Figure 4.2-12 Directed oil flow in core-form
construction.
Figure 4.2-13 Heat exchangers assembled on
transformer.
Figure 4.2-14 Radiators assembled on transformer.
4-6
Load losses and no-load losses may be further broken
down. Load losses consist of three components: winding
I
2
R losses, winding eddy losses, and stray losses. Wind-
ing I
2
R losses are the Ohmic losses due to the load cur-
rent flowing in the windings. Skin effect is generally
ignored, but would be included in the I
2
R losses.
Depending upon the conductor dimensions, the increase
in winding I
2
R losses may or may not be negligible.
In addition to the intended load current, unintended
eddy currents flow within the winding conductor due to
the presence of leakage flux, or flux that does not flow
within the core. These eddy currents result in additional
Ohmic losses in the winding conductors. The eddy losses
can be reduced by using a stranded conductor, breaking
the path for eddy current flow into smaller regions.
If parallel strands or conductors are used, circulating
currents between parallel sections can flow due to differ-
ing amounts of flux linkage. These losses can be reduced
by transposing the cable, as in the cases of CTC (contin-
uously transposed cable) or transposing the winding
conductors at various points along the winding.
Stray losses comprise the remainder of the load losses.
Stray losses are losses due to currents induced in various
metallic structures of the transformer, including the
winding clamping structure, tie plates, core bolts, etc.
Stray losses are parasitic in nature, and difficult to cal-
culate. In some cases, the losses can result in high local-
ized temperatures of various metal components. Often,
in larger power transformers and GSUs (generation step
ups), this is a serious concern and must be carefully
examined during design and construction.
No-load losses are composed mainly of losses in the
steel core. The steel that comprises the core contains
many microscopic regions of like magnetic polarity,
called magnetic domains. When magnetic flux passes
through the core, these domains shift to orient their
magnetic poles in the direction of the flux. Since the flux
varies sinusoidally, the domains must change direction
as the flux changes direction. This continuous move-
ment of the magnetic domains produces losses. These
losses are termed hysteresis losses.
Like the winding conductors, the flux passing through
the steel core, which is electrically conductive, induces a
current flow around the core. The magnitude of these
induced currents is reduced by dividing the core cross sec-
tion into thin sheets or laminations. Each sheet is coated
with an insulating film. This reduces the total eddy cur-
rent, and therefore the eddy loss, in the steel core.
The core losses in any transformer are dependent on the
core flux density. The core flux density is determined by
the magnitude and frequency of the primary voltage
and the cross-sectional area of the core leg according to
Equation 4.2-1.
4.2-1
Where:
B
m
= maximum flux density in Tesla.
f = operating frequency.
A = cross-sectional area of core in mm
2
.
V = primary voltage (RMS) per phase.
T = number of primary winding turns per phase.
The flux density, and therefore the principal core losses,
are thus independent of load. There may be some local-
ized heating of the outer core laminations or the core
support structure by stray flux due to winding current
that will be exacerbated by overloads. These losses are
usually of little consequence to either short-term or
long-term reliability of the transformer because they are
removed from the winding insulation and affect the oil
only in the local area involved.
If a transformer has displayed a tendency, in normal
service, to produce methane or ethane close to or in
excess of condition #1 as described in ANSI C57.104,
dissolved gas analyses should be performed at daily
intervals when overloading to a level that has not been
previously experienced. Continued operation at that
level may normally be considered safe from this point of
view if gas accumulation rates do not exceed condition
#2 as detailed in ANSI C57.104.
4.2.4 Factory Testing
Little information on a particular power transformer is
available to the user. Generally, the only information a
user has on a unit is the information given on the fac-
tory test report and on the nameplate. Combined, these
reports give the user key performance characteristics
needed to operate the transformer. Included in this
information is the heat run data, giving the temperature
rises of the windings and the oil at the benchmark load-
ing of the nameplate rating, and the loss data. In order
Figure 4.2-15 Transformer equivalent circuit.
3
10
225

=
T
V
fA
B
M
4-7
to interpret the data given on a factory test report prop-
erly, at least a passing understanding of the underlying
test is necessary. This understanding will assist the
reader in interpreting the numbers and any remarks that
may be associated with the test values.
Factory Loss Tests
No-Load Loss Testing
No-load loss testing is performed with one winding
open-circuited and rated voltage applied to the other
winding. For practical reasons, voltage is usually applied
to the low-voltage winding. This results in full excitation
on the transformer, with only the exciting current flow-
ing through the excited winding (Figure 4.2-16).
Load Loss Testing
The load losses in an operating transformer consist of:
I
2
R losses in the current carrying components due to
load current in all active windings.
Joule losses caused by circulating currents in winding
conductors due to impedance inequalities in parallel
conductors.
Eddy currents in winding conductors due to radial
gradients in the induced voltages.
Stray losses due to induced currents in the tank walls
and internal support structures.
Load losses are measured by short-circuiting either the
high-voltage or low-voltage winding and then applying
a voltage to the other winding that results in rated cur-
rent flowing through the windings (Figure 4.2-17). The
power loss measured, as measured by the wattmeter, is
then the load loss at the temperature of the windings
during the test. Since losses are dependent upon temper-
ature, the load losses must be corrected to a standard
reference temperature.
Prior to the test, the transformer is allowed to sit de-
energized long enough for the oil temperature to stabi-
lize. The temperature of the windings is then assumed to
be equal to the average oil temperature. The load losses
are corrected to a standard reference temperature of
20C plus the rated average winding rise. For 55C or
55/65C transformers, the reference temperature is gen-
erally 75C. For 65C rise transformers, the reference
temperature is generally 85C. This should be explicitly
specified on the test report.
The I
2
R losses vary with temperature in a predictable
fashion. The difficulty is in determining the appropriate
temperature for an operating condition since the tem-
perature variation between parts of a winding can be
considerable.
ANSI standards require reporting of transformer losses
at a reference temperature of 85C for 65C transformer
ratings and 75C for 55C transformer ratings.
For copper, these I
2
R losses are corrected to an operat-
ing temperature (T) as shown in Equation 4.2-2.
4.2-2
Where:
Pr(T) = I
2
R loss due to load current at operating
temperature T (C).
T
m
= reference temperature of the test report
losses.
P
R
(T
m
) = test report I
2
R losses at T
m
.
(The test report will give total load losses at the refer-
ence temperature and the measured winding resistances
corrected to this temperature. The tested I
2
R can be cal-
culated using the test report resistance at the reference
temperature and rated winding current. Note that the
test report resistance values for each winding are equal
to the sum of the three phases as required by ANSI
standards. Use of rated winding phase current with the
test report resistance allows calculation of the three-
phase I
2
R loss for each winding.)
The stray and eddy losses vary inversely with tempera-
ture. For copper, these losses are corrected as shown in
Equation 4.2-3.
P
s
(T) = P
s
(T
m
) (234.5 + T
m
)/(234.5 + T) 4.2-3
Figure 4.2-16 Basic circuit for no-load loss testing. Figure 4.2-17 Basic circuit for load loss testing.
m m
PR(T) PR(T ) (234.5 T)/(234.5 T ) = + +
4-8
Where:
Ps(T) =stray losses at operating temperature
T (C)
T
m
=reference temperature of the test report
losses
P
s
(T
m
) =test report load losses at T
m
minus P
R
(T
m
)
calculated above.
The load losses at operating temperature T are the sum
of P
R
(T) and P
s
(T).
Factory Temperature Rise Tests
The basis for all practical transformer thermal models
used in the evaluation of transformer loading is the fac-
tory heat run. The factory heat run provides a direct
measurement of the thermal performance of a trans-
former at a particular benchmark, the nameplate rating.
Given the temperature rises at this benchmark loading,
it is possible to estimate the temperature rises under
other loading conditions. Therefore, to understand the
thermal models and their application to transformer
loading and rating calculations, one must know the
basic principles behind the factory heat run. This
knowledge will prove invaluable when interpreting the
data presented on the certified test report.
This test must be performed on the tap position and
winding connection that gives the highest winding tem-
perature as required by ANSI standards. The tempera-
ture rises over ambient should be given in the certified
test report for each windingalong with the tap posi-
tion, total losses, and line currents. The calculated wind-
ing hot spot temperature rise over ambient should given
for the maximum transformer rating.
Two methods are used to perform factory heat runs: the
short-circuit method and the loading-back method.
Of these two methods, the short-circuit method is by far
the most common and will be covered in the greatest
depth here. In addition to these two methods, it is also
possible that a heat run may be performed by connect-
ing the transformer to a load and loading the trans-
former at rated voltage and current simultaneously. This
method is generally only practical for small power trans-
formers and is not often used.
Short-Circuit Method
The short-circuit method, also sometimes referred to as
loss injection, consists of two segments. In both seg-
ments of the test, one winding is short-circuited and a
voltage is applied to the other winding. In this manner,
each winding temperature can be checked. For trans-
formers with multiple windings, the test load levels are
usually agreed upon before hand, and each winding is
tested with the other windings short-circuited.
For the initial segment of the test, one winding is short-
circuited and a voltage is applied to the other winding
such that the active power supplied equals the total
losses (load losses plus no-load losses) previously mea-
sured by other means (refer to IEEE C57.12.90 or IEC
76-1). Note that temperature rise tests are usually per-
formed at the combination of taps that gives the highest
losses. The power is applied continuously to bring the
oil temperatures up to their steady-state values. Steady-
state is defined in the respective test standards, but usu-
ally is such that the average oil temperature has not
increased more than 1C (or 2.5%, whichever is greater)
for three consecutive hours. The purpose of this segment
of the test is to establish the oil temperature rise over
ambient.
The second segment of the test is to measure the winding
temperature rises. Once a steady-state average oil tem-
perature has been reached, the applied voltage is reduced
such that rated current flows in the windings under test.
This load is held continuously for one hour. At the con-
clusion of one hour, the winding resistance is measured
shortly after the windings have been disconnected,
allowing just enough time for the inductance effects to
dissipate. The winding resistance is measured several dif-
ferent times after shutdown, recording the resistance and
time after shutdown each time (Figure 4.2-18). These
points are then used to extrapolate the winding resis-
tance back to the instant of shutdown. With this wind-
ing resistance and the cold winding resistance (measured
prior to the heat run with the windings at a known ambi-
ent temperature), the average winding rise at the instant
of shutdown can then be determined. This procedure
may have to be repeated for each winding unless the
available test equipment and transformer design allow
simultaneous resistance readings on all windings.
After completion of the above procedures, the oil tem-
perature rises and average winding rises for each wind-
ing are known, given the test conditions. Since it is not
Figure 4.2-18 Example winding cooldown curve illustrating
extrapolation of resistance measurements to shutdown.
4-9
always possible to maintain the exact loadings required
to give the total loss or rated current, the raw results
obtained above must be corrected to the rated losses and
loads. This is done using simple empirical equations as
shown in Equations 4.2-4 and 4.2-5.
For oil temperatures:
4.2-4
For winding temperatures:
4.2-5
Where:
o,corr
= oil temperature corrected to rated losses.
o,meas
= oil temperature measured during the heat
run.
P
t,rated
= total losses at rated load.
P
t,meas
= total losses applied during oil rise portion
of the heat run.
m = oil rise exponent (0.8 for ONAN, ONAF,
OFAF; 1.0 for ODAF).
w,corr
= average winding rise corrected to rated
load.
w,meas
= average winding rise measured during
test.
I
w,rated
= rated winding current.
I
w,meas
= winding current measured during the
winding rise portion of the test.
n = winding rise exponent (0.8 or ONAN; 0.9
for ONAF; 1.0 for OFAF, ODAF).
Loading-Back Method
This method is rarely used, and as such will not be
described in any great detail here. For more informa-
tion, consult the relevant test standards, IEEE
C57.12.90 and IEC 76-2.
4.3 RISKS OF INCREASED LOADING
Any energized transformer has a finite risk of failure.
This risk of failure is generally very low. Loading trans-
formers above the nameplate rating results in higher
risks, eventually reaching the extreme where failure
would be certain. The challenge is to determine loading
levels that represent an acceptable level of risk, given the
circumstances.
There are several risks related to the loading of trans-
formers. The most significant of these risks are long-
term degradation of the insulation paper and short-term
reduction in dielectric strength due to bubble formation.
Understanding these risks and their mechanisms is
essential to any discussion on transformer loading.
Users must be aware of the risks, take steps to assess
these risks, and develop a methodology for maintaining
risks at acceptable levels.
4.3.1 Short-Term Risks
Risks classified as short-term are those that would
result in the immediate failure of the transformer.
Reduction in dielectric strength due to gas evolution
from the winding conductor insulation is the principal
short-term risk. Additional short-term risks include gas
evolution from lead conductors, gas evolution from
wood and cellulose materials adjacent to metallic hot
spots from stray flux heating, delamination of phenolic
resin tap boards, and reduction of short-circuit with-
stand capability in epoxy-bonded CTC. Some of these
risks will not result in failure in and of themselves, but
when combined with an abnormal transient voltage or
throughfault, may result in failure.
Bubble Evolution
Through destructive testing of model coils, researchers
have discovered that under certain loading conditions, a
significant decrease in dielectric strength of the winding
assembly occurs. In the transformers considered here,
specially refined petroleum oil is used as an electrical
insulation material and, concurrently, as a heat transfer
fluid. Gas-filled bubbles can be formed in the trans-
former oil in units in service. The electrical breakdown
strengths of these gases are significantly lower than that
of the transformer oil surrounding these bubbles. Bub-
bles that find their way into an electrically stressed oil
gap can significantly reduce the breakdown voltage of
the gap and interfere with the functioning of a trans-
former (Figure 4.3-1).
m
meas t
rated t
meas o corr o
P
P
=
,
,
, ,

n
meas w
rated w
meas w corr w
I
I
2
,
,
, ,
=
Si xty-Hertz Breakdown Voltage in Percent of 25C Strength
0
10
20
30
40
50
60
70
80
90
100
0 50 100 150 200 250 300
Hot Spot Temperature, C
B
r
e
a
k
d
o
w
n

V
o
l
t
a
g
e
,
%
Paper Insulated
Winding Conductor -
Dry (<0.5%)
Paper Insulated
Cable - Dry
(<0.5%)
Paper Insulated
WindingConductor -
Wet (3%)
Figure 4.3-1 Example of 60-Hz breakdown vs.
temperature illustrating bubble formation.
4-10
Henrys Law states that the concentration, C
i
, of a single
ideal gas, i, dissolved in a liquid is proportional to the
partial pressure of that gas, P
i
, over the liquid:
C
i
= K
i
* P
i
4.3-1
At equilibrium, the drive of the dissolved gas to escape
is balanced by the drive of the gas over the liquid to
enter the liquid. K
i
is a constant dependent on tempera-
ture, T, in a typical exponential manner:
K
i
~ exp (B
i
/T) 4.3-2
The magnitude of K
i
and its dependence on temperature
reflect the particular intermolecular interactions of a
given gas liquid combination and are unique to that
combination. B
i
can be either positive or negative alge-
braically; it is negative for gases whose solubility
increases as the temperature decreases.
If, at constant temperature, the partial pressure of the
gas over a liquid is decreased, the liquid will be supersat-
urated and gas will vaporize in order to reach a new
equilibrium concentration as defined by Henrys Law.
Likewise, if the temperature changes so that K
i
becomes
smaller while the pressure of the gas over the liquid is
held constant, additional gas will make its escape. In the
reverse of these happenings, gas will dissolve.
When the reduction in pressure of the gas over the liq-
uid is slow or the local mass transport processes are rel-
atively fast, the dissolved gas will diffuse through the
liquid to and across the gas-liquid interface. The gas will
steam away. The escape of the gas tracks the slow
decrease in the external partial pressure.
If, however, the pressure of the gas over the liquid drops
so rapidly that dissolved gas does not have time to move
to the surface, local supersaturation can occur and bub-
bles can nucleate. The bubble presses aside the sur-
rounding liquid, and continues to grow as it absorbs
supersaturating gas from the liquid. A bubble will grow
so long as the pressure within the bubble exceeds the
external pressure. The primary constraint to bubble
growth is the pressure of the atmosphere over liquid.
The effect of surface tension of the liquid at the bubble
surface and the hydrostatic pressure of the liquid over
the site of bubble growth are minor.
An example warm beer bubbles and foams when the
bottle cap is removed and the pressure of carbon dioxide
(CO
2
) in the neck of the bottle is suddenly released.
Colder beer bubbles and foams less, in part because the
pressure to be released in the neck of the bottle is less,
but also because CO
2
is more soluble in cold water
(B
CO2
is algebraically negative).
Oil in transformers normally contains more than one
gas. Some transformers are designed to allow access of
air to the interior. The gases dissolved in the oil are
those present in air: nitrogen, oxygen, carbon dioxide,
and traces of other gases. Other transformers are sealed
and evacuated and then filled with nitrogen. The pri-
mary gas dissolved in the oil is nitrogen. Only small
amounts of oxygen and traces of carbon dioxide remain.
In general, gases dissolved in transformer oil behave as
ideal dilute solutions in that the solubility of one gas is
independent of the presence of any other gas; e.g., K
N2
is the same whether CO
2
is present or not, and C
N2
is
not affected by P
CO2
. However, in forming bubbles,
gases behave collectively. Each gas dissolved in the oil
contributes to the internal pressure of the bubble
according to Henrys Law, shown in Equation 4.3-3.
4.3-3
A bubble forms when the sum of the partial pressures
resulting from local concentrations of gases in the liquid
exceeds the pressure over the surface of the liquid phase
P
atm
.
The major solid electrical insulation in typical trans-
formers is cellulose in the form of paper and press board
impregnated with transformer oil. Water is absorbed by
the cellulose. It distributes itself between the bulk oil
and paper so that the concentrations of water in the
solid and liquid media are in equilibrium with the water
vapor pressure over the insulation system. Unlike nitro-
gen and oxygen, water vapor is a condensable gas.
A very high percentage of the total mass of water in the
system is contained in the paper. In most transformers
the amounts of free oil and impregnated cellulose are
comparable. In a subsystem containing equal volumes
of paper (with 0.5% moisture content by weight) and
transformer oil and equilibrated at 80C, the paper will
contain approximately 99.9% of the total weight of
water. In effect, the paper in a transformer acts as an
almost infinite reservoir supplying additional water to
the oil at the higher internal temperatures occurring
during operation. Water will re-absorbed on the paper
when the electrical load on the transformer is reduced
and will again be available for desorption when the
transformer is next loaded.
Desorption of water from the cellulose in the oil/paper
insulations system is an example of a physical reaction
that can increase the local concentration of gases in the
oil to produce bubbling. Local concentrations can also
change as a result of chemical reactions within the system.
bubble i i i atm
P = P = (C / K ) > P
4-11
Thermal decomposition of cellulosic insulation can
occur in an operating transformer. The decomposition
reaction can increase the local concentration of gases in
the oil. The primary products of thermal decomposition
of cellulose are water vapor, CO
2
, and CO. Decomposi-
tion of the oil/paper insulation at a sufficiently high
temperature generates gases more rapidly than they can
diffuse away into the surrounding oil. Again, the local
concentration of gases builds to the point where the
summation of partial pressures exceeds the pressure in
the transformer and a bubble forms.
To summarize, formation of bubbles in transformer oil
reflects a complex interplay of variables in an operating
transformer. Bubbles result from an abrupt departure
from equilibrium between dissolved gas and oil result-
ing from some change in pressure and/or local tempera-
ture in a transformer. Bubbles form when a local
increase in gas concentration results in a local pressure
that exceeds the pressure over the oil.
Evaluation of bubble formation is difficult. The mecha-
nisms are complex and dependent upon local condi-
tions. The most influential factor is the moisture
content of the paper. The temperature at which bubbles
form is highly dependent upon moisture content, as
illustrated in Figure 4.3-2. Note that this plot is for an
example loading condition, and therefore, the values
depicted are not to be taken as absolute truths.
Given the importance of moisture in the determination
of bubble evolution risk, accurate determination of
moisture content would be necessary to use any sort of
analytical model. Currently, there are no means of
determining the moisture content distribution through-
out an operating transformer with the required preci-
sion. Moisture content in oil is often related to moisture
content in the paper using equilibrium curves. This,
however, has two problems. First, the moisture distrib-
utes throughout the winding inversely proportional to
temperature. The moisture in oil could only give an
average content of the entire bulk insulation. Second,
the paper-oil system is rarely at equilibrium. The time
constant of moisture diffusion is on the order of several
days or weeks.
Without any practical means to precisely assess the risk
of bubble evolution, we must fall back on general guide-
lines that have been proven by experience. The general
consensus in the industry is that bubble evolution does
not occur at hot spot temperatures up to 140C. This is
probably valid for moisture contents up to 1.52%.
Newer, dryer transformers could probably be safely
loaded to temperatures approaching 160C. Transform-
ers in service for less than 6 months should be limited to
temperatures of 120C.
Oil Expansion
As with any fluid, oil expands as temperature increases.
Specifically, the coefficient of expansion for oil is
0.000756, or the oil expands 0.08% for every degree C
rise in temperature. This expansion is anticipated, and
transformers are equipped with devices to handle oil
expansion to some degree. IEEE C57.12.10, the IEEE
standard for power transformers below 230 kV, requires
that a transformer be capable of operating with temper-
ature to 105C. If the oil temperature exceeds 105C, the
oil may expand beyond the capacity of the gas space or
conservator tank, resulting in operation of the mechani-
cal pressure relief device and the expelling of oil. Not
only does this present a maintenance and environmental
headache, this can cause problems if enough oil is
expelled such that upon cooling the oil level drops to a
level exposing the active parts. This could result in a
dielectric failure.
4.3.2 Long-Term Risks
Aging of Cellulose and Oil in Transformers
For liquid-immersed transformers, the insulation system
consists of the paper and pressboard solid insulation
(cellulose) and the oil liquid insulation. Both the cellu-
lose and liquid insulation are adversely affected by heat.
However, while the liquid insulation can be replaced or
reprocessed relatively easily, the properties of the solid
insulation cannot be easily or effectively restored, nor
can the insulation be economically replaced. Therefore,
the loss of the paper insulation life is of utmost impor-
tance when considering transformer load capability.
Oil Aging
Oils in transformers are mainly mineral oils. The oils
can be grouped in two types: inhibited oils, which are
heavily refined, reducing the amounts of aromatics and
100
110
120
130
140
150
160
170
180
190
200
0 0.5 1 1.5 2 2.5 3 3.5
Moisture Content (%dry wei ght)
G
a
s

E
v
o
l
u
t
i
o
n

T
e
m
p
e
r
a
t
u
r
e

(
d
e
g

C
)
Figure 4.3-2 Gas evolution temperature vs. moisture
content for example conditions.
4-12
polyaromatics, and then inhibited with an antioxidant
to reduce oxidation, and the uninhibited oils, where nat-
ural aromatics and polyaromatics are natural inhibitors.
When oils are subjected to oxygen and heat, they start
to oxidize: the aldehydes, ketones, and finally carboxylic
acids, presumably with a high molecular weight, are
formed. The neutralization value and loss factor of the
oil gradually increase, and the interfacial tension is
reduced. At an even later stage, sludge may form. The
oxidation rate increases with increasing temperature.
For an inhibited oil, first the inhibitor is consumed (the
period this takes is called the induction period). When it
is fully consumed, the oils quickly start to oxidize.
Before the inhibitor is gone, one may re-inhibit the oil,
but the induction period is now be shorter, unless the oil
is reclaimed.
The main risk from the oil aging is sludge formation and
clogging of cooling channels. The oil aging is not gov-
erned by the oxygen concentration but by catalytic pro-
cesses, where hydroxyl radicals are formed via interaction
with active metals like copper and iron.
The oil condition may be followed by several diagnostic
markers: Neutralization value, interfacial tension, loss
angle, inhibitor content, and sludge measurement. It is
important to take maintenance precautions before the
condition becomes too severe.
Cellulose Aging
Paper and pressboard are made from cellulose fibers
(Figure 4.3-3). They contain 93% cellulose, with mole-
cules that are long, straight polysaccharide chains,
which may form crystalline regions. The average number
of cellulose molecules, called the degree of polymeriza-
tion (DP), describes the length of the cellulose molecule.
The cellulose also contains 5-6% hemicelluloses, which
are polysaccharide chains with branches that do not eas-
ily crystallize, and lignin, which is a large branched mol-
ecule. The last two act as a thermoplastic.
When cellulose ages, the cellulose chains are cut in a
process called chain scission, reducing the average
length of the cellulose chains and resulting in shorter
fibers. The number of chain scissions () is defined as
= DP
new
/DP
old
1, where the DP value is taken before
and after the aging period.
Paper may be chemically modified to allow for opera-
tion at higher temperatures (thermally-upgraded paper).
Upgraded paper is paper that is chemically changed by
different chemical additives that reduces the aging rate
compared to untreated kraft paper. A paper is consid-
ered thermally upgraded if it meets the life criteria as
defined in ANSI/IEEE C57.10050% retention in ten-
sile strength after 65,000 h in a sealed tube at 110C.
However, upgrading can be done in various ways, so it is
not a uniform group and may react differently to chemi-
cal aging accelerators. To retard aging, one can ther-
mally upgrade the paper by linking bulky substituents
such as cyanoethyl ether groups to the HO-groups in the
cellulose and hemicellulose; or add weak, organic bases
such as dicyandiamide, urea, or melamine (a cyclic tri-
mer of urea) so as to neutralize acids produced by oxi-
dation of the oil and paper. Beginning in the late 1960s,
transformers in the U.S. were manufactured using ther-
mally-upgraded paper, whether specified or not.
Cellulose degrades at any temperature. The rate of deg-
radation is very slow at room temperature. At elevated
temperatures, however, the rate of degradation increases
exponentially, effectively doubling for approximately
every 8C increase in temperature. Management of insu-
lation degradation is necessary to maintain a reasonable
life expectancy for the transformer.
The shortening of the cellulose chains results in a reduc-
tion of the mechanical properties of paper and board,
such as tensile, bursting, and folding strength. It is well
known that the aging rate increases with increasing tem-
perature; Montsinger states that the aging rate doubles
(or life is halved) by every 6-8
o
C temperature increase
(Montsinger 1930).
The insulation aging rate has been studied extensively,
and several mathematical models have been proposed.
The most accepted of these models for insulation aging
is that given in IEEE C57.91-1995 (IEEE 1995b), which
Figure 4.3-3 Graphic depiction of cellulose chain (light
gray is carbon, dark gray is oxygen, white is hydrogen).
4-13
expresses insulation aging as an Arrhenius reaction rate
equation. The equation is as shown in Equation 4.3-4.
(IEEE 1995b, Section 5.2, Equation 2) 4.3-4
Where:
F
AA
is the insulation aging rate.
A is a constant equal to 15,000 for most insula-
tion types.
HS,R
is the reference hot spot temperature for the
insulation.
HS
is the hot spot temperature at which aging is
evaluated.
The IEEE insulation model requires two parameters,
the constant A and the reference hot spot temperature.
For all cases, it is safe to assume that A is 15,000. The
reference hot spot temperature varies depending upon
whether the insulation is thermally upgraded and what
standard is applied. For 55C insulation, the reference
hot spot temperature is 95C. For 65C insulation, the
reference hot spot temperature is 110C.
Equation 4.3-4 can also be expressed in a form for per
unit life:
4.3-5
Where:
B describes how sensitive the ageing is to tem-
perature (15,000 in C57.91).
T
Hot
is the hotspot temperature of the trans-
former.
A
t
is a factor describing the influence of the
environment on the change in tensile strength
(9.8 x 10
-18
in C57.91).
The base of all this is a definition of a material test
where the criterion is that paper aged under a certain
condition (e.g., T
Hot
= 110
o
C) should not lose more
than a certain percentage of its functional property (ten-
sile strength) within the period. This is referred to as the
expected life of a transformer (e.g., 30 years) at nominal
load under specified ambient conditions where the
hotspot condition has to be mastered.
The dependence of aging on temperature is fairly well
known; However, the big problem for a utility when
managing an aging transformer population is that
Equations 4.3-4 and 4.3-5 are based on experiments
under optimum clean conditions for the paper, and that
quantitative information on the importance of the
transformer condition on aging is hard to find. It is all
about getting information about the A
t
factor.
Chemical Description of Aging
For the purposes of insight, aging considerations can be
based not on changes of a functional property like ten-
sile strength, but on changes of the chemical properties
of the paperthe degree of polymerization. A disadvan-
tage of doing this is that the link is lost to the directly
relevant mechanical properties, though only applicable
on laboratory aged samples. The main advantages are
that the approach uses a property that can be measured
from samples from service-aged transformers and that it
is closer to considerations on the chemical kinetic pro-
cesses taking place. If the aging analyses are based on
DP changes, a correlation between the functional criti-
cal tensile strength and the DP, as shown in Figure
4.3-4, is needed.
One usually considers aging under normal conditions
(nameplate conditions and specified overload) to be due
either to hydrolysis or to oxidation, and that a third pro-
cess (pyrolysis) may take place under failure conditions
(thermal defects). The aging due to one kinetic mecha-
nism follows Equation 4.3-6
4.3-6
where the DP is given before and after an aging period
(t) at a temperature T. E is the so-called activation
energy, which describes the sensitivity to the tempera-
ture, A depends on the chemical environment, and R is
the molar gas constant. Note that this is of the same
form as the Arrhenius reaction rate equation first pro-
posed by Dakin in 1948 (Dakin 1948). Usually the
paper has a DP around 1000 in a transformer after
+
=
273 273
, HS R HS
A A
AA
e F

273
=
Hot
T
B
t
e A Life Unit Per
Figure 4.3-4 Correlation between tensile index and DP
value for nonthermally upgraded kraft paper.
( 273)
1 1
E
R T
old new
Ae t
DP DP
+
=
4-14
vapor phase drying, and often a DP of 200 is used as an
end-of-life criterion.
We can rearrange and change the equation to:
4.3-7
and further to
4.3-8
where is the chain scissions and k the aging rate. We
can see that the aging rate depends on temperature (T),
the activation energy (E), and the environment (A). For
a first-order single process, this dependence can be
described by an Arrhenius plot, plotting the natural log-
arithm of the rate versus the inverse absolute tempera-
ture (T + 273) as shown in Figure 4.3-5.
The principle here is that, if the points end along a
straight line, this aging can be described by one single
aging process.
Thus far, focus has been on the temperature depen-
dence. Recent investigations indicate that the environ-
ment may be of even greater importance. Let us first
consider non-upgraded kraft paper.
Kraft Paper
One major mechanism of paper aging is oxidation. The
oxidizing agent in this environment is oxygen from air
ingress. The ultimate end products of oxidation are the
same as for combustioni.e., water and carbon dioxide.
The mechanism of low-temperature oxidation is quite
different from that of combustion, though. The oxygen
concentration is, of course, an important parameter to
determine the rate of oxidation. However, most experi-
mental studies show that the aging rate is not so
strongly influenced by oxygen content. Typically, the
overall degradation rate no more than doubles in exper-
iments with oxygen present, compared to when oxygen
is totally excluded. We can, therefore, say that the
importance of oxygen is limited.
The other major mechanism of paper aging is hydroly-
sis. The significance of water content is paramount: a
humidity of 3-4% may increase the degradation rate of
paper by a factor of 10 or more, compared to dry paper.
In addition, it is commonly acknowledged that a high
acidity in the oil accelerates aging but in a nonspecific
way. Lately, models have been developed outside the
power engineering community (Lundgaard et al. 2004).
It is now understood that the hydrolysis is specific-acid
catalyzed and unimolecular. This means that it is cata-
lyzed exclusively by hydrogen ions from dissociated
acids; undissociated carboxylic acids do not depolymer-
ize cellulose. It is the hydrogen ion activity (or pH)
that matters, not the total (undissociated) acid concen-
tration. It also means that water does not participate in
the rate-controlling step. Water does, however, affect the
acidity by causing carboxylic acids to dissociate, and in
this way it exerts a profound influence upon the aging
process. During this process, carboxylic acids (like for-
mic, acetic, and laevulinic acids) are produced, which
makes the process auto-acceleratory. These acids mainly
reside in the cellulose and may to a large degree be
extracted from oil and paper using water extraction.
This applies to kraft paper. The processes for thermally-
upgraded papers are more complicated.
It is likely that the activation energy of oxidation is dif-
ferent and lower than for hydrolysis, and that for pyroly-
sis, it is even higher than for hydrolysis. This can be
illustrated by the plot in Figure 4.3-6, where lines with
different slopes are plotted for the different mechanisms.
( 273)
1
1
E
R T new
new Old
DP
Ae t
DP DP
+

=

1
new
k t
DP
=
Figure 4.3-5 Arrhenius plot and significance of
parameters for kraft paper at 0.5% (dry) and 3% (wet)
moisture.
1/ T
l
n

(
r
e
a
c
t
i
o
n

r
a
t
e
)
O
2
H
2
O
1/ T
l
n

(
r
e
a
c
t
i
o
n

r
a
t
e
)
O
2
H
2
O
Figure 4.3-6 Total aging as a sum of several
mechanisms.
4-15
Thermally Upgraded Paper
Upgraded paper is paper that is chemically changed by
different chemical additives that reduce the aging rate
compared to untreated kraft paper. A paper is consid-
ered thermally upgraded if it meets the life criteria as
defined in ANSI/IEEE C57.10050% retention in ten-
sile strength after 65,000 h in a sealed tube at 110C.
However, upgrading can be done in various ways, so it is
not a uniform group and may react differently to chemi-
cal aging accelerators. To retard aging, the paper can be
thermally upgraded by linking bulky substituents such
as cyanoethyl ether groups to the HO-groups in the cel-
lulose and hemicellulose, or weak, organic bases such as
dicyandiamide, urea or melamine (a cyclic trimer of
urea) can be added so as to neutralize acids produced by
oxidation of the oil and paper.
In one study it was found that the aging rate for one
upgraded paper type (Insuldur from Avery Dennison)
showed both lower reaction rates at a specific tempera-
ture, and that the acceleration due to presence of water
was lower, as shown in Figure 4.3-7 (IEEE 1995c). In
another study on an upgraded paper (Rotherm) from
Tullis Russel, a similarly reduced sensitivity to water was
observed (Schroff 1985).
The different sensitivity to water indicates that the acid
catalyzed hydrolysis is suppressed. In (Lundgaard et al.
2004), the experimental results indicated that aging for
oxygen-rich oil could increase by a factor of about seven
in the temperature range below 90
o
C, while water (3%)
increased it around 3 times as shown in Figure 4.3-7.
Plastic Deformation and Winding Slackness
The transformer winding is clamped with a pressure
strong enough to give friction forces between windings
that may withstand the forces from a short-circuit cur-
rent. The cellulose can undergo plastic deformation;
water will in this context act as a plasticizer. Changes in
pressure may result from the following reactions:
Increased moisture will cause swelling and increase
the pressure.
Drying may cause shrinking of the insulation and
reduced pressure.
Thermal cycling will cause expansion and retraction
of the materials, cyclic pressure variation, and loss of
pressure.
Diagnostics of the Cellulose
Accurately diagnosing the condition of the cellulose,
and therefore the remaining useful life of the trans-
former, is a most difficult, if not impossible, task. The
only way of accurately determining the condition of the
cellulose insulation system is to take samples from vari-
ous places throughout the transformer and measure the
DP. This is impractical for several reasons. First and
foremost, this involves disturbing the solid insulation in
areas of high electrical stress. Additionally, many areas
are simply inaccessible.
A limited sampling can be done by taking paper from
relatively accessible areas. This, however, involves drain-
ing all of the oil, physically breaching and entering the
tank, collecting the sample while avoiding contamina-
tion, and refilling the unit under vacuum. Breaching the
tank introduces the possibility of contamination by par-
ticulates or moisture and carries a finite risk.
Without entering the tank, the only remaining alterna-
tive for measurement is to draw an oil sample and mea-
sure the level of a chemical marker indicative of
cellulose aging. Fortunately, several furanic compounds,
measurable by High-Performance Liquid Chromatogra-
phy, are produced during the aging process. By measur-
ing the levels of these compounds in the oil, the average
condition of the entire insulation system can be esti-
mated. This method, however, carries with it several
drawbacks. First, the measurement represents only the
average of all of the cellulose in the unit. In reality, only
a small percentage of the cellulose is subject to high
temperatures during overload. It is this cellulose, and in
particular the cellulose that is exposed to high tempera-
tures and high electrical stress, that is important. In
addition, furan analysis is incapable of discerning local-
ized defects that result in high heating in a small area.
Furan production is also a function of the paper-to-oil
ratio of a transformer, which varies widely depending
upon winding construction, BIL, and rating. Finally, if
the oil is replaced or reclaimed, the aging information
derived from furan analysis up to that point is lost.
A final method of estimating the condition of the cellu-
lose is to simply examine the typical loading and ambient
Figure 4.3-7 Arrhenius plots of aging rate for standard
Kraft paper and thermally-upgraded insuldur paper
(IEEE 1995c).
4-16
air temperatures over the transformers service life and
approximate the cumulative insulation aging by the stan-
dard thermal models and insulation aging models. While
not precise, this method is remarkably simple and gives a
general feel for the serviceability of a particular unit.
Concluding Remarks on Thermal Aging of the Insulation
Historically, too little attention has been given to the
changes in thermal performance of a transformer due to
changes in the condition of the insulation system.
Depending on the materials used, the acceleration of
aging due to chemical contamination may exceed that
from increased temperature. Together with a poor con-
dition of the transformer, a long-term temperature
increase may result in a very short life. The life may be
estimated using the formulae below.
The life expectancy of the solid insulation is difficult to
define precisely. The end-of-life of any device or system
is when the system can no longer perform its intended
function. In the case of electrical insulation, this is when
the insulation no longer provides sufficient dielectric
strength to prevent dielectric breakdown. In the case of
cellulose insulation, such as that used in power trans-
former, the electrical properties of the paper do not sig-
nificantly degrade until well after the mechanical
properties become unworkable. The paper will became
extremely brittle and crack or tear at the slightest
mechanical disturbance. If this were to occur in an oper-
ating transformer, the loss of solid insulation by displace-
ment coul d detri mental ly weaken the i nsul ati on
structure. Therefore, the mechanical strength of the paper
is the defining evaluation criterion with respect to aging.
Several different mechanical strength properties have
been used by various investigators to quantity the deg-
radation of cellulose. The burst strength, fold strength,
elongation to break, abrasion resistance, and tensile
strength have all been used as indicators of mechanical
strength. In addition, degree of polymerization (a mea-
sure of average polymer chain length of the cellulose
molecule) has been used. Degree of polymerization, DP,
varies nearly proportionally to reduction and mechani-
cal strength. Of these measures, tensile strength and DP
are used most commonly to define end-of-life criteria.
Table 4.3-1 lists several different end points for insula-
tion thermal life based upon 50% tensile strength reten-
tion, 20% tensile strength retention, and a degree of
polymerization value of 200 (McNutt 1992). Also illus-
trated in this table are the detrimental effects of mois-
ture and oxygen.
Since there are many different valid end-of-life criteria,
it is ultimately up to the user to select a criterion. The
recommendation of this guide is that the DP of 200 end
point be used when evaluating life consumption. As an
alternative, the user could forgo the definition of an
insulation life end point and instead consider aging in
relative termsi.e. per unit aging rate (F
AA
) or equiva-
lent aging hours. However, the question still remains of
when to remove a unit from service.
The failure risk of transformers has often been repre-
sented by the bathtub curve, so named for its shape
(Figure 4.3-8). This curve represents the failure risk over
time for a particular transformer, with all else constant.
External factors, such as loading, throughfault fre-
Table 4.3-1 Insulation Thermal Life Expectancies for Various Endpoint Criteria, Moisture Contents, and Oxygen Levels
(McNutt 1992)
Basis Moisture in Paper (%) Oxygen Level Life (hrs)
50% Tensile
0.5% Low 65,020
1.0% Low 32,510
2.0% Low 16,255
0.5% High 26,000
1.0% High 13,000
2.0% High 6,500
20% Tensile
0.5% Low 152,000
1.0% Low 76,000
2.0% Low 38,000
0.5% High 60,800
1.0% High 30,400
2.0% High 15,200
200 D.P.
0.5% Low 158,000
1.0% Low 79,000
2.0% Low 39,500
0.5% High 63,200
1.0% High 31,600
2.0% High 15,800
4-17
quency, and maintenance (or lack thereof) all shift por-
tions of the curve in some fashion.
If the failure rate with time is represented by a bathtub
curve, the time axis can be viewed as a cumulative ther-
mal aging rather than actual age of the unit. As the
insulation system ages thermally, the system approaches
the wear out portion of the curve, where failure risk
increases, although failure is not necessarily imminent.
The transition point from a relatively flat failure risk to
an increasing failure risk is the thermal end-of-life, in
rough terms. Of course, this is only approximate and
varies with several factors, including condition, design,
vintage, etc.
One major point of confusion and misunderstanding
when performing loading calculations and making load-
ing decisions is the discussion of percent loss-of-life (in
percent total life expectancy) and of the end-of-life or
life remaining.
In rough terms, the end of life for an insulation sys-
tem is the point at which the insulation no longer per-
forms reasonably. For electrical insulation, this means
the point at which the insulation system no longer main-
tains a majority of its original dielectric strength. As
insulation ages, the dielectric strength of the paper does
not decrease significantly until well after the paper has
become brittle. Given the need for transformer insula-
tion to withstand mechanical forces, especially during
throughfault, the point at which the paper loses enough
strength to withstand the mechanical forces is the prac-
tical end of life for the insulation system.
Given this, a precise definition becomes difficult because
this is dependent upon the application of the material,
both electrically and mechanically. Therefore, general
measures of mechanical strength have been historically
used. Some criteria include 50% tensile strength reten-
tion, 20% tensile strength retention, and more recently,
a degree of polymerization of 200. In addition, func-
tional life testing has been performed on distribution
transformers subjected to impulse voltages and short-
circuits. Each of these criteria gives a different life end
point, ranging from 65,000 hrs to 180,000 hrs. These
values are largely artificial. A transformer that is not
subject to high throughfault duty may operate reliably
well after the 180,000-hr end point.
Therefore, if %loss of life (LOL) is discussed, the value
for a given temperature can vary by a factor of three.
It is well known that insulation aging rates increase with
the presence of increased moisture and oxygen. The
original equations in the IEEE Loading Guide (IEEE
1995b) are based mostly upon sealed tube aging tests
with moisture contents less than 0.5% and presumably
lower oxygen contents. This does not reflect the reality
of operating transformers and misleads the user on the
actual aging rate. The importance of keeping transform-
ers dry and oxygen free are not reflected in the guide,
and are best addressed by including quantitative esti-
mates of the effects of moisture and oxygen in the aging
equation. It is proposed that this be done by applying
multiplying factors to the Age Acceleration Factor
calculated as shown in Equation 4.3-9.
4.3-9
Where:
F
AA
is the insulation aging rate.
A is a constant equal to 15,000 for most insula-
tion types.
HS,R
is the reference hot spot temperature for the
insulation.
HS
is the hot spot temperature at which aging is
evaluated.
k
H2O
and k
O2

are determined from Tables 4.3-2 and 4.3-3,
respectively
Figure 4.3-8 Example bathtub curve demonstrating
thermal end-of-life.
e
k k =
F

273 +
B
-
273 +
B

O O H AA HS o
2 2
Table 4.3-2 Aging Rate Correction Factor for Moisture in
Paper
Moisture Content in Paper (roughly) k
H2O
Dry (< 0.5%) 1
Moist (0.5-2.5%) 2
Wet (> 2.5%) 4+
Table 4.3-3 Aging Rate Correction Factor for Oxygen
Content of Oil
Oxygen Content k
O2
Low 1
High 3-5
4-18
The factors developed in the tables were derived from
data published by Emsley, Lundgaard and McNutt
(Emsley and Stevens 1994; Lundgaard et al. 2004; and
McNutt 1992). They are approximations. However,
insulation aging calculations, as a whole, are gross
approximations.
By applying these factors directly in the equation, users
are well aware of the importance of moisture and oxy-
gen, and are given a method for evaluating the impact of
moisture and oxygen, as well as estimated life consump-
tion for transformers with moisture contents in paper
greater than 0.5% and with increased oxygen levels.
4.3-10
Where:
t is the number of hours at the hot spot tempera-
ture.
Expected life is a value selected from Table 4.3-1.
4.3.3 Additional Risks
Ancillary Components
Following insulation loss of life and bubble formation,
the most significant risk associated with transformer
overload is damage to the ancillary components, such as
bushings, on-load tap changers (LTCs), off-load tap
changers (DETCs), and current transformers (CTs).
This risk is often overlooked as a rating consideration.
Generally, these components are oversized due to their
relative cost when compared to the transformer itself.
However, it is possible that some limitation on the over-
load capability of a particular transformer may be due
to an ancillary component.
Bushings
Oil-impregnated, paper-insulated, capacitance-graded
bushings are designed with 105C bushing hot spot limit
at rated load and 95C top oil temperature average over
24 hrs. Overload, based upon the rated bushing current,
is possible. Overload risks of bushings include:
Pressure build-up due to oil expansion.
Deterioration of gaskets and seals.
Thermal deterioration of paper insulation.
Increase in dielectric loss, possibly resulting in ther-
mal runaway.
Gas evolution at extreme hot spots.
Overload limits, as specified in C57.19.100-1995:
40C ambient temperature
110C transformer top oil temperature
2x rated bushing current
150C bushing hot spot temperature
The bushing hot spot is usually located adjacent to the
central conductor in the area of the mounting flange. It
is here that the insulation is the thickest and the external
heating, via the bulk oil, is highest. The hot spot tem-
perature of the bushing conductor may be calculated as
shown in Equation 4.3-11 (IEEE 1995c).
4.3-11
Where:
HS
is the bushing hot spot temperature.
K
1
is a bushing-specific constant, typically rang-
ing from 15 to 32.
K
2
is a bushing-specific constant, typically rang-
ing from 0.6 to 0.8.
n is a bushing-specific constants, typically rang-
ing from 1.6 to 2.0 with 1.8 typical.
I is the bushing current in per unit of the rated
bushing current.
O
is the bushing immersion oil temperature
(transformer top oil).
The transient temperature rise may then be calculated as
shown in Equation 4.3-12.
4.3-12
Where:
HS,2
is the bushing hot spot rise at time t
1
+ t.
HS,1
is the bushing hot spot rise at time t
1
.
HS,U
is the steady-state temperature given in the
previous equation.
B
is the bushing time constant.
The determination of the constants is the greatest obsta-
cle to the use of these equations. Accurate determina-
tion can only be made through tests on duplicate
bushings. These tests are not performed routinely by the
bushing manufacturer. This i nformation may be
requested from the manufacturer. If the data is not
available, however, typical default values can be used
without a significant sacrifice in accuracy.
The paper used in paper-insulated bushings is not ther-
mally-upgraded, unless otherwise specified. As such, the
standard loss-of-life model given in C57.91-1995 for
65C rise insulation systems should not be used to cal-
Life Expected
t F
LOL
AA

= %
O
n
HS
K I K + =
2 1
+ =

B
t
HS U HS HS HS
e

1 ) (
1 , , 1 , 2 ,
4-19
culate the loss-of-life for bushing insulation. Instead,
Equation 4.3-13 should be used:
4.3-13
Where:
F
AA
is the aging acceleration factor.
Percent loss-of-life and cumulative aging can be calcu-
lated as outlined in C57.91-1995.
The above method may be applied to resin-bonded or
resin-impregnated bushings. The above method does
not apply to draw-lead connected bushings. Draw-lead
connected bushings are limited by the temperature rise
of the draw-lead. Therefore, the bushing temperature is
determined by the size and construction of the draw-
lead. Accurate determination of draw-lead current car-
rying capability would require detailed information on
the lead and its application. However, in a properly
designed application, the draw-lead does not limit the
thermal rating of the transformer.
On-Load Tap Changers (LTCs)
The critical component of an on-load tap changer, from
a thermal standpoint, is the contacts. LTCs are properly
designed with a contact rise over oil of less than 20C at
1.2 times rated load (rated load here is the rated load of
the LTC, not the transformer itself). In addition, LTCs
are designed to break twice rated current at least 40
times without detrimental degradation of the contacts.
In some instances of high overloads, it may also be
advisable to switch an LTC to manual control and mini-
mize or eliminate operation. Overload risks of LTCs
include:
Increased contact wear and ablation with increased
load during break operations
Increased contact temperature increases probability
and rate of coking of contacts (>120C)
Higher overloads result in prolonged arcing during
break operation. Dragging the arc across the contacts
could result in short-circuiting the regulating winding.
Overload limits, as specified in C57.131-1995 (IEEE
1995a):
120C contact temperature (higher is OK, but may
result in greater maintenance).
2x LTC rated load current (limit breaking operations
at high load level to few times/year).
The temperature rise of the contacts is calculated as
shown in Equation 4.3-14 (IEEE 1995c)
4.3-14
Where:
C
is the LTC contact temperature rise over oil.
C,R
is the LTC contact temperature rise over oil at
rated load.
I is the LTC current in per unit of the rated LTC
current.
n is an exponent that varies from 1.6 to 1.85,
with 1.8 as the default.
4.3-15
Where:
C
is the total contact temperature.
A
is the ambient air temperature.
TO
is the transformer top oil temperature.
K is a constant to account for the difference
between transformer top oil temperature and
LTC compartment oil temperature. This is typi-
cally around 0.8.
It is important to note that the main difficulty associ-
ated with elevated temperatures in LTCs is the increased
tendency toward contact coking. Since coking increases
the electrical resistance across the contact assembly,
coking can result in even higher temperatures and possi-
ble thermal runaway. If frequent high loading periods
are expected, the temperature differential between the
main tank oil and the LTC compartment oil should be
checked, if possible. In addition, the LTC oil should be
clean oil with high interfacial tension (IFT) and low
acidity.
Reduced Capacity Taps
The load dependent losses, and therefore the heating of
the transformer windings, is a function of the winding
current. Often the transformer load is expressed in
terms of apparent power, or MVA, since this number is
essentially the same for all windings (neglecting losses).
However, in the case of rated MVA when taps are
present, the rated current varies inversely proportional
to the voltage. The consequence for loading is that, for
lower voltage taps, the rating in MVA for a given size
conductor decreases.
Often, to maintain the same MVA load capability for all
taps, the conductors are sized according to the current
at the lowest voltage time. However, occasionally, to
reduce cost, the transformer may be specified with
reduced capacity taps. In this case, the conductors are
sized for the rated current at the neutral position. For
higher-voltage taps, the transformer is capable of carry-
ing full rated MVA. However, for lower voltage taps, the
273
000 , 15
368
000 , 15
+
=
HS
e F
AA
n
R C C
I
,
=
C TO A C
K + + =
4-20
transformer is capable of carrying only a reduced MVA,
corresponding to the rated current at the neutral tap.
Transformers with reduced capacity taps can easily be
identified by referring to the table of tap positions, volt-
ages, and rated currents on the nameplate. Transformers
with reduced capacity taps have the same rated current
for all taps at voltages below neutral. An example of this
is shown in Table 4.3-4.
It is important to examine the nameplate when deter-
mining the load capability of a particular transformer to
determine if the unit was designed with reduced capacity
taps. If it was designed with reduced capacity taps, and
the unit is operated at taps below neutral, the load capa-
bility in MVA will be lower than the nameplate MVA.
Table 4.3-4 Nameplate Showing Tap Positions, Voltages, and Currents
CONNECTIONS
WINDING VOLTAGE
MAXIMUM FOA
AMPERES
CRT LOAD TAP CHANGER
PCS
CONNECTS IN EACH PHASE
P1 TO P3 TO R TO
H
VOLTAGE WYE
120175 432 16L 11 11 3
119495 435 15L 11 10 3
118610 437 14L 10 10 3
118130 440 13L 10 9 3
117445 442 12L 9 9 3
116765 445 11L 9 8 3
116080 448 10L 8 8 3
115395 450 9L 8 7 3
114715 453 8L 7 7 3
114030 456 7L 7 6 3
113350 458 6L 6 6 3
112665 461 5L 6 5 3
111980 464 4L 5 5 3
111300 467 3L 5 4 3
110615 470 2L 4 4 3
109935 473 1L 4 3 3
109250 476 NEUT 11 11 11
108565 476 1R 11 10 11
107885 476 2R 10 10 11
107200 476 3R 10 9 11
106520 476 4R 9 9 11
105835 476 5R 9 8 11
105150 476 6R 8 8 11
104470 476 7R 8 7 11
103785 476 8R 7 7 11
103105 476 9R 7 6 11
102420 476 10R 6 6 11
101735 476 11R 6 5 11
101055 476 12R 5 5 11
100370 476 13R 5 4 11
99690 476 14R 4 4 11
99005 476 15R 4 3 11
98325 476 16R 3 3 11
WINDING VOLTAGE AMPERES
NO LOAD TAP CHANGER
POS
CONNECTS IN EACH PHASE
A B
X
VOLTAGE WYE
59000 753 2 1 TO 19 13 TO 15
06000 787 1 1 TO 18 13 TO 14
4-21
Bushing or Internal CTs
Detailed evaluation of bushing CTs is difficult, if not
impossible. The location of the CT can be in the bushing
adapter or turret, or mounted on the active assembly
in the tank. Bushing CTs are usually limited by trans-
former top oil temperature. Maintaining a top oil tem-
perature of less than 110C should avoid excessive
temperatures. Aging rates should be lower than that of
the winding insulation over the average of the loading
conditions.
De-Energized Tap Changers (DETCs)
De-energized tap changers are similar, in a thermal
sense, to on-load tap changers, albeit much simpler. The
same general considerations and risks apply to DETCs
as to LTCs. Coking may become a greater problem for
DETCs, as they are seldom operated. It may be advis-
able to periodically operate the DETC over the range of
taps to wipe the contacts clean of any deposits.
Lead Heating
Internal connecting leads may be a concern. The same
hot spot limits, and associated risks, apply for leads as
for windings. In general, it is impossible to evaluate lead
heating without detailed design data. In a properly
designed transformer, the connecting leads should not
limit the overload capability. However, lead hot spots
have been a limitation in the load carrying capability of
many transformers because of lack of attention in the
design stage or manufacturing practices that permitted
or even encouraged liberal application of insulating tape
at lead connection points.
Evaluation of lead heating problems outside of the fac-
tory is rather difficult. Even with dimensions of the
leads and construction details, it is difficult to assess the
oil flow conditions in the vicinity of the lead. The best
method for assessing lead heating problems is by exam-
ining the DGA records for the unit for any hint of heat-
ing gases involving cellulose decomposition, particularly
CO and CO
2
, and eventually CH
4
and C
2
H
6
. If these
gases are present in significant quantity or have a gener-
ation rate that increases sharply with load, load should
be reduced immediately. At the earliest convenience, an
internal inspection should be performed to locate the
source of the gassing. In addition to DGA, partial dis-
charge detection and perhaps Furan analysis may also
yield useful information.
Stray Flux Heating
Heating of non-current carrying metal components by
the leakage flux of the windings and leads is termed
stray flux heating. The leakage flux induces eddy cur-
rent in any conducting material that it passes through.
This includes the steel clamping structure, tie rods or tie
plates, metal core bands, and the tank wall itself. Since
leakage flux varies proportionally with load current, the
stray flux heating increases roughly with the square of
winding current. For larger power transformers and
GSUs in particular, the problem of stray flux heating
can be substantial. Special design measures must be
taken to reduce the induced currents.
The most common areas of stray flux heating problems
are tie plates, used to connect the upper and lower core
yoke clamping structure, and bushing penetrations. As
with lead heating, evaluation of stray flux heating
potential is difficult. The best method for assessing stray
flux heating problems is by examining the DGA records
for the unit for any hint of hot metal gases, specifically
methane and ethane. Should either of these gases exceed
120 ppm or 65 ppm, respectively, high loading should be
discontinued and the source of the gases investigated. In
addition, sudden increases in these gas levels concurrent
with increased loading should be considered additional
cause for alarm.
In addition to DGA, stray flux heating in the tank wall
can be evaluated by examining the transformer with an
infrared camera. Particular attention should be paid in
the area of bushing penetrations and bushing adapters.
A visual examination for discolored paint could also
reveal stray flux heating.
4.4 THERMAL MODELING
Given the extreme complexities of heat transfer within a
transformer, as well as the lack of information typically
available, drastic simplifications must be made. It is
important to recognize the simplifications and to know
the areas in which these simplifying assumptions are not
sufficiently accurate. Therefore, this section begins with
a brief and somewhat academic overview of the various
heat transfer mechanisms involved. Following this, the
section describes available thermal models.
4.4.1 Mechanisms of Heat Transfer
In a power transformer, two heat transfer mechanisms
are predominant: radiation and convection. Conduction
plays a smaller role and will not be described at length
here. It should also be noted that even the most modern
treatments of the subject are largely empirical. This ini-
tial discussion parallels that of Montsinger in (Mon-
tsinger 1951), with additional emphasis on variables
that may not be constant.
Radiation
All bodies at a temperature above their surroundings
radiate heat energy to the surroundings. The mathemati-
cal relationship between heat lost due to radiation and
4-22
the temperature rise are fairly well known, as shown in
Equation 4.4-1 (Cengel 1997)
4.4-1
This relationship can be simplified to a simple power
relationship of the form shown in Equation 4.4-2.
4.4-2
This relationship, for a given exponent x, is generally
valid over a limited temperature range (Figure 4.4-1).
The curve fits in Figure 4.4-1 are generally valid from a
0C rise to about a 70C rise. The exponents vary from
1.133 for a 0C ambient to 1.117 for a 40C ambient.
Note that the emissivity will change the multiplicative
constant but not the exponent.
This relationship could also be expressed inversely as:
4.4-3
where m = 1/x. For the ranges of x derived above, m var-
ies from 0.883 for a 0C ambient to 0.895 for a 40C
ambient.
The absorption rate in the infrared region for transpar-
ent liquids such as transformer oils is relatively high.
Therefore, the radiant energy from transformers wind-
ings is absorbed by even a thin layer of oil, and is there-
fore negligible. Air, on the other hand, has a negligible
absorption rate, and radiant energy loss must be consid-
ered when determining the heat loss from the tank, and
therefore the oil.
Forced Convection
When heat is transferred from a solid to a moving fluid
medium, the adjacent fluid is heated by conduction.
Since the fluid medium is moving, the heated fluid is
quickly replaced by cooler fluid. This results in an
increased rate of heat transfer than would occur by con-
duction alone. The heat loss due to convection can be
expressed by Newtons Law of Cooling, as shown in
Equation 4.4-4 (Cengel 1997).
4.4-4
Where:
Q is the rate of heat loss.
h is the heat transfer coefficient.
T
c
is the temperature at the surface of the object
(solid).
T
a
is the temperature of the surrounding fluid.
For cross-flow over a cylinder or similar-shaped object
(Cengel 1997):
4.4-5
Where:
C is a constant.
Re is the Reynolds Number.
Pr is the Prandtl Number.
k is the thermal conductivity of the fluid.
D is the characteristic length of the immersed
object (diameter of cylinder).
is the density of the fluid.
V is the velocity of the fluid.
is the viscosity of the fluid (varies with tem-
perature).
C
p
is the specific heat of the fluid.
m, n are constants.
From this, it can be seen that h is a function of several
parameters, including geometry, fluid velocity, and fluid
viscosity. The exponent n is typically 1/3. The exponent
m, determined empirically, ranges from about 0.3 to 0.8,
depending upon the geometry. Note that if the expo-
nents are equal, the viscosity cancels out and is no
longer a factor. Summarizing, the heat loss can be
expressed by:
4.4-6
or, solving for the temperature rise:
4.4-7
Natural Convection
Natural convection is slightly more complex than forced
convection. The heat transfer is still a function of the
) (
4 4
a b rad
T T A Q =
x
rad
K Q =
Figure 4.4-1 Radiative heat transfer at various ambient
temperatures.
m
rad
CQ =
) (
a c
T T hA Q =
n
p
m
n m
k
C
VD
C
D
k
C h
= =

Pr Re
= K Q
KQ =
4-23
velocity of the fluid, as described above. However, now
the velocity of the fluid is the result a varying fluid den-
sity with temperature. This is known as thermosiphon
flow. The basic relationship is the same as for forced
convection.
4.4-8
However, h is now a function of the temperature rise as
well. For natural convection along a vertical plate (also
applicable to a vertical cylinder), h is given by:
4.4-9
m varies from 1/3 to 1/4. Note that h is also a function of
viscosity, , to the -m power.
Combining the temperature component of h with the
equation for convection heat transfer:
4.4-10
Taking the inverse to solve for the temperature rise:
4.4-11
Since m varies from 1/3 to 1/4, the exponent in Equation
4.4-11 varies from 0.75 to 0.80.
Combined Heat Transfer
Summarizing the above discussion, the temperature rise
(or, conversely, the heat loss) for a particular mechanism
of heat transfer can be expressed in the following form:
4.4-12
or
4.4-13
Where:
is the temperature rise above the surrounding
medium.
K is a constant.
Q is the heat loss (or gain).
x is an exponent that varies with mechanism as
The constant K is generally constant over the range of
temperatures and materials considered in transformer
thermal modeling. The only variable that changes
appreciably with temperature is the viscosity of oil. For
forced convection, the viscosity is generally not a factor.
For natural convection, the temperature rise is a func-
tion of viscosity as follows:
4.4-14
Transient Formulation
For any temperature rise, the temperature at any time, t,
can be determined by equating the heat input to the sum
of the change in heat storage and the heat loss:
Heat generated = Change in heat storage + Heat loss.
4.4-15
Traditionally, transient temperatures have been calcu-
lated by approximating an instant of time as a step-
change in load and calculating the temperature at the
end of the time step as follows:
4.4-16
Where:
is the calculated temperature at time t + t.
i
is the initial temperature at time t.
u
is the ultimate steady-state temperature.
t is the time step.
e is the time constant.
This closed form solution to the transient differential
equation has several drawbacks. First, it is difficult to
correct for the change in losses with temperature. The
losses at the ultimate temperature rise are a function of
the ultimate temperature. The ultimate temperature is,
of course, a function of these losses, requiring an itera-
tive calculation at each time step to find the ultimate
temperature rise. In addition, the closed form solution
above is only truly correct for heat transfer function
with an exponent of 1.0 (i.e., forced convection).
A more robust approach is to solve the differential
equation using a finite difference approach, as taken by
Pierce (Pierce 1994) and Lesieutre (Lesieutre et al. 1997)
) (
a c
T T hA Q =
( )
( )
m
p
a c m
k
C
L T T g
L
k
C Gr
L
k
C h
= =

2
3 2
Pr
1 +
=
m
K Q
) 1 ( 1 +
=
m
nat
KQ
x
KQ =
1
x
Q K =
Table 4.4-1 Exponents for Radiation, and Forced and
Natural Convection
Mechanism Exponent
Radiation 0.883-0.895
Forced Convection 1.0
Natural Convection 0.75-0.8
m m
nat
KQ
+
=
) 1 ( 1
x
p gen
K
dt
d
mC Q
+ =
( )

+ =

t
i u i
e 1
4-24
Referring to the differential equation above, the differ-
ence equation would take the form:
4.4-17
4.4.2 Top Oil Model (IEEE C57.91-1995, Clause 7)
The following is an outline of the temperature calcula-
tion methods presented in Clause 7 of IEEE C57.91-
1995 (IEEE 1995b) (Figure 4.4-2).
The thermal mechanisms of a transformer are extremely
complex and difficult to model. To reduce complexity,
the transformer is analyzed as a lumped system, and
several assumptions are made. The transformer is essen-
tially reduced to two systems, the bulk oil and the wind-
ing, with an additional hot spot temperature located
at the top of the winding.
The oil temperature is presumed to be lowest at the bot-
tom of the winding. As the oil rises upward along the
winding, the oil adjacent to the winding is heated at a
constant rate. Therefore, the oil temperature is assumed
to increase linearly from the bottom of the winding to
the top of the winding, with the highest oil temperature
located at the top of the winding.
At rated load, the winding temperature distribution is
assumed to be higher than the oil temperature distribu-
tion by a constant value,
w
, and thus parallel to the oil
temperature distribution. Due to stray flux concentra-
tion near the top of the winding, a further increase in
temperature, the hot spot temperature, is located at the
top of the winding. This hot spot temperature represents
the hottest temperature endured by the insulation and
therefore the highest aging rate (1995b).
The IEEE top oil method relates the steady-state oil
temperature rises to the total losses by a power function.
The steady-state winding rises are also related to the
winding current by a power function. In addition, cer-
tain corrections are made to account for additional con-
siderations such as increased mixing due to forced-oil
cooling systems and change in winding resistance with
temperature.
The following is an outline of the temperature calcula-
tion method presented in IEEE C57.91-1995, Clause 7.
Symbols:
C is the thermal capacity (Watt-hours/deg C).
W
CC
is the weight of the core and coils (kg).
W
Tank
is the weight of the tank and fittings in contact
with the oil (kg).
V
Fluid
is the volume of oil (L).
O
is the oil thermal time constant (min).
w
is the winding time constant (min).
t is the time (min.).
t is the calculation time step (min).
TO
is the top oil rise over ambient temperature
(deg C).
x
p gen
K
t
mC Q

+
=
1 2
2 ,
Figure 4.4-2 Graph of temperature rises vs. height for IEEE Clause 7 Model.
4-25
HS
is the hot spot rise over the top oil temperature
(deg C).
P
T
is the total loss (W).
K is the load (per unit of the nameplate rating).
R is the ratio of load loss to no-load loss
at the nameplate rating.
m is the winding temperature rise exponent.
n is the oil temperature rise exponent.
HS
is the hot spot temperature (deg C).
A
is the ambient temperature (deg C).
Subscripts:
R indicates a rated quantity.
i indicates an initial quantity.
U indicates an ultimate or steady-state quantity.
Oil Thermal Time Constant:
In order to calculate the transient response of the bulk
oil, it is necessary to calculate an oil thermal time con-
stant. The thermal time constant of the bulk oil at rated
temperature equals:
O,R
= C *
TO,R
/ P
T,R
4.4-18
Where:
O,R
is the oil thermal time constant at rated load
(min).
C is the combined thermal capacity of the
transformer (W-min/lbs-C).
TO,R
is the measured oil temperature rise over
ambient at rated load (C).
P
T,R
is the total loss at rated load (W).
The thermal capacity, C, is the combined thermal
capacity of all transformer components in contact with
the bulk oil. This includes the core, windings, tank and
fittings, and the oil itself. The specific heat for each com-
ponent material is shown in Table 4.4-2.
A combined thermal capacity can be calculated by com-
bining the specific heats of the various component mate-
ri al s wi th the component wei ghts. However, the
components are not at a uniform temperature. The oil
temperature varies from the bottom of the tank to the
top. For OA (ONAN) and FA (ONAF), the assumption
has been made that the mean oil temperature is 76% of
the maximum top oil temperature. In addition, 2/3 of
the tank weight is used.
For OA and FA (ONAN and OFAF), the thermal
capacity, C, equals:
C = 0.06 * W
CC
+ 0.04 * W
Tank
+ 1.33 * V
Fluid
4.4-19
For DFOA and NDFOA (ODAF and OFAF), the ther-
mal capacity, C, equals:
C = 0.06 * W
CC
+ 0.06 * W
Tank
+ 1.93 * V
Fluid
4.4-20
At top oil temperatures other than rated, the time con-
stant must be corrected as follows:
4.4-21
Initial Temperatures:
The following equations are used to calculate the initial
temperatures based upon the assumption that the load
prior to the calculation period was constant long
enough for the temperatures to reach their steady-state
limits. This assumption is reasonable if the load is fairly
constant for a period of time prior to the overload equal
to two to three times the oil thermal time constant.
4.4-22
Transient Temperatures:
The following equations are used to calculate the tran-
sient temperature rises based upon the oil time constant
calculated above, the user-determined winding time con-
stant, load, losses, and initial and ultimate temperature
rises. For a graphical representation of the various tem-
perature rises, see Figure 4.4-2.
Table 4.4-2 Specific Heat for Copper, Steel, and Oil
Material Specific Heat (W-min/lbs-C)
Copper 0.05
Steel 0.06
Oil 14.6
n
R TO
i TO
n
R TO
U TO
R TO
i TO
R TO
U TO
R O O
1
,
,
1
,
,
,
,
,
,
,

( )
( )
m
i R HS i HS
n
i
R TO i TO
K
R
R K
2
, ,
2
, ,
1
1
=

+
+
=

4-26
For all cooling modes:
The top oil rise at time t
2
= t
1
+ t is given by the follow-
ing:
4.4-23
4.4-24
The hot-spot rise over top oil at time t
2
= t
1
+ t is given
by the following:
4.4-25
4.4-26
The hot-spot temperature at time t
2
= t
1
+ t is given by:
4.4-27
The above calculation is repeated for each time step of
the overload period, with the initial temperatures rises
equal to the temperature rises from the previous time
step. In addition, loss-of-life calculations are performed
at each time step (IEEE 1995b).
4.4.3 Bottom Oil Model (IEEE C57.91-1995,
Annex G)
This model was originally developed by Linden Pierce
with GE in the early 1990s based upon detailed temper-
ature measurements on a model coil (Pierce 1992, 1994).
The model was included in the 1995 revision of IEEE
C57.91 as Annex G (IEEE 1995b).
In the top oil model, the hot spot temperature consisted
of three components: ambient temperature, top oil rise
over ambient, and hot spot rise over top oil (Figure
4.4-3). In the bottom oil model, which is the approach
that will be developed here, the hot spot temperature
consists of four components: ambient temperature, bot-
tom oil rise over ambient, top-of-duct oil to bottom oil
gradient, and hot spot rise over top of duct oil. These
components are shown schematically below. In addition
to these components, the average winding rise over aver-
age duct oil needs to be developed, as the winding losses
are a function of this temperature and not the hot spot
temperature.
4.4-28
Average Winding Rise over Average Duct Oil
It will be assumed that the temperatures of the winding
and the duct oil vary linearly from bottom to top, and
both are parallel. This results in a winding temperature
rise over adjacent duct oil that does not vary over the
height of the winding, and a constant heat flux from the
winding to the duct oil. In line with the finite difference
transient formulation described above, a heat balance
will be developed for the winding-duct oil system.
The heat generated by the winding at time t
2
= t
1
+ t is
given by:
4.4-29
Where:
Q
gen,W
is the average heat generated by the
windings.
L is the winding current, in per unit rated.
P
I2R
is the Ohmic losses in the windings due to
the winding current at rated current.
P
E
is the eddy losses in the windings at rated
current.
4.4-30
Where:
W,1
is the average winding temperature calculated
at the previous time step, t
1.
W,R
is the average winding temperature at rated
load.
K
is 234.5C for copper windings and 225C for
aluminum windings.
The heat lost by the windings to the duct oil is depen-
dent upon the mechanism of heat transfer. For ONAN,
ONAF, and OFAF, the predominant mode of heat
( )
( )
n
U
R TO U TO
R
R K
+
+
=
1
1
2
, ,

( )
1 , 1 , , 2 ,
1
TO
t
TO U TO TO
O
e

+
m
U R HS U HS
K
2
, ,
=
( )
1 , 1 , , 2 ,
1
HS
t
HS U HS HS
W
e

+
2 , 2 , 2 , 2 , HS TO A HS
+ + =
DO HS BO DO BO A HS / /
+ + + =
Figure 4.4-3 Components of bottom oil model.
BO
DO
HS
( )
W E W R I W gen
K P K P L Q + =
2
2
,
+
+
=
K R W
K W
W
K

,
1 ,
4-27
transfer is natural convection. As outlined above, the
heat loss for natural convection can be expressed by:
4.4-31
Where:
Q
loss,W
is the average heat lost by the windings to the
average duct oil.
W,1
is the average winding temperature at the pre-
vious time step.
DO,1
is the average duct oil temperature at the pre-
vious time step.
W,R
current.
DO,R
is the average duct oil temperature at rated
current.
W,R
is the oil viscosity at a temperature equal to
the average of the average winding tempera-
ture and average duct oil temperature at rated
load.
W,1
the average of the average winding tempera-
ture and average duct oil temperature at the
previous time step.
m is a constant, usually between
1
/
3
and
1
/
4
.
For ODAF transformers (directed forced oil), the pre-
dominant mode of heat transfer is forced convection.
The heat loss for OFAF can then be expressed by:
4.4-32
Note that it is assumed that the effect of the change in
oil viscosity with temperature is negligible for this case.
Completing the heat balance, the heat absorbed by the
windings is given by:
4.4-33
Where:
Q
abs,W
is the heat absorbed by the winding during
incremental time step, t.
m
W
is the mass of the windings.
C
P,W
is the effective specific heat of the windings.
W,1
is the average winding temperature at time
t
1.
W,2
t
2
= t
1
+ t.
t is the incremental time step.
The term m
W
C
P,W
is usually not known directly. Instead,
this term can be calculated from a winding thermal time
constant calculated from the winding cooldown curves
taken during the factory heat run. From the definition
of a thermal time constant:
4.4-34
Summing the three terms of the heat balance and solv-
ing for
W,2
:
4.4-35
Duct Oil Gradient over Bottom Oil
The oil entering the winding ducts is at a temperature
equal to the bottom oil temperature. The oil is then
heated as it rises up the winding duct, either by forced
oil flow or natural thermosiphon flow. The temperature
at the duct exit, for a constant heat flux along the duct,
is dependent upon the mass flow rate, as follows:
4.4-36
Where:
T
e
is the temperature of the fluid exiting the duct.
T
i
is the temperature of the fluid entering the
duct.
q
s
is the constant heat flux along the duct.
A
s
is the surface area of the duct.
is the density of the fluid.
V is the fluid velocity (average).
A
c
is the cross-sectional area of the duct.
C
p
is the specific heat of the fluid.
For ODAF, the mass flow rate through the duct remains
essentially constant for varying load levels. For ONAN,
ONAF, and OFAF, the oil flow through the duct is by
natural thermosiphon flow. The flow rate through the
duct in these cases is a complex function of the tempera-
ture rise in the duct, the temperature drop through the
radiators, and the relative height of the windings and the
radiators. The temperature rise in the duct, in turn, is a
function of the mass flow rate through the duct (Figure
4.4-4).
Lacking sufficient data, the following simple approach
will be taken, relating the temperature rise through the
vertical ducts (top-of-duct oil bottom oil) to the loss
by an exponent, y, as follows:
4.4-37
Where:
DO/BO
is the duct oil rise over bottom oil.
( )
E R I
m
W
R W
m
R DO R W
DO W
W loss
P P Q +
=
+
2
1 ,
,
1
, ,
1 , 1 ,
,

( )
E R I
R DO R W
DO W
W loss
P P Q +
=
2
, ,
1 , 1 ,
,

( )
t
C m Q
W W
W P W W abs
=
1 , 2 ,
, ,

( )
( )
R DO R W
E R I W
W P W
P P
C m
, ,
2
,

+
=
( )
1 ,
,
, ,
2 , W
W P W
W loss W gen
W
C m
t Q Q
+

=
P c
s s
i e
C VA
A q
T T
+ =
y
R BO TDO HS BO DO
L H
2
, / 2 , /
* =
4-28
H
HS
is the height of the winding hot spot in per
unit of the winding height (1.0 = top).
TDO/BO,R
is the top-of-duct oil rise over bottom oil
at rated load (see discussion below).
L is the per unit load.
y is an exponent.
In almost all cases, it can be assumed that the hot spot is
at the top of the winding, and hence adjacent to the top
duct oil temperature, making H
HS
equal to 1.0. The data
from tests done on model coils in Pierce (1994) suggests
that, for most cases, the hot spot is within the top 10%
of the winding. When the increased eddy loss at the top
of the winding is factored in, the hot spot should be
even closer to the top, most often within a few turns or
disc sections. However, H
HS
has been included to allow
the engineer to assume a hot spot position below the top
of the winding.
The exponent, y, is as yet undetermined. Lacking mea-
sured data, the values given by Pierce (1994) can be
used. For ONAN, ONAF, and OFAF, y is given as 0.5.
For ODAF, y is given as 1.0. Since the mass flow rate
does not change for ODAF over varying loads and the
heat flux is proportional to the losses, the exponent of
1.0 for ODAF should be correct. For the other cooling
modes, however, the mass flow rate is a function of both
the oil temperatures within the duct and the oil temper-
atures in the bulk oil and radiators. Therefore, a more
complex relationship is needed.
The temperature of the oil exiting the ducts, expressed
as the rise over bottom oil
DO/BO
, at rated load is diffi-
cult to determine. It cannot be measured easily. Even
when measurements are made, the values vary consider-
ably for the various vertical ducts around the circumfer-
ence of the winding (Pierce 1992). Pierce (1994) suggests
that, for ONAN, ONAF, and ODAF, this value is equal
to the rated top oil rise. For OFAF, Pierce suggests using
the average winding rise. These recommendations
appear to be based upon the tests of the model coil in
(Pierce 1992).
Hot Spot Rise over Top Duct Oil
By following a development similar to the average wind-
ing rise over average duct oil, only considering an incre-
mental portion of the winding at the hot spot (top of
winding), the hot spot temperature rise over top duct oil
can be calculated.
The heat generated by the winding at the hot spot at
time t
2
= t
1
+ t is given by:
4.4-38
Where:
Q
gen,HS
is the heat generated by the windings at the
hot spot.
P
I2R
P
E
current.
4.4-39
Where:
HS,1
is the winding hot spot temperature calculated
1.
HS,R
is the winding hot spot temperature at rated
load.
K
aluminum windings.
The heat lost by the windings to the duct oil is depen-
dent upon the mechanism of heat transfer. For ONAN,
ONAF, and OFAF, the predominant mode of heat
transfer is natural convection. As outlined above, the
heat loss for natural convection can be expressed by:
4.4-40
Where:
Q
loss,HS
is the heat lost by the winding hot spot to
the top duct oil.
HS,1
is the winding hot spot temperature at the
previous time step.
DO,1
is the duct oil temperature at the previous
time step.
HS,R
is the winding hot spot temperature at rated
current.
Figure 4.4-4 Oil flow throughout the duct. ( )
HS E HS R I HS gen
K P K P L Q + =
2
2
,
+
+
=
K R HS
K HS
HS
K

,
1 ,
( )
E R I
m
HS
R HS
m
R DO R HS
DO HS
HS loss
P P Q +
=
+
2
1 ,
,
1
, ,
1 , 1 ,
,

4-29
DO,R
is the duct oil temperature at rated current.
HS,R
the average of the winding hot spot tempera-
ture and top.
HS,1
the average of the winding hot spot tempera-
ture and top duct oil temperature at the pre-
vious time step.
m is a constant, usually between 1/3 and 1/4.
For ODAF transformers (directed forced oil), the pre-
dominant mode of heat transfer is forced convection.
The heat loss for OFAF can then be expressed by:
4.4-41
Note that it is assumed that the effect of the change in
oil viscosity with temperature is negligible for this case.
windings is given by:
4.4-42
Where,
Q
abs,HS
is the heat absorbed by the winding during
incremental time step, t.
m
W
is the mass of the windings.
C
P,W
is the effective specific heat of the windings.
HS,1
t
1.
HS,2
is the average winding temperature at time t
2
= t
1
+ t.
ing for
HS,2
:
4.4-43
Bottom and Top Bulk Oil Rise
Pierce (1994) calculates the average oil rise and then
applies a top-to-bottom oil gradient to that average oil
rise. This gradient is calculated as a function of the
losses raised to an exponent. This exponent is given by
Pierce (1994) as 0.5 for ONAN and ONAF, and 1.0 for
OFAF and ODAF.
The heat generated by the active assembly (core and
coils) at the hot spot at time t
2
= t
1
+ t is given by:
4.4-44
Where:
Q
gen,O
is the heat generated within the transformer.
P
I2R
P
E
current.
P
S
is the stray loss.
P
C
is the core (no-load) loss.
4.4-45
Where:
W,1
is the average winding temperature calculated
1.
W,R
load.
K
aluminum windings.
The heat loss is given by:
4.4-46
Where:
Q
loss,AO
is the heat lost by the bottom oil to the sur-
rounding air.
A,1
is the ambient temperature at the previous
time step.
A,R
is the ambient temperature during the heat
run.
AO,1
is the average oil temperature at the previous
time step.
AO,R
is the average oil rise at rated load.
y is an exponent equal to 0.8 for ONAN, 0.9
for ONAF and OFAF, 1.0 for ODAF.
bulk oil is given by:
4.4-47
Where:
Q
abs,AO
is the heat absorbed by the average bulk oil
during incremental time step, t.
AO,1
is the average bulk oil temperature at time t
1.
AO,2
is the average bulk oil temperature at time t
2
= t
1
+ t.
( )
E R I
R DO R HS
DO HS
HS loss
P P Q +
=
2
, ,
1 , 1 ,
,

( )
t
C m Q
HS HS
W P W HS abs
=
1 , 2 ,
, ,

( )
1 ,
,
, ,
2 , HS
W P W
HS loss HS gen
HS
C m
t Q Q
+

=
( ) ( )
C W S E W R I O gen
P K P P K P L Q + + + =
2
2
,
+
+
=
K R W
K W
W
K

,
1 ,
T
y
R A R AO
A AO
AO loss
P Q
1
, ,
1 , 1 ,
,
=

( )
t
mC Q
AO AO
P AO abs
=

1 , 2 ,
,

4-30
The value mC
P
is a sum of product of the masses of the
transformer in contact with the oil, as well as the oil
itself, and the specific heat of those components:
4.4-48
ing for
AO,2
:
4.4-49
Once the average bulk oil temperature has been deter-
mined, a top-to-bottom bulk oil gradient must be calcu-
lated. This is done by the following equation:
4.4-50
Where,
TO/BO
is the top-to-bottom bulk oil gradient, C.
TO,R
is the top oil rise at rated load, C.
BO,R
is the bottom oil rise at rated load, C.
z is an exponent (0.5 for ONAN, ONAF; 1.0
for OFAF, ODAF).
The top and bottom oil temperatures are then solved for
by adding and subtracting, respectively, half of the top-
to-bottom bulk oil gradient:
4.4-51
4.4-52
4.4.4 IEC Model (IEC 354-1991)
The IEC 354-1991 (McNutt 1992) model is similar to
the IEEE Clause 7 model (Figure 4.4-5). The two mod-
els differ in a few important respects. First, the calcula-
tion of the rated hot spot rise is done by applying a hot
spot factor to the average winding rise over average oil
temperature gradient. This hot spot factor is a design
specific constant that generally varies between 1.0 and
1.4, with 1.2 being typical for power transformers.
Note: The winding temperature rise exponent in this
document differs slightly from that described in IEC
354. The equations listed below apply a multiple of 2 to
this exponent, whereas in IEC 354, this multiple is
included in the exponent definition. Therefore, the wind-
ing temperature rise exponent listed here is equal to half
of the winding temperature rise exponent in IEC 354.
Oil Thermal Time Constant:
In order to calculate the transient response of the bulk
oil, it is necessary to calculate an oil thermal time con-
stant. Since IEC 354 does not include this information,
the IEEE C57.91-1995 equations can be offered as an
option:
For OA and FA (ONAN and OFAF), the thermal
capacity, C, equals:
C = 0.0272 * W
CC
+ 0.01814 * W
Tank
+ 5.034 * V
Fluid
(1) (IEEE Clause 7) 4.4-53
For DFOA and NDFOA (ODAF and OFAF), the ther-
mal capacity, C, equals:
C = 0.0272 * W
CC
+ 0.0272 * W
Tank
+ 7.305 * V
Fluid
(2) (IEEE Clause 7) 4.4-54
( )
+ + =
fluid fluid P Tank C C steel P P
W C W W C mC
, & ,
( )
1 ,
, ,
2 , AO
P
AO loss O gen
AO
mC
t Q Q
+

=
( )
z
O gen
AO loss
R BO R TO BO TO
t Q
Q
=
,
,
, , /

2
/ BO TO
AO TO
+ =
2
/ BO TO
AO BO
=
Figure 4.4-5 Diagram illustrating assumed temperature profile within transformer for IEC model.
0.1
( )
1.108
4-31
The thermal time constant of the bulk oil at rated tem-
perature equals:
O,R
= C *
TO,R
/ P
T,R

(3) (IEEE Clause 7) 4.4-55
At top oil temperatures other than rated, the time con-
stant must be corrected as follows:
(4) (IEEE Clause 7) 4.4-56
Initial Temperatures:
The following equations are used to calculate the initial
temperatures based upon the assumption that the load
prior to the calculation period was constant long
enough for the temperatures to reach their steady-state
limits. This assumption is reasonable if the load is fairly
constant for a period of time prior to the overload equal
to two to three times the oil thermal time constant.
(5) (IEC 2.4.1, Equation 1) 4.4-57
(6) (IEC 2.4.2, Equation 2) 4.4-58
(7) (IEC 2.4, Equation 1+2) 4.4-59
g is difference between average winding and average oil
temperatures.
H is a multiplier that equals 1.3.
Hg is the hot spot rise above top oil.
Transient Temperatures:
The following equations are used to calculate the tran-
sient temperature rises based upon the oil time constant
calculated above, the user-determined winding time con-
stant, load, losses, and initial and ultimate temperature
rises. For a graphical representation of the various tem-
perature rises, see Figure 4.4.5.
For OA and FA (ONAN and ONAF) cooling modes:
The top oil rise at time t
2
= t
1
+ t is given by the follow-
ing:
(8) (IEC 2.4.1, Equation 1) 4.4-60
(9) (IEC 2.5) 4.4-61
The hot-spot rise over top oil at time t
2
= t
1
+ t is given
by the following:
(10) (IEC 2.4, Equation 1+2) 4.4-62
In addition, for DFOA (ODAF) cooling modes a cor-
rection factor is applied to account for the change in
winding Ohmic losses due to temperature as follows:
(11) (IEC 2.4.3, Equation 3) 4.4-63
For other types of cooling, the change in winding
Ohmic losses is negligible. Note that the IEEE method
neglects the change in winding resistance with tempera-
ture for all cases, assuming that this change is always
offset by the change in oil viscosity with temperature.
The transient hot spot rise is then:
(12) (IEC 2.5) 4.4-64
The equation form is the same as IEEE if the hot spot
rise over top oil term is calculated with the IEC method.
Note that the IEC method calculates the hot spot rise
over top oil temperature at rated load through the use of
a multiplier, H, applied to the difference between the
average winding temperature from test and the average
top oil temperature from test. For power transformers,
H is typically approximately 1.3.
2
= t
1
+ t is given by:
(13) (IEC 2.4.1, Equation 1) 4.4-65
n
R TO
i TO
n
R TO
U TO
R TO
i TO
R TO
U TO
R O O
1
,
,
1
,
,
,
,
,
,
,

( )
( )
n
i
R TO i TO
R
R K
+
+
=
1
1
2
, ,

( )
( )
n
i
R BO i BO
R
R K
+
+
=
1
1
2
, ,

m
i R i HS
K Hg
2
,
=
( )
( )
n
U
R TO U TO
R
R K
+
+
=
1
1
2
, ,

( )
1 , 1 , , 2 ,
1
TO
t
TO U TO TO
O
e

+
m
U R U HS
K Hg
2
,
=
( )
R HS U HS U HS U HS , , , ,
15 . 0 + =
( )
1 , 1 , , 2 ,
1
HS
t
HS U HS HS
W
e

+
2 , 2 , 2 , 2 , HS TO A HS
+ + =
4-32
For NDFOA and DFOA (OFAF and ODAF) cooling
modes:
For forced-oil cooling modes, the increased mixing
introduced by the forced circulation of the oil increases
the complexity of the model. The top oil temperature
leaving the winding is greater than the measured top oil.
Therefore, the top oil temperature leaving the winding is
calculated using the sum of the ambient temperature,
the bottom oil temperature rise, and the difference
between the top oil and the bottom oil rises. Note that
this differs from the IEEE methodology. The IEEE
method assumes that the measured top oil temperature
is equal to the temperature of the oil leaving the top of
the winding for all cooling modes.
The bottom oil rise at time t
2
= t
1
+ t is given by:
(14) (IEC 2.4.2, Equation 2) 4.4-66
(15) (IEC 2.5) 4.4-67
The top-to-bottom oil rise at time t
2
= t
1
+ t is given by
the following:
(16) (IEC 2.4.2, Equation 2) 4.4-68
(17) (IEC 2.5) 4.4-69
The hot-spot rise above top oil,
HS
, is calculated using
the same equations as used for OA and FA cooling
modes described above.
2
= t
1
+ t is given by:
(18) (IEC 2.4.2, Equation 2) 4.4-70
The above calculation is repeated for each time step of
the overload period, with the initial temperatures rises
equal to the temperature rises from the previous time
step. In addition, loss-of-life calculations are performed
at each time step (IEEE 1995b).
4.4.5 Proposed IEC Model
The proposed IEC model (IEC 1991) takes a bit of a rad-
ical departure from the earlier models. The authors of
this new model recognized that the hot spot temperature
increases more rapidly initially than is currently predicted
with traditional equations. This is commonly attributed
to an increase in the duct oil that occurs more quickly
than the increase in bulk oil. This phenomenon they
have dubbed the "hot spot bump" (see Figure 4.4-6).
Rather than taking an analytical approach to the prob-
lem, as had been taken by Pierce with the IEEE Annex
G model, the authors of the model took the approach of
applying corrective factors and multiple time constants.
The constants would be adjusted empirically to fit mea-
sured temperature data. Like the IEEE Clause 7 model,
the revised IEC model lumps the hot spot temperature
into a top oil rise over ambient and a hot spot rise over
top oil.
4.4-71
The computation of the constituent rises is a complex
arrangement of factors and time constants. The intent
of these factors is to simulate the hot spot bump by
inclusion of an additional exponential term in the calcu-
lation of the hot spot rise over top oil.
The calculation of the top oil temperatures proceeds in a
similar fashion to the IEEE Clause 7 model. First, the
ultimate top oil temperature for the given load level is
calculated:
4.4-72
Where:
TO,U
is the ultimate top oil temperature rise, C.
TO,R
is the rated top oil temperature rise, C.
K is the per unit load.
R is the ratio of load losses to no-load losses.
x is a constant exponent.
( )
( )
n
U
R BO U BO
R
R K
+
+
=
1
1
2
, ,

( )
1 , 1 , , 2 ,
1
BO
t
BO U BO BO
O
e

+
m
U R BO T U BO T
K
2
, ,
=
( )
1 , 1 , , 2 ,
1
BO T
t
BO T U BO T BO T
W
e

=

2 , 2 , 2 , 2 , 2 , HS BO T BO A HS
+ + + =

Figure 4.4-6 Example of hot spot bump.
TO HS TO A HS /
+ + =
x
R TO U TO
R
R K
+
+
=
1
1
2
, ,

4-33
The transient top oil rise is then calculated by the fol-
lowing equation (note the additional constant applied to
the time constant):
4.4-73
Where:
TO,2
is the top oil temperature rise at the current
time step, C.
TO,1
is the top oil temperature rise at the previous
time step, C.
t is the time step, minutes.
k
11
is a transformer specific constant.
O
is the oil thermal time constant, minutes.
The hot spot rise over top oil is then calculated. The
ultimate hot spot rise over top oil is calculated in the
traditional manner:
4.4-74
Where:
HS/TO,U
is the ultimate hot spot rise over top oil, C.
HS/TO,R
is the rated hot spot rise over top oil, C
(IEC defines this as H*g, where H is a hot spot
factor between 1.1 and 1.5 and g is the average
winding to average oil gradient).
K is the load in per unit.
y is a constant exponent.
The most significant difference between this new IEC
model and more traditional calculations is in the tran-
sient formulation of the hot spot rise over top oil. The
new IEC model breaks the rise into two components
with different time constant, one on the order of the
winding time constant and the other on the order of the
oil time constant. The transient hot spot rise over top oil
is calculated as follows for increasing load:
4.4-75
Where:
HS/TO,2
is the hot spot rise over top oil at the
current time step, C.
HS/TO,1
is the hot spot rise over top oil at the
previous time step, C.
k
21
& k
22
are transformer specific constants.
t is the time step, minutes.
W
is the winding thermal time constant,
minutes.
O
is the oil thermal time constant, minutes.
For decreasing loads, the effects of thermal capacity are
ignored, and the hot spot rise over top oil is taken as the
ultimate hot spot rise over top oil. The individual terms
are then summed to give the hot spot temperature:
4.4-76
The various k-constants and the exponents x and y are
transformer specific. The recommended method for
determining these values is via extrapolation from the
heating curves of a prolonged factory heat run test. This
information would not be available for existing trans-
formers, and it is unclear whether it could even be derived
using standard factory heat run procedures. In lieu of
measured values, Table 4.4-3 provides suggested values.
4.5 THERMAL RATINGS
To this point, this chapter has examined the risks of
increased loading and determined ways to predict the
pertinent equipment temperatures. Now a method for
maintaining reasonable risk levels under everyday oper-
ation must be developed. The manner in which trans-
former ratings are calculated and communicated varies
from utility to utility, depending upon operating proce-
dures. Thermal ratings are used for various purposes,
ranging from planning to operation. The form and com-
plexity of the presentation of calculated ratings vary
with the preference of the user.
The loading of transformers is ultimately an economical
decision. The increased risk of failure and reduced
usable life must be balanced with the capital investment
of the unit and, to a smaller degree, increased mainte-
nance costs. In addition, the impact of the loss of the
unit on the integrity of the system must be factored in.
Therefore, the ultimate decision on rating limits is up to
the individual utility. Presented here are general recom-
mendations.
Ratings can be viewed as a function of several factors:
Ambient temperature
magnitude
diversity
Load Shape (diversity)
Pre-Load
( )

+ =

O
k
t
TO U TO TO TO
e

11
1
1 , , 1 , 2 ,
y
R TO HS U TO HS
K
, / , /
=
( )
( )

+ =

22 22
/
21 21
1 , / , / 1 , / 2 , /
1 1 1
k
t
k
t
TO HS U TO HS TO HS TO HS
O W
e k e k

TO HS TO A HS /
+ + =
Table 4.4-3 Suggested Values for Constants in Revised IEC
Model
ONAN ONAF OFAF ODAF
Oil exponent, x 0.8 0.8 1.0 1.0
Winding exponent, y 1.3 1.3 1.3 2.0
Constant k
11
0.5 0.5 1.0 1.0
Constant k
21
2.0 2.0 1.3 1.0
Constant k
22 2.0 2.0 1.0 1.0
Oil Time Constant (min) 210 150 90 90
Winding Time Constant
(min)
10 7 7 7
4-34
Rating Duration
Power transformer ratings are traditionally calculated
over a 24-hour period. The general procedure is to cal-
culate the temperatures and loss-of-life over the 24-hour
period and compare these to pre-selected limits. The
load is adjusted accordingly such that none of the limits
is exceeded and one or more limits are met. This is then
the definition of the rating. The steps to calculating the
rating are as follows:
1. Determine initial temperatures.
Since transformers have a significant thermal capac-
ity, the initial temperatures prior to the rating become
an important factor, particularly with rating dura-
tions less than twice the oil thermal time constant.
Therefore, the initial operating temperatures of the
transformer must be estimated by some means. The
initial temperatures can be estimated directly (possi-
bly from measured temperatures) or calculated by
calculating the temperatures at the end of a 24-hour
period of assumed loading.
2. Calculate equipment temperatures and loss-of-life for
rating period.
Using any available means, such as those outlined in
Section 4.4, the operating temperatures, principally
hot spot and top oil, are calculated throughout the
rating period. In addition, loss-of-life is calculated.
3. Compare maximum temperatures and loss-of-life
from Step 2 to selected limits.
4. Adjust load level and return to Step 2.
4.5.1 Ambient Air Temperature
Obviously, the ambient air temperature is a major factor
in the load capacity of a power transformer (Figure
4.5-1). Each degree that average 24-hour air tempera-
ture is below 30C represents additional capacity. There-
fore, seasonal variations in air temperature can be used
to give higher thermal ratings for cooler periods.
Seasonal Variation
Depending upon geographical location, the temperature
can vary widely between seasons. This allows for
increased capacity during cooler months. Depending
upon the nature of the load, this can be a significant
advantage, especially in cooler climates. If the load is
largely heating and lighting loads, then the peak system
loading tends to occur in the winter, when ambients are
the lowest and thermal capacity the highest. On the flip
side, in areas with summer peaking loads, the peak load
coincides with the peak ambient temperatures and con-
sequently the minimum capacity periods.
Often times, users produce thermal ratings based upon
seasonal peak or average ambients. This allows the user
to take advantage of the additional capacity during peri-
ods where the ambient is below the rated ambient. Rat-
ings of this kind are simple and safe. Significant
additional capacity can be realized without even exceed-
ing the rated hot spot temperature of 110C.
Diurnal Variation
Diurnal variations in ambient temperature are often
predictable. Due to the thermal capacity of the trans-
former, cyclic variations in ambient temperature can be
used to advantage.
Measurement
Ambient temperature should be measured as close as
possible to the transformer. Ideally, one would want the
temperature of the air in the vicinity of the radiators or
heat exchangers. This is generally not practical. Temper-
atures measured within the substation should be suffi-
cient. Weather bureau data may be used with the
understanding that temperatures can vary significantly
over a short distance.
For transformers located in any sort of enclosure such
as a kiosk or vault, temperatures should be used that are
measured in the vicinity of the transformer. If this is not
possible, an offset should be added to the ambient tem-
perature to account for the increased local temperatures.
4.5.2 Load
Pre-load
For transient ratings, the thermal mass of the trans-
former makes the rating dependent upon the initial tem-
peratures at the start of the rating period. These initial
temperatures are governed by the loading for the previ-
ous 24 hours. Therefore, for rating durations less than
approximately twice the oil thermal time constant, pre-
load is a significant factor. An illustration of this is
Load Shape
Also due to the long thermal time constant of the trans-
former, load shape can be a factor in rating calculations.
Figure 4.5-1 Example plot of rating vs. air temperature
for various durations with pre-load = 0.7 PU.
4-35
A load shape that is flat will achieve higher tempera-
tures than a diverse (large difference between minimum
and maximum) load shape for a given peak magnitude.
This can often be used to advantage if the loading on a
particular unit is cyclic and predictable.
4.5.3 Rating Type and Duration
A thermal rating is a statement of the load capability of
a transformer under a given set of conditions. These
conditions represent various typical operating scenarios
that a system operator may run into. The system opera-
tor then selects the rating scenario that most represents
the current operating scenario to determine the load
capability of the transformer for the current operating
scenario. Depending upon the application of the trans-
former, the system conditions impacting the trans-
former, and company operating policies, various rating
scenarios may be developed. However, these rating sce-
narios can generally be lumped into three categories:
- Normal life-expectancy loading
- Long-term emergency (LTE)
- Short-term emergency (STE)
Normal Life-Expectancy Loading
Normal life-expectancy loading represents the normal
operating state of the transformer. A transformer repre-
sents a large capital investment, and therefore it makes
good business sense to get a full life expectancy out of
the unit. Increased temperatures result in accelerated
aging rates, and therefore a lower return on investment.
The normal rating represents the load limit for con-
tinuous operation.
Long-Term Emergency
The definition of a long-term emergency rating (LTE)
varies within the industry. Typically, it denotes a rating
where the thermal capacity of the equipment does not
greatly impact the rating. For power transformers, the
length of the oil time constant makes this definition a
bit unclear given the cyclic variation in load and air tem-
perature. For the sake of discussion and throughout this
chapter, a long-term emergency rating is defined as a
rating greater than 4 hours in duration.
Short-Term Emergency
As with LTEs, the definition of short-term emergency
ratings (STE) varies. Short-term ratings are usually
defined as extremely short duration ratings that take
advantage of the thermal capacity of the equipment.
These ratings range from a few minutes in duration to a
few hours. In this document, STEs are defined as ratings
4 hours and less. An example of a 4-hour STE rating is
shown in Figure 4.5-3, demonstrating the pre-load, rat-
ing period, and post contingency period.
It is important to note here that short-duration ratings
are highly sensitive to the inaccuracies in the transient
temperature modeling of transformers. Therefore, cau-
tion should be exercised when calculating ratings less
than 1 hour in duration.
4.5.4 Rating Procedure
As a guideline, the following procedure for initial rating
or re-rating of a transformer is proposed.
1. Gather information
Information essential to the loading calculations
must be gathered. This information must include, at
minimum, the factory test report and the nameplate
drawing. It is highly recommended that maintenance
history (DGA, oil quality, etc.) and loading history
be obtained. Outline drawings may be of use as well.
2. Assess condition
The full scope of this step is beyond this guide; how-
ever, it is important to recognize the importance of
condition. Guidelines for condition assessment with
regard to loading are given below.
Figure 4.5-2 Example plot of rating vs. pre-load for
various rating durations.
Figure 4.5-3 Example 4 Hr STE rating.
4-36
3. Examine long-term risks
When considering a particular loading event, the first
things examined are the long-term risks. This mainly
consists of evaluating the loss of insulation life out-
lined above. This is an economic decision, balancing
the cost benefit of the increased load levels with the
decreased return on capital investment. If the loading
is short duration, the loss of life may be negligible,
and therefore the only concern is avoiding immediate
failure.
4. Examine short-term risks
If the increased life consumption is acceptable, the
concern then shifts to avoiding immediate failure.
This means limiting operating temperatures, particu-
larly hot spot temperature.
5. Check auxiliary equipment
Most of the attention when presented with a poten-
tial overload scenario is on the oil and winding tem-
peratures. However, the auxiliary equipment
(bushing, LTCs, etc.) should not be neglected. While
these should be sized properly to avoid limiting the
transformer rating, this is not guaranteed. Before
proceeding with increased loading, it is often advis-
able to verify the thermal capability of these devices.
6. Follow-up post loading
Once the decision has been made to permit with
increased loading, any significant overload events
should be followed up with a site inspection and oil
analysis. This should be done to verify that there were
no unintended consequences of the increased loading
event and to determine suitability for future overload.
4.5.5 Condition-Based Loading
Rather than proposing a static set of limits for various
loading situations, a condition-based methodology is
presented here. Limits are first given for healthy trans-
formers. These are to be viewed as maximum limits.
These maximum limits are then reduced based upon the
assessed condition of the unit. Conditions necessitating
the reduction of the temperature limits include:
moisture content of bulk insulation
oxygen content
gassing history
criticality of service
For this guide, three condition categories are proposed:
Good, Moderate, and Marginal. Transformers are
placed in each category using the several different crite-
ria outlined above. The lowest category for any particu-
lar criterion is the category for the unit. For example, if
a unit has an Ethylene content of less than 36 ppm, but
has a moisture content of 2%, the unit falls in the Mar-
ginal category (Table 4.5-1).
Once the unit has been placed in a condition category,
the criticality of service, or the impact of failure, should
be considered. If the unit is crucial to system integrity,
or is the only GSU for a generation plant with no spare,
it may be advisable to use the rating limits for a Mod-
erate or Marginal unit, even if the condition suggests
Good. The rating limits for each condition category
are given in Table 4.5-2. These should be taken as guide-
Table 4.5-1 Criteria for Condition Categories Used to
Determine Rating Limits
Good Moderate Marginal
Moisture < 0.5% 0.5%-1.5% > 1.5%
Oxygen < 3% TDG 3%-5% TDG > 5% TDG
Methane < 120 ppm 120-400 ppm > 400 ppm
Ethane < 65 ppm 65-100 ppm > 100 ppm
Ethylene < 50 ppm 50-100 ppm >100 ppm
Table 4.5-2 Condition-Based Rating Limits for Transformers
Condition Normal LTE (> 4 hrs) STE (< 4 hrs)
GOOD
Top Oil 95 105 110
Hot Spot 120 140 160
LOL (hrs) 24 - -
Normal LTE (> 4 hrs) STE (< 4 hrs)
MODERATE
Top Oil 95 105 105
Hot Spot 120 130 140
LOL (hrs) 24 - -
Normal LTE (> 4 hrs) STE (< 4 hrs)
MARGINAL
Top Oil 95 100 100
Hot Spot 110 120 120
LOL (hrs) 24 - -
4-37
lines and not necessarily as God-given rules. Each user
has different policies and risk tolerances.
4.5.6 Maintenance Considerations
Given the multitude of unknowns in loading power
transformers, it is essential that increased vigilance be
exercised. The reasons for this are twofold. First, the
assumptions made in evaluating the risks of loading
need to be verified. Second, the unit must be maintained
in top working order to prevent conditions that may be
exacerbated by increased loading.
Throughout the thermal rating process, assumptions are
made to turn a problem of enormous complexity into a
manageable process. This is what distinguishes engineer-
ing from science. The most important of these assump-
ti ons are the l ocati on of the thermal ly l i mi ti ng
temperature near the top of the winding, adequately
sized and applied leads, and proper construction and
design.
These assumptions need to be verified in some manner.
Direct measurement of the temperatures of the winding
is impossible unless fiber optic probes had been installed
during manufacture. Even with the probes installed, this
does not preclude the existence of an unintended hot
spot elsewhere in the transformer.
Increased Preventive Maintenance
One often neglected consideration in the loading of
transformers is the need for increased maintenance. The
cooling equipment must be maintained in top working
order. Loss of even partial cooling could be disastrous.
If a unit were loaded to the nameplate rating of the
highest cooling mode of an OA/FA/FA and all cooling
fans were lost, the unit would be overloaded by roughly
66%, resulting in temperatures exceeding 200C. In
addition, various components will age or wear more
quickly.
Prior to initial re-rating or increased load, a field inspec-
tion should be scheduled. At this time, the unit should
be subjected to careful scrutiny. In particular, the fol-
lowing should be checked:
Make sure all pumps and fans are operational. Manu-
ally switch the cooling on and off to ensure proper
operation. Listen to the pumps, if possible, for
sounds of cavitation. Occasionally, phases on the
pumps are reconnected in the wrong phase order,
causing the pump to run in reverse. Note the position
of all radiator valves. If a radiator or bank of radia-
tors is valved off, do not open the valve. Make a note
of this, and derate the unit accordingly.
Calibrate and check all gauges. This is particularly
important if the cooling equipment is switched on/off
by the temperature gauges. Oil temperature should be
double-checked by placing a handheld thermocouple
on the tank wall as close to the thermal well as safety
allows. Once the oil gauge has been calibrated, the
estimated hot spot temperature should be calculated
for the measured top oil temperature and the given
load. The ratio of the CT feeding the heating element
of the analog WTI (winding temperature indicator)
(if equipped) should be adjusted until the gauge reads
the calculated value. Finally, check that all snap
switches or setpoints for cooling switching and
alarms are properly set. Be sure the alarm and trip
settings coincide with appropriate rating limits. Cal-
culated ratings do no good if the unit will trip at
lower temperatures!
Check the oil level. Make sure the oil is at the level
specified in the operating manual for the measured
oil temperature. As mentioned previously, oil will
expand with increasing temperature. Too much oil
will result in operating of the pressure relief device
and expulsion of oil. Too little oil could result in
exposing the active assembly upon cooling.
Check all gasketed surfaces for leaks. Specifically,
check the bushings, both at the tank cover and at the
top of the bushing, and any other tank penetrations.
Gaskets in particular will harden and become brittle
more rapidly than at normal operating temperatures.
As this happens, the gaskets can begin to leak, result-
ing in reduced oil levels and ingress of moisture and
oxygen. Should this occur, the unit must then be
taken out of service at some point to replace the gas-
kets. If this occurs with sufficient frequency, it may be
advisable to replace the gaskets with a higher temper-
ature material such as Viton.
Check the tank wall for discolored paint. Areas of high
temperatures due to stray flux heating in the tank
wall may result in noticeable discoloration of the
paint. Pay close attention to bushing penetrations.
Check conservator and gas blanket systems. Make sure
the oil conservator, if equipped, is filled to the proper
level and all piping is leak free. Make sure any desic-
cant canisters are filled with a suitable desiccant. If
equipped with an inert gas blanket system, make sure
that the pressure regulator is operating properly and
that sufficient gas quantity is available in the canister.
Gas over-pressure can cause supersaturation of the oil
at high temperatures. Upon cooling, the gas would
come out of solution and free bubbles would form.
Check the LTC for proper operation, if equipped. If
possible, measure the temperature of the LTC com-
partment oil. If the differential between the LTC oil
4-38
and the main tank exceeds 15C, consider scheduling
an outage to investigate the cause. If system condi-
tions permit, consider cycling the LTC throughout
the tap range, in particular exercising the reversing
switch.
Draw an oil sample for DGA and oil quality. Safe over-
loading requires the oil to be in good condition. In
addition, DGA provides a useful tool for spotting
lead heating or stray flux heating problems, or any
other unknown thermal conditions. In addition, units
with possible incipient problems should never be con-
sidered for overload. If an evaluation of insulation
condition with respect to thermal aging is desired,
Furan analysis could be performed as well.
The above checks should be repeated as part of a
preventive maintenance program at the usual scheduled
interval. If a unit is frequently loading near or above
nameplate, and increased preventive maintenance sched-
ule should be considered. In addition, following more
severe emergency overloads, a site inspection should be
scheduled at the earliest convenient time to ensure that
no damage was done and that the unit is operating
normally.
4.6 WINDING TEMPERATURE MEASUREMENT
Temperature monitoring and measurement is an impor-
tant tool in loading power transformers. It provides
feedback on the cooling performance of the trans-
former, confirmation of calculated temperatures, and a
reliable assessment of the current operating condition of
the transformer. That said, accurate temperature mea-
surement is difficult and can be expensive. Depending
upon the technology, it may not be economical for
smaller units.
Simulated WTI
Simulated WTIs (winding temperature indicators) are
by far the most common devices for measuring winding
temperatures (Figure 4.6-1). In reality, however, these
devices do not actually measure the winding tempera-
ture. They simply measure the temperature of a specially
calibrated heating element that is immersed in the top
bulk oil near the tank wall. These devices simulate the
actual winding temperature.
These most common analog devices consist of a brass
tube, or well, that is mounted on the side of the tank,
such that it is immersed in the top oil. A separate heat-
ing element is the placed in the tube, and a current pro-
portional to the winding current is passed through the
heating element. This current produces a temperature
rise that, at rate load, equals the expected hot spot tem-
perature.
Since these devices do not directly measure the winding
temperature, they are inherently inaccurate. The ther-
mal characteristics of the devices do not exactly match
that of the transformer windings. At higher loads, the
error can be significant. In addition, these devices have a
longer time constant, making them inaccurate during
transient shifts in load (Teetsel 2003). This can be signif-
icant, especially if the cooling equipment is switched by
the WTI temperature.
Fiber Optic Temperature Measurement
First introduced in the early 1980s, use of fiber optic
temperature probes provides a method for direct tem-
perature measurement of a point on the surface of the
winding insulation. This is possible due to the inherent
dielectric strength of the silica fiber optic material.
There are essentially two viable competing technologies
in this area: fluoroptic thermometry and abortion shift
of GaAs semiconductor.
Fluoroptic thermometry probes utilize a phosphor coat-
ing at the tip. When the coating is excited with a pulse of
light sent down the fiber optic, the coating fluoresces.
The rate of decay of this fluorescence is temperature
dependent. Therefore, by measuring the rate of decay at
the far end of the fiber optic, the temperature of the
probe tip can be determined.
GaAs semiconductor devices utilize a GaAs semicon-
ducting wafer topped with a reflective coating at the end
of the probe. A pulse of white light is sent down the
fiber, passing through the GaAs wafer. Some of the light
is absorbed by the GaAs, while the remainder passes
through and reflects off the reflective coating. The light
then returns down the fiber where the magnitude of the
received light is measured for the various wavelengths of
the spectrum of the light pulse. Light at different wave-
lengths is absorbed by the GaAs, with the magnitude at
a particular wavelength a function of temperature of the
GaAs semiconductor. Therefore, the absorption spec-
Figure 4.6-1 Schematic drawing of winding temperature
indicator.
4-39
trum is a function of the temperature of the GaAs. The
characteristic curve of the absorption spectrum shifts
toward higher wavelengths with increasing tempera-
tures. By measuring the shift in the absorption spec-
trum, the temperature of the probe is determined.
Regardless of the technology used, fiber optic tempera-
ture probes have a few significant disadvantages. First,
these probes must be installed during manufacture. The
probes must be inserted in the winding at the expected
hot spot location. Since this location is difficult to pin-
point, several probes at different locations must be used.
In addition, hot spot locations can move with differing
oil flow regimes, making precise hot spot determination
impossible. These probes are also extremely delicate.
Even with the newer, ruggedized probes, breakage often
occurs. The manufacturer must take extra precautions
during winding and final assembly not to break the
fibers. This in turn increases the cost of manufacture for
the unit as a whole. The measuring units for reading the
probes are also rather expensive, making them economi-
cal only for larger, critical units.
4.7 MODEST INCREASES IN CAPACITY FROM
EXISTING TRANSFORMERS
When additional capacity is required, and higher tem-
peratures are not practical or acceptable, there are lim-
ited options for increasing the rating of a transformer by
adding additional cooling. This can be effective to some
extent, but care must be taken to ensure that the wind-
ing hotspot is not excessive.
There are essentially two ways to reduce transformer
temperatures: reduce the losses or increase the heat
transfer. The former would require a redesign and
rewind of the transformer and is therefore not economi-
cal unless the unit is already being rewound following a
failure. Even then, other design constraints will limit the
amount of additional capacity. The only viable option
is, therefore, to increase the heat transfer. To be cost
effective, any cooling upgrades must be installable in the
field. This precludes any internal modifications.
Additional pumps and fans may be added, with limited
increase in capacity. There is a practical maximum to
pump flow rates or fan capacities, above which there will
be no increase in heat transfer. For units employing heat
exchangers, larger heat exchangers may be added if the
cost of the retrofit is justified by the increased capacity.
Methods of increasing the cooling capacity include:
addition of fans, or higher flow rate fans, to radiators
or heat exchangers
retrofit radiators or heat exchangers with higher cool-
ing capacity units
water spray cooling over radiator fins
retrofit with oil-water heat exchangers
All of the methods listed above increase the heat trans-
fer from the oil to the surrounding environment. The
effect would then be to decrease both the average oil rise
and the top-to-bottom oil gradient. In non-directed
flow designs, this reduction in oil temperature increases
the thermal pressure of the natural thermosiphon flow
through the windings. Assuming that the duct size is not
limiting the oil flow in the area of the hot spot, the hot
spot temperature is then decrease.
However, this leads to a major caveat in the application
of supplemental cooling: the drop in oil temperature
does not necessarily give a corresponding drop in the
hot spot temperature. In other words, if the oil tempera-
ture is reduced 10C, the hot spot temperature will be
reduced somewhere between 0 and 10C. Without
detailed design data or embedded fiber optics, it is
impossible to determine the actual decrease in hot spot
temperature. Therefore, the excess capacity, if any,
gained by applying supplemental cooling should not be
relied upon for planning or operating purposes unless
the manufacturer has reviewed the application.
4.8 EXAMPLES
Example 1:
This first example will illustrate a simple 6 hr transient
rating with a flat preload and load cycle, followed by a
6 hr transient rating with a cyclical preload and load
cycle. The unit is a 220/69-kV autotransformer rated at
224 MVA with directed forced-oil cooling (DFOA).
For this unit, the bottom oil temperature rise at rated
load is known, so the more accurate IEEE Annex G
model is used. The parameters for the thermal calcula-
tion are as shown in Table 4.8-1.
As mentioned previously, the preload and load cycles
are flat. A flat preload enforces a conservative assump-
tion that the transformer temperatures have reached
steady-state (and therefore the maximum) prior to the
onset of the contingency. The contingency onset is
assumed to occur at 12 pm to coincide with the peak
ambient. The load shape during the contingency is also
assumed to be flat, again giving a conservative answer
(Figure 4.8-1).
The ambient temperature used for the calculation is a
representative peak day for the summer months mea-
sured at the substation. This represents the worst-case
4-40
scenario where the contingency occurs on the hottest
day of the season (Figure 4.8-2).
This particular unit was manufactured in 1997 and is in
good working order. Therefore, the temperature limits
for this LTE rating shall be set at 140 hot spot tempera-
ture and 110C top oil temperature. Given the above
load and air temperature cycles and the temperature
limits, the load over the 6 hour rating interval is
increased until either the hot spot temperature or the
top oil temperature meets the temperature limit. This is
demonstrated in Figure 4.8-3. The 6 hr LTE rating for
this specific scenario is 278 MVA.
In the above scenario, the conservative approximation
of assuming a flat preload and load cycle was used.
Where the load follows a predictable cyclical pattern,
increased capacity can be safely realized by using the
actual load shapes. Recalculating the rating above utiliz-
ing the preload and load cycles, shown in Figure 4.8-4,
reveals a slight increase in the rating from 278 MVA to
291 MVA.
Note that the preload in this example peaks higher at
0.88 pu than the flat preload of 0.7 pu assumed above.
However, the preload is below 0.7 the majority of the 24
hour cycle, with an average load of 0.53. Given the ther-
mal time constant of the transformer bulk oil, the oper-
ating temperatures with a diverse load shape will be
fractionally less than the temperatures with a flat load
shape (Figure 4.8-5).
Table 4.8-1 Parameters for Example 1
MVA Base for Loss Data 200 MVA
Temperature Base for Loss Data 75 C
Winding I2R Losses 525072 W
Winding Eddy Losses 0 W
Stray Losses 0 W
Core Losses 54560 W
Cooling Mode Type ODAF
Nameplate MVA 224 MVA
Guaranteed Average Winding Rise 65 C
Rated Average Winding Rise 50.6 C
Rated Hot Spot Rise 62.2 C
Rated Top Oil Rise 32.2 C
Rated Bottom Oil Rise 29.2 C
Rated Ambient Temperature 30 C
Winding Conductor Copper
Per Unit Eddy Loss at Winding Hot Spot 0
Winding Time Constant 5 min
Per Unit Winding Height to Hot Spot 1
Weight of Core & Coils 225500 lbs
Weight of Tank & Fittings 102600 lbs
Fluid Type Mineral Oil
Oil Volume 21696 gals
Load
0
0.2
0.4
0.6
0.8
1
1.2
1.4
0 10 20 30 40 50 60
Time (hrs)
L
o
a
d

(
P
e
r

U
n
i
t
)
Figure 4.8-1 Preload and load cycle for Example 1.
Ambient Temperature
0
5
10
15
20
25
30
35
0 10 20 30 40 50 60
Time (hr s)
A
m
b
i
e
n
t

T
e
m
p

(
d
e
g

C
)
Figure 4.8-2 Air temperature cycles for Example 1.
Figure 4.8-3 Calculated transformer temperatures for
Example 1.
Load
0
0.2
0.4
0.6
0.8
1
1.2
1.4
0 10 20 30 40 50 60
Ti me (hr s)
L
o
a
d

(
P
e
r

U
n
i
t
)
Figure 4.8-4 Preload and load cycles for Example 1 with
cyclical load.
4-41
Example 2
This second example will answer a question: A substa-
tion transformer is feeding a circuit in parallel with two
other transformers. One of the three transformers must
be removed from service for 48 hours. Can the remain-
ing two units handle the increased load safely?
The unit in this example is a 220/138-kV autotrans-
former rated at 180/240/300MVA OA/FA/FA. Again,
the Annex G model is used with the parameters shown
in Table 4.8-2.
The predicted load for the days of the outage is shown
in Figure 4.8-6. The lower curve shows the predicted
load with all units in service. The upper curve shows the
load with two units in service. Figure 4.8-7 shows the
predicted air temperatures in the vicinity of the substa-
tion for the 48 hour period under question.
Figure 4.8-8 reveals that the hot spot temperature peaks
at about 140C and the top oil temperature reaches
approximately 100C. Given the short duration of the
event, these temperatures should be safe, assuming the
in-service units are in good working order.
Figure 4.8-5 Calculated transformer temperatures for
Example 1 with cyclical load.
Load
0
0.2
0.4
0.6
0.8
1
1.2
1.4
0 10 20 30 40 50 60
Ti me (hrs)
L
o
a
d

(
P
e
r

U
n
i
t
)
Load (PU)
Original Load
Figure 4.8-6 Predicted load during outage for Example 2.
Ambient Temperature
0
10
20
30
40
50
0 10 20 30 40 50 60
Time (hr s)
A
m
b
i
e
n
t

T
e
m
p

(
d
e
g

C
)
Figure 4.8-7 Predicted air temperature for outage.
Table 4.8-2 Parameters for Example 2
MVA Base for Loss Data 180 MVA
Temperature Base for Loss Data 65 C
Winding I2R Losses 146265 W
Winding Eddy Losses 0 W
Stray Losses 0 W
Core Losses 47894 W
Cooling Mode Type FA
Nameplate MVA 336 MVA
Guaranteed Average Winding Rise 55 C
Rated Average Winding Rise 53.6 C
Rated Hot Spot Rise 68.6 C
Rated Top Oil Rise 38.2 C
Rated Bottom Oil Rise 12.4 C
Rated Ambient Temperature 40 C
Winding Conductor Copper
Per Unit Eddy Loss at Winding Hot
Spot
0
Winding Time Constant 5 min
Per Unit Winding Height to Hot Spot 1
Weight of Core & Coils 190600 lbs
Weight of Tank & Fittings 145750 lbs
Fluid Type Mineral Oil
Oil Volume 26973 gals
Figure 4.8-8 Calculated transformer temperatures during
outage
4-42
Figure 4.8-9 shows the cumulative aging during the out-
age, in hours, for the in-service units. The total aging
over the 48 hour period is approximately 220 hours. If
this event is relatively rare, this increased aging should
be acceptable over the expected 180,000 hour life of the
unit.
Insulation Aging
0
50
100
150
200
250
0 10 20 30 40 50 60
Ti me (hrs)
C
u
m
u
la
t
iv
e

A
g
in
g
(
h
r
s
)
0
5
10
15
20
25
A
g
e
A
c
c
e
le
r
a
t
i
o
n

R
a
t
e
Total Aging(hrs) Age Acceleration Rate
Figure 4.8-9 Insulation aging during outage.
4-43
REFERENCES
Aubin, J. and T. Langhame. 1992. Effect of Oil Viscos-
ity on Transformer Loading Capability at Low Ambient
Temperatures. IEEE Trans. on Power Delivery. Vol. 7.
No. 2. pp. 516-524. April.
Cengel, Y. A. 1997. Introduction to Thermodynamics and
Heat Transfer. McGraw-Hill. Boston, MA.
Dakin, T. W. 1948. Electrical Insulation Deterioration
Treated as a Chemical Rate Phenomenon. AIEE
Transactions. Vol. 67. pp. 113-122. November.
Emsley, A. M. and G. C. Stevens. 1994. Review of
Chemical Indicators of Degradation of Cellulosic Elec-
trical Paper Insulation in Oil-filled Transformers. IEE
Proc. Sci. Meas. Technol. Vol. 141. No. 5. pp. 324-334.
September.
IEC. 1991. IEC 60354: 1991. Loading Guide for Oil-
immersed Power Transformers. January.
IEC. 2004. IEC 60076-7. Power Transformers Part 7:
Loading Guide for Oil-immersed Power Transformers.
Committee Draft 2.
IEEE. 1995a. Standard Requirements for Load Tap
Changers. IEEE Standard C57.131-1995. March.
IEEE. 1995b. Guide for Loading Mineral-Oil-Immersed
Transformers. IEEE Standard C57.91-1995. June.
IEEE. 1995c Guide for Application of Power Apparatus
Bushings. IEEE Standard C57.19.100-1995. August.
Lundgaard, L., W. Hansen, D. Linhjell, and T. Painter.
2004. Ageing of Oil Impregnated Paper in Power
Transformers. IEEE Trans. on Power Delivery. Vol. 19.
No. 1. pp. 230-239. January.
Lesieutre, B. C. Hagman, W.H. and J. L Kirtley Jr. 1997.
An Improved Transformer Top Oil Temperature
Model for Use in An On-Line Monitoring and Diag-
nostic System. IEEE Trans. on Power Delivery. Vol. 12
No. 1. pp. 249-256.
McNutt, W. J. 1992. Insulation Thermal Life Consid-
erations for Transformer Loading Guides. IEEE
Transactions on Power Delivery. Vol. 7. No. 1. pp 392-
401. January.
Montsinger, V. M. 1930. Loading Transformers by
Temperature. Trans. AIEE. Vol. 49. pp 776-790.
Montsinger, V. M. 1951. Transformer Engineering. John
Wiley & Sons. New York. pp. 275-351.
Pierce, L. W. 1992.An Investigation of the Thermal
Performance of an Oil Filled Transformer Winding.
IEEE Trans. on Power Delivery. Vol. 7. No. 3. pp. 1347-
1358. July.
Pierce, L. W. 1994. Predicting Liquid Filled Trans-
former Loading Capability. IEEE Trans. on Industry
Applications. Vol. 30. No. 1. pp. 170-178. January/Feb-
ruary.
Schroff, D. H. and A. W. Stannett. 1985. A Review of
Paper Ageing in Power Transformers. IEE Proc. Vol.
132. Pt. C. No. 6. pp. 312- 319. November.
Teetsel, M. 2003. Winding Temperature Measurement:
Techniques, Devices and Operation. From presentation
before IEEE/PES Transformers Committee. October.

5-1
CHAPTER 5 Substation Terminal
Equipment
5.1 INTRODUCTION
In comparison to overhead lines, underground cables, and power transformers, substa-
tion terminal equipment is usually much less expensive to replace, and since terminal
equipment is within the utility substation boundaries, its replacement does not require
public hearings or regulatory approval. Nonetheless, in many cases, the benefits of
increasing the rating of lines, cables, and transformers may be limited by one or more
breakers, line traps, or current transformers, and replacement of such equipment can be
both time consuming and disruptive due to required circuit outages. Finally, just as with
lines, cables, and transformers, when relatively modest increases in terminal equipment
rating are required, more detailed knowledge of thermal behavior can often be obtained
quite easily.
Substation terminal equipment consists of many different types and designs of power
equipment. Included in this classification are line traps, oil circuit breakers, SF
6
circuit
breakers, rigid tubular bus, line disconnects, current transformers, bolted connectors, and
insulator bushings.
In a recent Electra article entitled Dynamic Loading of Transmission Equipment An
Overview (CIGRE 2002), representatives of Study Committee 23 concluded, there is
scope for implementing dynamic loading principles for a wide range of transmission
assets. The need for increased power flow in substation terminal equipment is illustrated
in Figure 5.1-1 taken from (New York Power Pool 1982). It shows that the thermal rating
of over 50% of the transmission circuits in New York State are thermally limited by sub-
station equipment.
Figure 5.1-1 Thermally limiting transmission circuit equipment.
Chapter 5: Substation Terminal Equipment Increased Power Flow Guidebook
5-2
The increase in circuit rating, resulting from applying
the various methods of increasing power flow in over-
head transmission lines, underground cable, and power
transformers is often limited by terminal equipment, as
shown in Figure 5.1-2. This figure illustrates the unex-
pected conclusion that relatively modest investments in
terminal equipment (replacement of the CT in Circuit
A) yields an increase in that circuit rating and a 50-MW
increase in the rating of the complex interface. Here a
large increase in circuit rating is obtained for a very
modest expenditure on terminal equipment rather than
a relatively large investment in lines, cables, or trans-
formers.
This chapter is limited to studying practical, rather sim-
ple methods of increasing the power flow through less
capital-intensive equipment such as switches, bus, line
traps, breakers, and power transformer auxiliary equip-
ment. Because the substation equipment being uprated
is generally less expensive to replace than lines, cables,
and transformers, some of the more elaborate methods
of monitoring are difficult or even impossible to justify
economically. Also, because of the large number of
switches, circuit breakers, etc., in any power system, and
the variety of designs, both the thermal models that rep-
resent the equipment and the requirements for weather
monitoring must be kept simple.
This chapter includes four sections:
Section 5.2, Summary: Equipment Types and IPF Oppor-
tunities, is a summary of terminal equipment types, their
thermal response to changes in weather and current, and
the risks associated with high current loading.
Section 5.3, Thermal Models for Terminal Equipment,
describes specific thermal models for each type of
equipment and suggests common limits on temperature.
Section 5.4, Uprating of Substation Terminal Equip-
ment, reviews dynamic thermal rating of terminal
equipment, including a discussion of the need for
weather and load data.
Section 5.5, Thermal Parameters for Terminal Equip-
ment, describes methods of determining specific ther-
mal parameters from field test, laboratory test, and
manufacturer heat-run tests.
5.2 SUMMARYEQUIPMENT TYPES AND IPF
OPPORTUNITIES
Substation terminal equipment includes a wide variety
of equipment types with varying opportunities for
increased power flow. This section provides a broad
overview of the types of equipment that might limit
power circuit thermal ratings.
5.2.1 Equipment Rating Parameters
For each type of terminal equipment, the following
issues are compared:
Primary reasons for temperature and deterioration
limits
Type of thermal model used in rating calculations
Consequences of over-temperature
Degree of thermal interaction with other equipment
Sensitivity to weather parameters
Response to short-time emergency loads
The comparisons included here are not intended to be
exhaustive, but rather to be an initial guide as to what
can be expected to result from the various methods of
increasing power flow.
Temperature and Deterioration Limits
Manufacturers of terminal equipment usually follow
ANSI or IEEE or IEC standard recommendations with
regard to maximum operating temperatures of substa-
tion terminal equipment. One clear exception is bus.
While there are manufacturing standards for strain bus
and tubular bus, temperature limits and thermal models
are not typically included in the standards.
Figure 5.1-2 Diagram showing the limiting element for
each of multiple circuits making up a complex power flow
interface. The total interface transfer limit is shown at the
top labeled Transfer (MW). Note that replacing the CT in
Circuit A yields an increase in the transfer limit from 450
to 500 MW! (Diagram courtesy of N. Dag Reppen,
Niskayuna Power Consultants, Inc.)
5-3
Increased Power Flow Guidebook Chapter 5: Substation Terminal Equipment
One obvious way to increase the rating of substation ter-
minal equipment involves the use of higher than recom-
mended equipment component temperatures, especially
when this is done for limited periods of time and when
such events occur infrequently. However, when the
exceedence of normally recommended maximum equip-
ment temperatures are to be allowed, the consequence
of such events on the life and proper function of termi-
nal equipment must be known.
Thermal Models
Thermal models for substation terminal equipment fall
into one of two categories. The first category is similar
to the power transformer top oil model (see Chapter
4). In this ambient-adjusted model, the temperature
rise above ambient for critical components of the equip-
ment (e.g., switch contact temperature), determined by
reference to the appropriate standard, by a manufac-
turer test report (if available), or by field or laboratory
measurement, is specified for a known current (typically
the rated current of the equipment). This reference
temperature rise is then adjusted for other currents
according to an equation of the form:
5.2-1
Where:
2
is the temperature rise to be calculated.
R
is the reference temperature rise.
I
2
is current for which the temperature rise is to be
calculated.
I
R
is the reference current which causes
R.
n is an exponent, generally close to 1.0.
The second category of thermal model consists of an
actual heat balance similar to that used for overhead
lines and underground cables. In this model, the temper-
ature rise is calculated with a heat balance equation of
the form:
5.2-2
Where:
I is current in amps.
R is the ac resistance of the component.
q
s
is the solar heat gain.
q
r
is the radiation heat loss.
q
c
is the convective heat loss.
With either sort of thermal model, heat storage in the
equipment can be included in order to simulate tran-
sient thermal response to changes in current flow.
Circuit breakers, CTs, and line traps are usually mod-
eled with the ambient adjusted thermal model. Strain
and tubular bus, bolted connectors are usually modeled
with the heat balance approach. Switches and line traps
can be modeled either way, but the heat balance
approach usually requires too many dimensional and
material parameters to be practical.
Determination of Equipment Thermal Parameters
As discussed in Section 5.5, the determination of other
than default thermal parameters for substation terminal
equipment is one of the most challenging parts of deter-
mining and increasing power flow through them. Unlike
power transformers, for which certified heat run data is
typically available, thermal test data from the manufac-
turer is seldom required by the utility and, if it was orig-
inally supplied, it may no longer be available for older
equipment. For newer equipment, it should be possible
to obtain documentation of a thermal design test. This
documentation will contain measurements of the tem-
perature rises above ambient of critical equipment parts
at rated current. Thermal time constants and exponents
are not typically available and must be determined by
measurement or assumed.
Estimate and Consequences of Over-temperature
Unless one chooses to be extremely conservative, the
magnitude and consequences of equipment over-tem-
perature must be evaluated when rating substation
equipment. For example, strain bus is seldom rated for
still air conditions, because this would yield extremely
low thermal ratings. But when strain bus is rated at
100
o
C on the basis of a 3-foot-per-second crosswind,
then the temperature that it might obtain under still air
conditions (the temperature rise) must be estimated,
and the consequences of occasionally attaining such a
temperature on bus strength and clearance evaluated.
Degree of Thermal Interaction with Other Equipment
Spacing of equipment in most substation designs is
driven by electrical clearance considerations. At dis-
tances sufficient to meet these electrical clearance needs,
there is little or no thermal interaction by means of con-
vection or radiation. On the other hand, substation
equipment is connected by electrical conductors that
may conduct heat as well as current. The source of such
heat may be either other electrically connected equip-
ment or the conductor itself. Fortunately, the conduc-
tion of heat between equipment by means of typical bus
conductors is unlikely to be significant. The impact of
heat from conductors that are themselves hot, however,
is a source of concern.
Sensitivity to Weather
The thermal rating of most substation equipment is sen-
sitive to air temperature, solar heating, and wind speed
and direction. Nonetheless, within the typical substa-
tion, because of the many equipment orientations and
2
2
2
n
R
R
I
I

=

2
s r c
I R q q q + = +
5-4
the degree of sheltering by other equipment and build-
ings, determination of reasonable values for solar heat-
ing, wind speed, and wind direction is very difficult. Air
temperature, on the other hand, is easily determined for
all equipment at a particular location. Therefore, with
the possible exception of strain and rigid bus work, wind
and solar effects are typically ignored.
Response to Short Time Emergency Overloads
In transmission power systems, normal power flows in
most circuits are modest (i.e., less than 30% of the cir-
cuit thermal capacity). This occurs because the system
must be capable of transmitting sudden increases in
power flows due to the sudden loss (outage) of key com-
ponents (e.g., generators and bulk transmission cir-
cuits). In order to limit the magnitude of such sudden
emergency loads, the operator may intervene within a
short period of time (e.g., 15 minutes) in order to reduce
power flow levels to normal continuous ratings or
below. In such situations, short-time emergency ratings
of substation terminal equipment may be useful.
Two factors determine the short-time emergency (STE)
rating of substation terminal equipment (and other
power equipment). These factors are the equipments
thermal time constant and its ability to withstand occa-
sional high temperature events with an acceptable
degree of deterioration. The thermal time constant is
defined as that time, after a sudden increase in electrical
current, after which the equipment temperature rise
equals 63% of that which will ultimately occur if the new
higher load continues indefinitely. An example of with-
standing occasional high temperature exposure is the
aging of free-standing current transformer insulation,
which may shorten the life but not cause short-term cat-
astrophic failure.
Maximum Multiple of Nameplate Rating
For short-time and long-time emergency ratings, the
thermal rating calculation formulas may allow for oper-
ation at many times the continuous nameplate rating,
but there may be perfectly good engineering reasons to
limit these transient rating to a multiple of the name-
plate rating. For example, this is done with power trans-
formers, which are normally limited to 200% of
nameplate regardless of the STE or long-time emer-
gency (LTE) rating calculations.
5.2.2 Thermal Rating Parameter Comparison
Tables 5.2-1 and 5.2-2 compare thermal rating parame-
ters of substation terminal equipment.
5.3 THERMAL MODELS FOR TERMINAL
EQUIPMENT
As noted in the preceding section of this chapter, there
are many types and designs of terminal equipment, and
detailed thermal test data, particularly for older equip-
ment, is unlikely to be available. As a result, simplicity is
preferred in modeling terminal equipment.
5.3.1 Bus Conductors
Bus conductors in substations come in a wide variety of
sizes and types. To keep things reasonably simple, three
Table 5.2-1 Summary of IPF Characteristics for Substation Terminal Equipment (Part I)
Substation Terminal
Equipment Type
Temperature or Temp
Rise Limits (
o
C) Thermal Models
Consequence of Over-
Temperature
Thermal Interaction with
Other Equipment
Strain Bus 75 to 125 (cont.) Heat balance
Loss of strength, sag
clearance
Possible
Rigid Bus 75 to 125 (cont.) Heat balance Loss of strength Possible
Switches
(Air Disconnects)
70/93 rise normal
105/120 rise LTE
Ambient Adjusted
Contact damage,
annealing of parts
None
Line Traps 90 to 115 rise (cont.) Ambient Adjusted
Damage to Insulation or
reduction in tensile
strength of aluminum
None
Bushings 150 conductor temp
Adjusted for top oil of PT
or OCB.
Reduction in insulation
life, overpressure, gas-
ket deterioration
Directly influenced by oil
temp in OCB or PT
CTs - Bushing 120 hot spot
or OCB.
Decrease in insulation
life
Can be directly influ-
enced by oil temp in
OCB or PT
CTs Free-standing
45 rise oil
55 to 80 winding rise.
PT model, ambient
adjusted
Decrease in insulation
life
None
Circuit Breakers
90 (metal in oil)
80 (top oil)
Ambient Adjusted
Damage to contacts,
annealing of parts
None
Current Limiting Reactors 55

or 80 rise Ambient Adjusted Damage to insulation None
5-5
types of substation bus are recognized: rigid bus, strain
bus, and jumpers:
Rigid bus is normally tubular, functionally similar to
copper or aluminum conduit or pipe, but some older
rigid bus may be square or L shaped in cross-sec-
tion.
Strain bus is under tension (thus the name strain),
and usually identical to stranded conductor used in
overhead transmission lines. It usually is stranded
aluminum wires with a steel wire core (i.e., ACSR).
Jumpers are also made from stranded transmission
conductor but are not under tension.
The three types of bus conductor are shown in Figure
5.3-1.
The thermal model for these bus types are similar to
that of an overhead line (CIGRE 1997, IEEE 1993),
consisting of a heat balance between Ohmic and solar
heat input and convective and radiation heat losses.
The steady state temperature given a constant load,
ambient temperature, and effective wind speed must be
solved by iteration so as to satisfy the following heat
balance equation:
5.3-1
Where:
I is current in amps.
R is the ac resistance at temperature T in
ohms/meter.
q
s
is the solar heat gain (W/m).
q
r
is the radiation heat loss (W/m).
q
c
is the convective heat loss (W/m).
q
cond
is heat loss/gain due to conduction (W/m).
One difference between substation bus and overhead
lines is that reflected solar heating is negligible for lines,
but not for substation bus, where the conductor are
somewhat closer to the ground.
There are other important thermal rating differences,
even for the same conductor applied as substation strain
Table 5.2-2 Summary of IPF Characteristics for Substation Terminal Equipment (Part II)
Substation Terminal
Equipment Type
Practical Sensitivity to
weather
Thermal Time Constant
(min)
Sensitivity of cont.
rating to air temp.
(% change per
o
C)
Maximum Multiple of
Nameplate
Strain Bus
Wind speed and direc-
tion, air temp, solar
heating
5 to 15
0.6% to 0.8%
10% per fps wind
None
Rigid Bus
Wind speed and direc-
tion, air temp, solar
heating
10 to 30
1.0% to 0.8%
10% per fps wind
None
Switches
(Air Disconnects)
Air temp 30
0.8% for new 53
o
C rise
1.2% for older 30
o
C rise
200%
Line Traps Air temp 15 0.2% None
Bushings
Indirectly through trans-
former or breaker oil
temperature
or OCB.
Reduction in insulation
life
Directly influenced by oil
temp in OCB or PT
CTs - Bushing Air temp 15 Same as PT Same as PT
CTs Free-standing Air temp 15
Similar to PTs with OA
cooling.
Circuit Breakers Air temp 30 1% 200%
Current Limiting Reac-
tors (Dry-type)
Air temp 15 to 30
0.8% for 55
o
C rise
0.4% for 80
o
C rise
200%
2
s r c cond
I R q q q q + = + +
Figure 5.3-1 Three types of substation bus conductor.
5-6
bus and as a phase conductor in an overhead line. These
differences include:
The decrease in electrical clearance at high tempera-
ture is less likely to be a problem for substation bus
where strain bus spans are short.
The issue of loss in strength due to annealing is less
likely to be a concern in bus applications since the
increase in tension under ice and wind load is less
than for lines.
High electrical losses in bus are not a concern
because of the short length involved.
Inspection of connectors is much simpler in a substa-
tion than in a line, which might be 50 or more miles
in length.
On the other hand, the temperature attained under high
load conditions for both strain bus and line conductors
is very sensitive to wind cooling (forced convection).
Given these observations, it seems reasonable that sub-
station bus could be rated less conservatively than
overhead lines.
Consider Drake ACSR used as both substation bus and
as the phase conductor in a line. One end of the line ter-
minates at the substation with the Drake bus conductor.
Assume that both conductors are rated at 990 A for a
conductor temperature of 100
o
C, wind speed 2 ft/sec
perpendicular to the conductor, air temperature of
40
o
C, and full sun. If the wind drops to 0 ft/sec, the con-
ductor temperature with full rated load would increase
to 130
o
C. This is acceptable in both applications.
Now consider the impact of increasing the assumed
wind speed from 2 ft/sec wind to 3 ft/sec. The risk asso-
ciated with this change in the assumed rating conditions
appears to be greater for the line than for the strain bus
in the substation. Any possible deterioration in the
physical conductor and associated hardware is easier to
spot by a single trip to the substation. A line inspection
is far more expensive, requiring more time and travel.
Any permanent increase in sag is a genuine safety con-
cern along the line, not within the substation. Any loss
in conductor strength is more likely to result in a high
tension failure of the line during the next severe ice
storm than in the shorter substation span.
Oddly enough, however, substation bus is normally
rated more conservatively than lines in terms of weather
assumptions. Thus, one simple possible approach to
increasing power flow in substation bus might be the use
of less conservative weather assumptions.
5.3.2 Switch (Air Disconnect)
ANSI standards (ANSI 1979) specify certain require-
ments for high-voltage air disconnect switches. The
standards specify the rated current (thermal rating) of
the switch and the weather conditions and equipment
temperatures under which the rating is calculated. For
example, modern switches, produced after 1971, with sil-
ver contacts, are rated for continuous operation at a
temperature rise of 53
o
C, whereas those manufactured
after 1971 are rated for continuous operation at a rise of
30
o
C. In both cases, the continuous rating is calculated
for an air temperature of 40
o
C. A typical, rather simple
switch design is shown in Figure 5.3-2.
The PJM Interconnection has published detailed rating
data (PJM 1999) for air disconnects. The conclusions
drawn in the PJM documents reflect the operating phi-
losophy of PJM and should be considered by anyone
utilizing the analysis. For example (New York Power
Pool 1995), the NY Power Pool (presently NY ISO) uti-
lizes limits of 93
o
C, 120
o
C, and 140
o
C in calculating the
rating of pre-1971 air disconnects for normal continu-
ous, long-time emergency (LTE) and short-time emer-
ge nc y ( STE) r at i ngs. The mos t re c e nt PJ M
recommendations are 93
o
C, 115
o
C, and 125
o
C for the
same ratings. The PJM standard also considers the tem-
perature of conducting material joints, switch terminals
with bolted connections, and flexible connectors.
The PJM discussion also considers annealing of copper
and aluminum component parts as a factor in high-tem-
perature limits, while the NY ISO discussion does not.
Figure 5.3-2 Typical air disconnect (switch).
5-7
Simple Dynamic Rating Switch Model
The adjustment of steady-state switch rating, I
R
, with
air temperature, T
A
, may be approximated as follows:
5.3-2
Table 5.3-1 shows the variation of steady-state switch
rating with air temperature, using this simple equation.
Notice that the variation in the rating of the newer
switches, having a higher allowable contact temperature
rise over air temperature, is less. In any event, Table
5.3-1 indicates that the switch rating can be 5% to 20%
higher on a cool day.
The following, more general equations, are given in
ANSI C37.30. They allow adjustment of manufacturer
nameplate rating for both steady-state and transient
loads. The critical contact temperature may also be
tracked with these equations as load and air tempera-
ture vary over time.
5.3-3
5.3-4
5.3-5
5.3-6
Where:
U
is the ultimate contact temperature rise.
R
is the rated contact temperature rise.
I
2
is switch current at the present time step, t
2
.
I
R
is the rated switch current.
n is an exponent, generally between 0.7 and 1.0
(default 0.8).
2
is the contact temperature rise at the present
time step, t
2
.
T
A
is the ambient temperature.
1
is the contact temperature rise at the previous
time step, t
1
.
t is the time step.
is the switch thermal time constant (default
30.0 min).
The switch rating can be determined by one of several
observable temperature rises with different limiting
temperatures. It seems likely that the switch contacts
will most often be the limiting temperature. In addition,
it must be assumed that the contacts are kept in good
condition such that there is not an appreciable increase
in contact resistance.
Thermodynamic Dynamic Rating Switch Models
As part of the development of the EPRI DTCR soft-
ware, thermodynamic models of certain switches were
developed (Coneybeer 1992) and verified through labo-
ratory testing. For example, Figures 5.3-3, 5.3-4, and
5.3-5 show a comparison of contact temperature mea-
1
2
2
40
R A
R
R
T T
I I
T

=

Table 5.3-1 Impact of Air Temperature on the Normal
Rating of Air Disconnects
Ambient
Temperature
(
o
C)
Thermal Rating
(53C rise in silver
contact temp)
%Nameplate (after
1971)
Thermal Rating
(30C rise in silver
contact temp)
%Nameplate
(before 1971)
45 95 91
40 100 100
35 105 108
30 109 115
25 113 122
20 117 129
1
2
2
40
n
R A
R
R
T T
I I
T

=

2
2
n
U R
R
I
I

=

( )( )
2 1 1
1
t
U
e

= +
2 2 A
T T = +
Figure 5.3-3 Laboratory current step-sequence for
switch tests.
Figure 5.3-4 Comparison of laboratory measurements
of contact segment temperature to IEEE/ANSI and EPRI
Dynamp thermodynamic model with adjusted
parameters.
5-8
sured in a laboratory to temperature calculated with a
thermodynamic switch model.
The thermodynamic switch model consists of modeling
three switch segments separatelythe contact segment,
the bus bar segment, and the shunt. The model is basi-
cally a heat balance much like that used to model a bare
overhead line. The detailed model had the advantage
that it adjusted ratings for wind cooling, but it had a
number of disadvantages, primarily consisting of the
requirement for detailed geometrical dimensions and
electrical and thermal parameters as illustrated by the
following list:
Outer and inner diameters of the bus segment.
Contact material and surface emissivity/absorptivity.
Switch rated current, contact temperature at rated
current, and ratio of contact resistance to that of a
new contact.
Dimensions of contacts.
Shunt material and emissivity/absorptivity.
Shunt width and thickness.
Utility advisors on the project concluded that these
requirements were onerous and impractical, given the
number of switch designs being used in large utilities. In
addition to these problems, the tests indicated that the
switch contact temperature calculation was adequately
modeled with the simpler ANSI/IEEE equations and
with previously developed utility models (Bendo et al.
1979). This is illustrated in Figure 5.3-5.
In addition, the laboratory testing of an old weathered
switch with contacts that were in poor condition showed
that the ANSI/IEEE model (or presumably the Dynamp
thermal model with default parameters) underestimated
the contact temperature, as shown in Figure 5.3-4. The
good agreement between the measured temperature and
the EPRI Dynamp thermal model was accomplished by
noting the poor condition of the switch contacts and
adjusting the parameters accordinglyhardly a practi-
cal solution to a bad switch. In reality, this switch
should have been de-rated or replaced if part of a
heavily loaded power circuit.
Field Testing of Switches
Both as part of the original series of substation terminal
tests and as part of more recent field measurements of
switch temperature in operating substations, it was con-
cluded that even in heavily loaded circuits, switch tem-
peratures rarely reach levels that allow for meaningful
measurements. The older DTCR tests concluded that the
temperature rise due to solar heating was generally
higher than the temperature rise due to electrical current.
The more recent tests utilized infrared (IR) imaging
cameras and prepared (white painted) switch surfaces. It
was found that this approach to measurement provides
a noncontact measurement of temperature with an
accuracy of within 1 to 2 C. The camera used was not
unusual, but the experience of the operator was.
The primary impediment to field testing involves the rel-
atively low current levels that most switches and other
substation equipment experience. At a current equal to
30% of the switch's thermal rating, the temperature rise
is only about 10% of the rated rise.
5.3.3 Air-core Reactor
Series-connected air-core reactors are governed by
ANSI standards (ANSI 1996, 1965). The rating is lim-
ited by the hot-spot temperature rise of the conductor in
contact with the insulation or encapsulation material.
The limiting temperature varies depending upon the
insulation material (as indicated by the temperature
index). For specific limits, refer to Table 5.3-2. No ther-
Figure 5.3-5 Comparison of laboratory test results to the
ANSI/IEEE model and the EPRI Dynamp thermodynamic
model.
Table 5.3-2 Temperature Limits for Air-Core Reactors
Insulation Tem-
perature Index
(C)
Average Winding
Rise by Resistance
(C)
Hottest-spot Winding
Temperature Rise (C)
105 55 85
130 80 110
155 100 135
180 115 160
220 140 200
5-9
mal model is specifically outlined in the applicable stan-
dards. The following simple model should reflect a
reasonable compromise between accuracy and efficiency.
5.3-7
5.3-8
5.3-9
Where:
U
is the ultimate winding hot spot temperature
rise.
R
is the rated winding hot spot temperature rise.
I
2
is winding current at the present time step, t
2.
I
R
is the rated current.
(default 0.8).
2
is the winding hot spot temperature rise at the
present time step, t
2.
1
is the winding hot spot temperature rise at the
previous time step, t
1.
t is the time step.
is the winding thermal time constant (default
5.0 min).
T
2
is the winding hot spot temperature at the
2.
T
A
5.3.4 Oil Circuit Breaker
ANSI standard (ANSI 1998) gives an expression for
allowable continuous current at different ambients. This
expression can be rearranged to give the temperature
rise as a function of the current as follows:
5.3-10
Equation 5.3-10 is for steady state. To calculate the tem-
perature during transient loading periods, it is necessary
to break the contact temperature rise over ambient into
two components with different time constants: contact
rise over oil and oil rise over ambient (ANSI 1979). This
may cause some difficulty in application, as the rated
contact rise over oil may not be available. In addition,
an expression or some guidance needs to be developed
in estimating the time constant.
The transient formulation is as follows:
5.3-11
5.3-12
5.3-13
5.3-14
5.3-15
Where:
O,U
is the ultimate oil temperature rise.
O,R
is the rated oil temperature rise.
I
2
is current at the present time step, t
2
.
I
R
is the rated current.
m is an exponent, generally between 1.5 and 2.0
(default 1.8).
O,2
is the oil temperature rise at the present time
step, t
2
.
O,1
is the oil temperature rise at the previous time
step, t
1
.
t is the time step.
O
is the oil thermal time constant.
HS,U
is the ultimate hot spot temperature rise over
oil.
HS,R
is the rated hot spot rise over oil.
(default 1.8).
HS,2
is the hot spot rise over oil at the present time
step, t
2
.
2
2
n
U R
R
I
I

=

( )( )
2 1 1
1
t
U
e

= +
2 2 A
T T = +
1.8
, C C R
R
I
I

=

Figure 5.3-6 Oil circuit breaker.
2
, ,
m
O U O R
R
I
I

=

( )( ) ,2 ,1 , ,1
1
O
t
O O O U O
e

= +
2
, ,
n
HS U HS R
R
I
I

=

( )( ) ,2 ,1 , ,1
1
W
t
HS HS HS U HS
e

= +
,2 ,2 ,2 HS A O HS
T T = + +
5-10
HS,1
is the hot spot rise over oil at the previous
time step, t
1
.
W
is the winding thermal time constant (default
5.0 min).
T
HS,2
is hot spot temperature at the present time
step, t
2
.
T
A
5.3.5 SF
6
Circuit Breaker
The equations given in (ANSI 1998) and (ANSI 1979)
also apply to SF
6
breakers. However, whereas it is neces-
sary to divide the contact temperature rise into two
components for oil circuit breakers, it should be suffi-
cient to consider only the temperature rise of the con-
tacts over ambient for SF
6
breakers.
There should be no appreciable thermal capacitance
between the contacts and the ambient air.
5.3-16
5.3-17
Where:
U
is the ultimate contact temperature rise.
R
is the rated contact temperature rise.
I
2
is breaker current at the present time step, t
2
.
I
R
is the continuous current rating of the breaker.
(default 0.8).
2
is the contact temperature rise at the present
time step, t
2
.
T
A
1
is the contact temperature rise at the previous
time step, t
1
.
t is the time step.
is the breaker contact thermal time constant
(default 5.0 min).
5.3.6 Bushings (Oil-immersed Equipment Only)
This model (ANSI 1995a) applies to capacitance graded
(condenser) bushings with oil-impregnated paper or
resin-impregnated paper. Draw-lead bushing applica-
tions are not considered, because the temperature rises
will depend upon the size of the draw lead conductor
and the amount of insulation on the draw lead.
Note that use of this model requires tested values for K
1
,
K
2
, and n. Figure 5.3-7 SF
6
circuit breaker.
2
2
n
U R
R
I
I

=

( )( )
2 1 1
1
t
A U
T e

= + +
Figure 5.3-8 Bushings.
5-11
The bushing model is as follows:
5.3-18
5.3-19
Note: Equations 5.3-18 and 5.3-19 are for the calcula-
tion of the oil temperature that the bushing is immersed
in, and are included for the sake of completeness. Equa-
tions 5.3-18 and 5.3-19 can be substituted by simply
specifying the oil temperature.
5.3-20
5.3-21
5.3-22
Where:
O,U
is the ultimate oil temperature rise.
O,R
is the rated oil temperature rise.
I
2
is current at the present time step, t
2
.
I
E,R
is the rated current of the equipment (trans-
former, OCB, etc.).
m is an exponent, generally between 0.7 and 1.0
(default 0.8).
TO,2
is the oil temperature rise at the present time
step, t
2
.
TO,1
is the oil temperature rise at the previous time
step, t
1
.
t is the time step.
O
is the oil thermal time constant.
HS,U
is the ultimate bushing hot spot temperature
rise over oil.
K
1
is constant equal to the rated bushing hot spot
rise over oil (15-32).
I
B,R
is the rated bushing current.
(default 1.8).
HS,2
is the bushing hot spot rise over oil at the
2
.
HS,1
is the bushing hot spot rise over oil at the pre-
vious time step, t
1
.
b
is the bushing thermal time constant (default
5.0 min).
T
HS,2
is bushing hot spot temperature at the present
time step, t
2
.
T
A
K
2
is a bushing-specific constant between 0.6 and
0.8.
5.3.7 Current Transformers
The rating of CTs can be complex. They are rated accord-
ing to (ANSI 1993), but unlike other substation terminal
equipment, the limits on current are a function of the tap
selection and the secondary burden as well as the CT
itself. No single set of rating factors can be specified for
all applications, even in the same utility substation.
To develop continuous ratings, the continuous thermal
current rating factor (CTRCF), defined in (ANSI 1993),
must be used. The standard does not consider LTE or
STE ratings, so these must be determined by nonstand-
ard methods, which should be different for free-standing
and for bushing-type CTs.
In general the rating of bushing-type CTs is considered
equal to that of the circuit breaker or power transformer
in which the CT is installed.
If the tap setting on a bushing CT is less than its maxi-
mum ratio, then the thermal capacity may be greater
than that of the CT when set to its full winding tap posi-
tion. The adjustment of thermal capacity for tap posi-
tion can allow operation at currents above the rating for
full tap position. For example, at the 50% tap position,
the CT rating would be 140% of its full winding position
rating.
The adjustment of a free-standing CT whose tap posi-
tion rating is I
tap
, and whose rated maximum tempera-
ture rise is
R
, the rating for air temperature (
air
), can
be obtained in much the standard ambient adjustment
method using the tap rating as a basis:
5.3-23
Noncontinuous ratings can be calculated based on the
power transformer model loading guide with 55
o
C aver-
age winding rise and OA cooling mode parameters. The
winding rise exponent of 2 is typically used to be conser-
vative.
5.3.8 Line Traps
Line traps consist of an air-core inductance coil. They
are described by reference (ANSI 1981), but the stan-
dard does not make rating adjustments terribly clear,
and there is some disagreement between sources. Fol-
lowing the method and suggestions outlined in the PJM
document on rating of line traps (PJM 1999), suitable
temperature limits are a function of the manufacturer as
( )
2
2 ,
, ,
1
1
m
E R
TO U TO R
I I R
R

+

=

+

( )( ) ,2 ,1 , ,1
1
O
t
TO TO TO U TO
e

= +
( )
, 1 2 ,
n
HS U B R
K I I =
( )( ) ,2 ,1 , ,1
1
b
t
HS HS HS U HS
e

= +
,2 2 ,2 ,2 HS A TO HS
T T K = + +
1
2
30
*
R air
tap
R
I I

=

5-12
The adjustment of line trap continuous rating for air
temperature is obtained using the usual equation form
with an exponent of 2.0 in (PJM 1999). Other sources
use 1.8.
The calculation of an STE rating can be obtained using
the temperature limits shown in Table 5.3-3 in an equa-
tion of the form similar to transient calculations with
other substation terminal equipment. For a 2-hour STE
rating with a GE Type CF, pre-1965 line trap, having a
30-minute thermal time constant, the STE rating is:
5.3-24
Because of the lack of a clear rating adjustment method
in the ANSI standard, the engineer should review the
various assumptions before doing such adjustments.
5.3.9 Other Types of Terminal Equipment
Other types of substation terminal equipment, while not
specifically described here, are derived similarly. Two
excellent articles on increased power flow for substation
terminal equipment are noted as references (Cronin
1972, Conway et al. 1979).
5.4 UPRATING OF SUBSTATION TERMINAL
EQUIPMENT
Overhead lines and underground cables are not consid-
ered in this chapter but are considered elsewhere. The
replacement or physical modification of overhead or
underground transmission lines is very expensive and
can be extremely difficult to schedule given the need for
reliable power service. Failure of lines and cables occurs
outside of the restricted access of a substation and can
result in legal and safety issues. Because of this, real-
time monitoring and dynamic rating of lines and cables
is relatively easy to justify even though the monitoring
and communications can be expensive and complex.
Similarly, power transformers (the primary cost compo-
nent of substations) and their replacement typically
involve large capital outlays and extended service out-
ages. Failure of transformers can occur in a number of
ways, and cooling equipment can be complex. As with
lines and cables, real-time monitoring of transformers
and the development of dynamic thermal rating meth-
ods are often easily justified.
Substation terminal equipment considered here fall into
one of four categories:
Conductors that connect current-carrying (and non-
current-carrying) equipment (strain bus, jumpers,
rigid tubular bus, bolted and welded connectors).
Air-insulated terminal equipment (line disconnects,
free-standing current transformers, series reactors,
and power line carrier [PLC] line traps).
Oil-insulated equipment associated with a power
transformer (load tap changers, transformer bush-
ings).
Oil circuit breakers and associated bushings.
None of the terminal equipment considered have associ-
ated forced cooling equipment (fans, circulating pumps
for oil, etc.). All have much simpler failure modes than
cables, lines, and power transformers. All are consider-
ably less expensive to replace. As a result, as the come-
dian Rodney Dangerfield might have said, they dont get
the same respect.
There are several ways in which the rating of substation
terminal equipment is unique. A large substation may
have only a few power transformers and three or four
lines connected to it, yet it may have tens or hundreds of
line disconnects, bus segments, connectors, etc. Even a
moderately short overhead line has miles of conductor
that the public can stand under or ride under, whereas
all of the substation equipment at a location is enclosed
by a fence and warning signs. Imminent failures in
power transformer windings, underground cables, and
overhead conductor splices are not directly visible or
measurable, but in many cases overheated terminal
equipment can be detected with a simple infrared scan.
As shown previously, ANSI, IEEE, and IEC standards
for substation terminal equipment usually allow the
nameplate rating of the equipment to be adjusted for air
temperatures other than 40
o
C. Numerous technical
Table 5.3-3 PJM Recommended Temperature Limits for Line Traps
Line Trap Manufacturer
Limit of Rise for Rated
Continuous Current (C)
Normal Max
Temperature (C)
LTE (>24 hrs) Max
Temperature (C)
STE (<24 hrs) Max
Temperature (C)
GE Type CF (1954-1965) 90 130 145 160
Westinghouse Type M 110 150 165 180
Trench Type L 110 150 170 190
GE Type CF (after 1965) 115 155 170 190
30
160 130
1 0.0183
134%
90
STE
I

+

= =

5-13
publications suggest that the equipment thermal rating
can be further adjusted for heat storage capacity, and
that the attainment of equipment temperatures higher
than the continuous limits is acceptable for short peri-
ods of time.
In reference (Coneybeer 1992), basic thermodynamic
methods are applied to switches, bus, and wave traps.
The experiments (sponsored by EPRI) and resulting
thermal algorithms account for solar heating and forced
convection (wind) cooling. In these models, the terminal
substation equipment is modeled by multiple, thermo-
dynamically coupled components. While shown to be
accurate, the dynamic rating algorithms based on this
work required a great deal of detailed weather data, and
equipment parameters and dimensions not readily avail-
able to the utility engineer.
Choosing practical thermal models for the various
types, sizes, and designs of substation terminal equip-
ment is a matter of maximizing accuracy while minimiz-
ing complexity. As discussed in the following, the
complexity of thermal models and monitoring methods
must be balanced against the cost and complexity of
implementation, the practicality of maintenance proce-
dures, and the consequences of equipment failure.
Increasing the thermal rating of terminal equipment
can be accomplished by one or more of the following
methods:
1. Accepting increased deterioration rates by using
higher equipment temperature limits.
2. Using actual equipment temperature rise data from
manufacturer or field tests rather than conservative
estimates.
3. Adjusting the rating for actual weather conditions
(e.g., air temperature) and precontingency circuit
loading.
Unless detailed experimental data is available, the first
of these methods can result in unexpected equipment
failures. The use of manufacturer or field test data is dis-
cussed in Section 5.5.1. Adjusting ratings by monitoring
weather conditions and circuit load is discussed in Sec-
tion 5.4.1.
5.4.1 Monitoring and Communications
Communications is an essential part of dynamic ther-
mal monitoring and rating of any power circuit compo-
nents. Dynamic rating of overhead lines may require
communication of measured data from multiple remote
locations along the line route to a nearby substation,
where the data is collated and communicated to an
operations center by means of RTU channels. This pro-
cess can be complex and require frequent maintenance
visits to unprotected sites.
Dynamic rating of power transformers and terminal
substation equipment is much simpler, since any equip-
ment or weather monitoring equipment is kept within
the secure boundaries of the substation, and the com-
munications link to the utility operations center is near
at hand.
Given the number of switches and other terminal sub-
station equipment, the use of equipment monitors (e.g.,
a switch contact temperature monitor) is impractical
and almost certainly uneconomical. On the other hand,
a single weather station located in or near the substation
is probably sufficient to dynamically rate all equipment
in the station.
5.4.2 Maintenance and Inspection Procedures
An initial inspection and periodic inspection visits are
crucial to reliable operation of dynamically rated termi-
nal substation equipment, since it is not economic to
monitor the equipment in real-time. In contrast to over-
head lines, the inspection of most terminal equipment
can be performed quickly and easily with infrared imag-
ing equipment and a trip to the single substation loca-
tion. Clearly, algorithms for the dynamic rating assume
that the equipment is operating in excellent condition.
It may be very difficult to detect imminent failures of
overhead lines (particularly full tension splices) or
underground cables, but thermal problems in substation
terminal equipment can usually be spotted before an
unexpected outage can occur.
5.4.3 Reliability and Consequences of Failure
Substations are designed to be reliable with alternate
configurations available if certain equipment should
fail. Thus the consequences of failure of a single substa-
tion component may be less than for a critical line or
cable.
Failures may occur as the result of metallic deteriora-
tion (e.g., switch contact plating), annealing (e.g., strain
bus), or insulation aging (e.g., wave traps or free-stand-
ing CTs). The mechanism of failure depends on the type
of terminal equipment.
In any event, the consequence of failure may be less for
substation terminal equipment. Overhead lines and
underground cables are placed in corridors that are not
secured against public access. If either fails, the public
or property may be harmed. Substation equipment is
5-14
enclosed by fencing designed to limit access, and neither
the public nor nonutility property is likely to be dam-
aged in the event of a failure.
On the basis of this observation, certain dynamic rating
calculation methods that yield higher thermal ratings in
exchange for an increased (but low) probability of tem-
perature limit exceedence may be justified. Such
approaches can seldom be justified for overhead lines
where the public safety may be directly involved.
5.5 THERMAL PARAMETERS FOR TERMINAL
EQUIPMENT
Given the various thermal models for terminal equip-
ment, the calculation of thermal ratings depends on
having certain thermal parameters. There are three basic
methods by which these thermal parameters can be
found. In order of preference the methods are:
1. The manufacturer provides laboratory test data for
the device, including transient behavior.
2. Laboratory tests or field measurements are per-
formed to obtain thermal parameters.
3. Typical thermal parameters are selected from the
technical literature or from the appropriate stan-
dards.
5.5.1 Manufacturer Test Report Data
Two examples of manufacturer test reports are included
here. The first involves temperature measurements at
rated load for an SF
6
circuit breaker. The second is for
an air disconnect switch.
With the SF
6
circuit breaker, the design margins for the
various components vary somewhat, but are surpris-
ingly large for most parts. For example, the 44
o
C tem-
perature rise of the main contact is the highest measured
for the various CB components, yet it is considerably
less than the default rise of 65
o
C. The current in the cir-
cuit breaker would need to exceed the rated current by
more than 20% before the main contact temperature rise
reaches 65
o
C, but at this current level the bushing termi-
nal would exceed its rise limit of 50
o
C. Considering all
the measured temperature rises for all the circuit
breaker components, the equipment could be operated
at about 15% over nameplate without exceeding the nor-
mal ANSI limits.
For the air disconnect, the maximum temperature rise
occurs for TC #1. Assuming an allowable temperature
rise of 53
o
C, the measured rise of 42.5
o
C would allow for
a switch load above nameplate of approximately 15%.
5.6 CONCLUSIONS AND SUMMARY
In the preceding discussion of increasing power flow for
substation terminal equipment, certain methods of cal-
culation are presented that allow safe, reliable operation
of terminal equipment above its nameplate rating.
Thus a 1200 A switch may be operated at more than
1200 A by considering actual air temperature (rather
than 40
o
C), the manufacturers test data for temperature
rise at rated load (rather than default ANSI values), and
possible operation at higher than usual temperature lim-
its. In addition, for emergency ratings, the limited dura-
tion emergency rating may be higher if the heat storage
capacity of the switch is considered.
Figure 5.5-1 SF
6
circuit breaker with measuring locations
for laboratory tests.
Table 5.5-1 Steady-State Temperature Measurements for
SF
6
Breaker
Location of TC
Measurement
Measured
Temperature
Rise (
o
C)
Specified
Maximum
Temperature
Rise (
o
C)
1 Conductor for test 41 -
2 Bushing terminal 38 50
3 Bushing conductor 42 -
4 Conductor junction 41 65
5 Conductor junction 40 65
6 Finger contact 42 65
7 SF
6
gas 29 -
8 Enclosure 22 70
9 Main contact 44 65
10 - Conductor junction 40 65
11 - Conductor junction 41 65
12 - Bushing conductor 41 -
13 - Bushing terminal 38 50
14 - Conductor for test 40 -
Ambient temp 32
Loading duration 12 hours
5-15
If air temperature at the substation and the electrical
loading of the equipment are reported to SCADA/EMS
in real-time, substation terminal equipment can be
dynamically rated. This is particularly useful in increas-
ing the circuit rating when overhead lines, power trans-
formers, or underground cables, in series with the
terminal equipment, are also dynamically rated.
Field verification of substation equipment thermal
behavior is particularly important since the basic rating
of the equipment and all methods of increasing the rat-
ing depend upon the equipment being in excellent con-
dition. Field measurements made with infrared imaging
cameras are an excellent way to affirm the thermal con-
dition of substation equipment. Not only is it possible
to use infrared imaging cameras to spot high tempera-
tures and detect damaged equipment, but if small areas
of the surface of bus, switches, etc. are prepared by
painting, a high-quality camera in the hands of an expe-
rienced operator can be used to measure temperatures
within a few degrees centigrade. Under high current
load conditions, equipment temperature measurement
can serve to verify the thermal models and parameter
assumptions.
In summary, it is possible to operate substation terminal
equipment at current levels exceeding nameplate by
5% to 15% in most cases without reducing reliability,
but the condition of the equipment must be verified by
periodic inspections.
REFERENCES
ANSI. 1965. Appendix to C57.99-1965. Application
Guide for Loading Dry-type and Oil-Immersed Cur-
rent-Limiting Reactors.
ANSI. 1979. C37.37-1979. Loading Guide for AC High-
Voltage Air Switches (in excess of 1000 volts).
ANSI. 1981.C93.3-1981. Requirements for Power-Line
Carrier Line Traps.
ANSI. 1988. C37.010-1979. Application Guide for AC
High Voltage Circuit Breakers Rated on a Symmetrical
Current Basis. Reaffirmed 1988.
ANSI. 1993. C57.13-1993. Requirements for Instru-
ment Transformers.
ANSI. 1995a. C57.19.100-1995. Guide for the Applica-
tion of Power Apparatus Bushings.
ANSI. 1995b. C57.91-1995. Annex B. Effect of Load
Transformers Above Nameplate Rating on Bushings,
Tap Changers, and Auxiliary Components.
Table 5.6-1 Laboratory test data for switch at rated loading.
5-16
ANSI. 1995c. C57.131-1995. IEEE Standard Require-
ments for Load Tap Changers. March.
ANSI. 1996. C57.16-1996. Requirement, Technology,
and Test Code for Current-Limiting Reactors.
Bendo, I. S. et al. 1979. Loading of Substation Electri-
cal Equipment with Emphasis on Thermal Capability,
Part II Application. IEEE Transactions. PAS-98.
No.4. July/August. pp. 1403-1419.
CIGRE. 1997. Working Group 12-22. Thermal State of
Overhead Line Conductors. Electra. No. 121. pp.51-67.
CIGRE. 2000. Working Group 22.12. Description of
State of the Art Methods to Determine Thermal Rating
of Lines in Real-Time and Their Application in Opti-
mising Power Flow. CIGRE 2000. Paper 22-304.
CIGRE. 2002. Study Committee 23. Dynamic Load-
ing of Transmission EquipmentAn Overview. Electra.
No. 202. June.
Coneybeer, R. T. 1992. Transient Thermal Models for
Substation Transmission Components. Master's The-
sis, School of Mechanical Engineering, Georgia Insti-
tute of Technology.
Conway, B. J. et al. 1979. Loading of Substation Elec-
trical Equipment with Emphasis on Thermal Capabil-
ity. IEEE Transactions on Power Apparatus and
Systems. Vol. PAS-98. No. 4. July/August. pp. 1394
1419.
Cronin, J. 1972. Rate Substation Equipment for Short-
time Overloads. Electrical World Magazine. April 15.
Douglass, D. A. and Edris, A. 1999. Field Studies of
Douglass, D. A., Edris, A. et al. 2002. Dynamic Load-
ingLessons Learned. CIGRE Report. September
2002, Paris, France.
IEEE. 1993. Standard 738-93. IEEE Standard for Cal-
culation of Bare Overhead Conductor Temperatures.
New York Power Pool. 1982. New York Power Pool
Task Force on Tie Line Ratings. Final Report. Final
Issue. June.
New York Power Pool. 1995. Tie-Line Ratings Task
Force. Final Report on Tie-Line Ratings. November.
PJM. 1999. PJM Interconnection Planning & Engineer-
ing Committee. Air Disconnect Switch Ratings. Feb-
ruary.
PJM. 1999. PJM Interconnection Planning & Engineer-
ing Committee. Guide for Determination of Line Trap
Normal and Emergency Ratings. August.
Seppa, T. O. et al. 1998. Use of On-Line Tension Mon-
itoring Systems for Real Time Ratings, Ice Loads and
Other Environmental Effects. CIGRE Report 102-22.
September. Paris, France.
6-1
CHAPTER 6 Dynamic Thermal Ratings
Monitors and Calculation Methods
6.1 INTRODUCTION
The opening of the transmission system to independent power generators and the reduc-
tion in traditional regulation has led many utilities, both in the United States and around
the world, to employ methods where transmission lines and equipment can be operated
reliably at higher loadings. Since the mid-1980s, considerable attention has been paid to
increasing the power flow of overhead lines, power transformers, underground cables, and
substation terminal equipment by means of monitoring weather and the equipment ther-
mal state and by developing more accurate thermal models. The resulting dynamic ther-
mal rating techniques have typically yielded increases of 5 to 15% in capacity.
Chapter 6 provides an overview of dynamic thermal rating methods. Each of the other
chapters in this book includes a section dealing with the dynamic rating of each type of
equipment (lines, cables, power transformers, and substation terminal equipment). The
goal of this chapter is to present a balanced overall view of when dynamic rating methods
are appropriate, how they are best implemented in a practical operational application,
and how such methods can be applied to complex interconnections consisting of multiple
circuits and many circuit elements. Though the EPRI DTCR software is used to illustrate
practical methods of implementation, the observations and technical insights are gener-
ally applicable.
Section 6.2, Issues Related to Dynamic Thermal Rating Methods, discusses concerns
related to dynamic ratings, including where calculations should be performed, the
impact of ratings on engineering, planning and operations functions, and the connec-
tion between ratings and increased utilization.
Section 6.3 Power Equipment Condition Assessment and Real-Time Monitors, outlines
the need for inspections and/or real-time monitors and the problems that may arise
without them.
Section 6.4, Dynamic Thermal Rating Models for Power Equipment, provides an over-
view on models for overhead lines, transformers, underground cables, and substation
terminal equipment.
Section 6.5, EPRIs DTCR Technology, describes the use of DTCR software.
Section 6.6, Operating with Dynamic Thermal Ratings, identifies operating issues
related to dynamic thermal ratings.
Section 6.7, Field Studies of Dynamic Ratings, describes field studies of dynamic ratings
used for overhead lines, transformers, underground cables, substation terminal equip-
ment, and power circuits.
Section 6.8, Conclusions, summarizes a number of observations concerning dynamic
ratings.
Chapter 6: Dynamic Thermal Ratings Monitors and Calculation Methods Increased Power Flow Guidebook
6-2
6.2 ISSUES RELATED TO DYNAMIC THERMAL
RATING METHODS
Utilities around the world are undergoing a major
transformation, seeking to increase the utilization of
existing power equipment within the electric transmis-
sion network while attempting to increase, or at least
not decrease, service reliability. Traditionally, power util-
ities operated in a tightly regulated and predictable busi-
ness environment where capital investments in
generation and transmission were carefully planned
many years in advance and return on investment was
guaranteed. In this environment, there was little reason
to load power equipment to levels close to thermal lim-
its, and there was little need for exact thermal models.
More recently, in many countries, power utilities are
moving to an open access business environment in
which transmission rights are not assured to the
owner/investor in generation. Those who own and oper-
ate the transmission system must respond to the trans-
mission capacity needs of new generators with little or
no ability to make long-term plans, yet anticipating only
moderate and regulated return on investment. At the
same time, environmental hurdles and legal battles limit
the construction of new transmission media (overhead
lines and underground cables) and inhibit the placement
and construction of new substations. Thus significant
transmission network additions and reinforcements
have been at a virtual standstill for over 25 years while
load growth has continued.
As a result of the difficulties in financing, planning, and
gaining approval for new facilities, the normal and post-
conti ngency l oadi ng of exi sti ng equi pment has
increased. This has occurred both system-wide and
(with the addition of new generation in unplanned but
commercially or environmentally attractive locations)
on specific circuits. In response, dynamic loading and
real-time monitoring of power equipment has become
an important tool in attempting to maintain system reli-
ability while allowing increased power flows. The
improved thermal models of power equipment, devel-
oped in pursuit of higher utilization levels, can improve
our understanding of high-temperature operation.
In the mid-1980s, EPRI initiated research projects
involving the dynamic rating of both overhead lines and
underground cables. The overhead line project was
referred to as Dynamp. It produced a thermal rating
model for bare overhead conductor that gives results
similar to both the IEEE and CIGRE thermal models.
In 1993, EPRI initiated research into the real-time ther-
mal monitoring of transmission circuits. As described
in (Douglass and Edris 1996), certain existing thermal
models for underground cable, overhead lines, power
transformers, and substation equipment such as line
traps, circuit breakers, bus, switches, and current trans-
formers were included in an integrated software model
capable of calculating the dynamic thermal rating of
transmission circuits consisting of one or more such ele-
ments. The resulting software was tested and improved
as a result of an extensive series of field tests. The result
of the initial series of field tests, primarily on overhead
lines, was summarized in (Douglass and Edris 1999).
During the same period, a number of real-time tempera-
ture and condition monitoring devices have been devel-
oped and refined. These include the EPRI Sagometer,
the Valley Groups CAT-1 line tension monitor, many
types of oil monitoring devices for power transformers,
and optical time-domain reflectometry instruments to
measure temperature along underground cables. In
addition, low power communication methods have
matured, simplifying the process of real-time monitor-
ing for lines and cables that may extend over many
miles. Computerized database and storage techniques
used by utilities have also evolved until it has become
relatively easy to obtain real-time data and to store
equipment thermal parameters in utility SCADA and
maintenance systems.
This section explores a number of key issues related to
the use of dynamic ratings.
6.2.1 Where Should Dynamic Thermal Circuit
Rating Calculations Be Performed?
In early field tests, the dynamic thermal rating software,
developed by EPRI and others, was installed on com-
puters placed in the substation environment, close to the
power equipment being monitored. This generally
yielded unsatisfactory and unreliable results. Problems
included hardware thefts, travel to remote locations to
investigate problems, and unreliable communications. In
consequence, the software was modified in its most
recent version to obtain real-time input from, and send
calculated output to, volatile ASCII files on the PC.
These files are read and written by simple SCADA-
based software programs.
In this newer arrangement, real-time monitors are inter-
faced with the SCADA system, not with the dynamic
loading computer. Calculated dynamic thermal ratings,
based on real-time data obtained through SCADA, are
read back into the SCADA database from which they
can be displayed according to the needs of each utility
or utility function. The advantage of this arrangement is
primarily one of simplicity and reliability. Other benefits
6-3
Increased Power Flow Guidebook Chapter 6: Dynamic Thermal Ratings Monitors and Calculation Methods
include centralizing the DTCR application onto one
computer, and having this computer near computer
maintenance personnel.
Communication between remote monitors and the util-
ity SCADA database is handled by separate commercial
software and hardware that has nothing to do with the
dynamic rating calculation software. Utilities may have
existing interface methods, in use for other applications,
and may choose the most appropriate interface based
on maintenance and ease of use criteria.
This arrangement also allows the utility operations peo-
ple to display or utilize the calculation results in the
fashion they prefer, rather than being forced to use an
embedded display. The information may be dynamically
linked to transmission/generation scheduling or genera-
tion shedding systems. The calculated and measured
information is available for data storage by a compo-
nent of an Energy Management System (EMS) making
historical information available to dispatchers.
The advantage of this arrangement is primarily one of
simplicity and reliability. The communication between
remote monitors and the utility SCADA database is
handled by separate commercial software and hardware
that has nothing to do with the dynamic rating calcula-
tion software. This arrangement also allows the utility
operations people to display the calculation results in
the fashion they prefer rather than being forced to use
the output screen provided.
6.2.2 CostsCapital and Otherwise
Traditionally, in regulated power utilities, the engineer-
ing, planning, and operations functions are somewhat
segregated. Where dynamic loading (rating) methods
are implemented, it is likely to have an impact on all
three of these areas. This can make the implementation
and use of dynamic rating methods challenging.
Figure 6.2-1 shows a typical statistical distribution of
line current and dynamic ratings for a circuit consisting
of an overhead line. The figure applies equally well to
normal loadings and normal dynamic ratings or to post-
contingency loadings and long-time emergency ratings.
With reference to this figure, consider how such data
may be utilized differently within the utility:
Unless dynamic rating methods are employed, the
system operator does not have to take action for any
of the load levels shown since all are below the static
rating. If the line is dynamically rated, the operator
may have to take action very occasionally when the
line current exceeds the dynamic rating. If the average
line current increases over time, the operator will
need to take action more frequently.
Unless dynamic ratings are implemented, the system
planner would need to upgrade the capacity of this
circuit before the line current begins to exceed the
static line rating. If the line is dynamically rated, the
necessary upgrade may be postponed since the
dynamic rating is typically higher than the static.
Unless line monitors are used, the equipment engi-
neers thermal model for the line cannot be tested
experimentally. With line monitors installed, the
detailed temperature and sag clearance response of
the line can be improved by comparison with the
field data.
The implementation of line monitors and the calcula-
tion of dynamic ratings for the line have an impact on
the interaction of the three groups. The planner may be
able to postpone capital investment in uprating the line,
and the design engineer may be able to improve the
accuracy of existing thermal models, but the operator
may need to act more frequently to reduce line current.
The improved accuracy of the design engineers thermal
line model may improve reliability by avoiding flash-
overs to trees and other electrical circuits, but the system
planner may find that a reduction in line rating leads to
difficulties in operating the system and the need for
increased capital investment in line upgrades.
In summary, dynamic rating technology may allow the
postponement of capital investment and lead to better
thermal modeling and physical reliability, but its appli-
cation may also require increased operator actions and,
if the static rating is too high, it may lead to the need for
physical line upgrades.

Figure 6.2-1 Current load versus dynamic rating
probability distributions.
6-4
6.2.3 Why Dynamic Ratings Go With Increased
Utilization
The original EPRI dynamic thermal circuit rating
project (3022-7) was proposed as part of the EPRI Flex-
ible AC Transmission System (FACTS) technology
research program (Edris 2000). FACTS is a power elec-
tronic-based technology for enhancing controllability
and increasing power transfer capability of transmission
circuits. FACTS controllers provide the system operator
with the means of rapidly controlling loads on particu-
lar circuits in order to maximize power transfer capabil-
ity of transmission corridors. The ultimate goal of
FACTS controllers is to bring the power transfer stabil-
ity limit as close as possible to the thermal rating of
transmission circuits.
Dynamic thermal ratings of power equipment vary over
time, of course. As such, while dynamic equipment rat-
ings may often be higher than manufacturers ratings,
there may be times when thermal capacity is inadequate,
and the load needs to be reduced quickly in order to
prevent damage.
One of the most attractive applications for dynamic
equipment ratings involves automatic actions after
(post) contingency occurrences. Here, after loss of a
major system component, the thermal capacity in real-
time is tested against actual loads. Utilizing the thermal
time delay, certain pre-programmed actions can be
undertaken if the temperature limits might otherwise be
exceeded and no action is taken when capacity is ade-
quate.
Given the ability to control circuit load, the knowledge
of both predicted load and dynamic ratings may be uti-
lized to warn the operator well in advance of potential
overloads. The EPRI dynamic rating software does this
by providing a parameter called Time-To-Overload
(TTO). TTO indicates to the system operator how much
time is left until equipment temperatures exceed safe
limits. In the case of most distribution station trans-
formers and other transmission circuits, the load profile
is repeated daily with only a variation in magnitude.
With this predicted load and predicted weather for over-
head lines, the TTO can be calculated and updated in
real-time. This allows the dispatcher to shed load or
redispatch generation in a precise manner.
6.3 POWER EQUIPMENT CONDITION
ASSESSMENT AND REAL-TIME
MONITORS
As part of increasing power flow, is it necessary to track
the condition of power equipment? This section dis-
cusses the need for periodic inspections and monitoring.
Two of the results of utility deregulation and growing
public opposition to new lines and substations are the
aging of the transmission system and the increased utili-
zation of existing equipment. When building new facili-
ties was easier and when the regulated economic
environment allowed a fixed return on investment,
power equipment was typically operated at lower utili-
zation levels and was easily replaced as it aged.
While the goal of increased utilization of aging equip-
ment is understandable in the present utility environ-
ment, it can lead to a reduction in system reliability
unless the equipment is carefully maintained and moni-
tored, either by frequent inspections or by real-time
instrumentation or both. Real-time instrumentation has
an additional advantage in that field data from heavily
loaded equipment can be analyzed in order to improve
our thermal models.
Dynamic ratings of lines, transformers, cables, and sub-
station terminal equipment can be calculated on the
basis of weather data alone. For example, the operating
conductor temperature and dynamic rating of an over-
head line can be calculated based on air temperature
and wind data from a nearby weather station. This is
attractive because no instruments need to be installed
on the structures or energized high voltage conductors.
The resulting instrumentation cost is low and the instru-
ments can be installed without taking an outage.
The drawback to this approach is that there is no direct
means of confirming that the dynamic ratings are cor-
rect or that the estimated conductor temperature and
sag are correct. Some simple examples of the kind of
problems that can result are:
1. Electrical clearances may be inadequate, even though
the conductor temperature does not exceed the maxi-
mum allowable design temperature, because the line
sags are not in accordance with design plan-profile
drawings.
2. Conductor temperature may exceed the calculated
conductor temperature because of broken or cor-
roded strands at certain points along the line. This
may, in turn, lead to local annealing and/or tensile
failure.
6-5
3. The conductor sag at high temperature may be
greater than that calculated in the original design
because of incorrect ruling span assumptions or
errors in the estimate of thermal elongation rates.
The first problems can be prevented by continuing phys-
ical inspections both before the dynamic rating calcula-
tion method is initiated and as long as it continues. The
second problem is difficult to detect without climbing
inspections since IR scans and video recordings may not
show such damage. The last problem, cannot be
detected by physical inspection of the line at normal
electrical load levels since the unfortunate result (i.e.,
inadequate electrical clearance) only occurs during
extremely high line loadings. Continuous sag or tension
monitoring can be used to detect such problems.
6.4 DYNAMIC THERMAL RATING MODELS
FOR POWER EQUIPMENT
A number of distinctions can be made between real-time
monitoring and the determination of dynamic thermal
equipment ratings. Equipment monitors provide the sys-
tem operator with an indication of the thermal state
(e.g., critical temperatures) of remote power equipment
but offer no guide to safe load levels. Weather monitors
provide air temperature, wind speed, wind direction,
and solar heating data. Dynamic ratings provide the sys-
tem operator, and system security assessment programs,
with an upper bound on circuit electrical loading that, if
adhered to, keeps equipment temperatures below user-
specified maximums. Monitor parameters may be in any
one of a variety of units (degrees C or pounds of line
tension). Dynamic ratings are expressed in amps or
MVA, the same units as electrical load, so the compari-
son is obvious.
Real-time equipment monitors may report equipment
temperatures (or for overhead lines, possibly sag or ten-
sion), electrical current, and/or weather data such as air
temperature, solar heating, and wind speed and direc-
tion. When communicated to the dispatcher, this data
may be useful in establishing the present thermal state
of the monitored equipment (e.g., the top oil tempera-
ture of a power transformer) or in developing a rough
idea of dynamic load capability. For example, many util-
ities may allow substation equipment to be loaded
above normal operating limits if the air temperature is
relatively low or if the critical equipment temperatures
are well below safe limits during an emergency. Analysis
of monitor data is often dependent on the operators
knowing specific equipment temperature limits, how-
ever, and is seldom useful to system planners making
capital investment decisions.
Dynamic thermal ratings are calculated with mathemat-
ical thermal models, which require equipment parame-
ters (e.g., maximum conductor temperature, resistance
per unit length), real-time equipment monitor data (e.g.,
underground cable shield temperature or sag-tension for
lines), real-time weather data (air temperature, wind
speed, solar heating, etc.), and real-time load current.
The calculation determining a safe dynamic load limit
typically consists of repeated calculation of equipment
temperature(s) for some period of time into the future to
determine the load current that just meets the safe limits
on equipment temperature. These calculations include
the present equipment temperature(s) and use predicted
weather data. The resulting maximum allowable load
current, expressed in amperes or MVA, is the dynamic
load limit (thermal rating) and may be compared
directly to the actual equipment load.
Certain types of real-time monitors (line sag or tension
monitors, transformer oil or winding monitors, and
underground cable thermocouples) are essential to the
verification of any dynamic safe loading method. They
provide verification that the dynamic rating method
works. Without such verification, incorrect or inaccu-
rate dynamic rating methods could damage power
equipment. In some cases this process of verification
may only be required during startup, after which the
real-time equipment monitors may be removed.
Therefore, real-time equipment monitors are important
to the verification of dynamic load limits but are not of
themselves a reliable guide to the safe loading of equip-
ment. Both real-time monitors and dynamic load limit
calculation methods are important to the reliable opera-
tion of transmission power equipment at high load levels.
6.4.1 Accounting for Heat Storage (Pre-load
Monitoring)
Thermal time constants of power transmission equip-
ment vary significantly. For example, Figure 6.4-1 shows
the relative time constants of typical lines and cables.
The time constant of the overhead line is about 10 min-
utes, while the buried cable time constants are greater
than 1000 minutes. These time constants are important
when considering emergency ratings (temporary appli-
cation of load beyond the normal continuous rating).
The long thermal time constant of buried cables yields
very high, short-duration, emergency ratings as com-
pared to overhead lines. Transformers are somewhere in
between.
6-6
6.4.2 Overhead Lines
For overhead lines, dynamic ratings can be calculated
based on weather monitors alone or in combination
with conductor temperature, tension or sag monitors.
Each type of monitoring has its own advantages and
disadvantages but dynamic ratings based on data from
sag or tension monitors appear to have significant
advantages over the use of weather monitoring alone.
Weather-Based Dynamic Line Ratings
Weather-based dynamic line ratings can be based on:
Air temperature alone
Air temperature and solar heating
Air temperature, solar heating, and wind speed
Air temperature, solar heating, wind speed, and wind
direction.
In all cases, the dynamic line rating is calculated with a
heat balance method and if the real-time current is mea-
sured, then the line conductor temperature can also be
calculated.
Equation 6.4-1 is the heat balance equation, consisting
of heat input, output, and storage terms. It is based on
the fact that at steady state, the heat input must equal the
heat output. In the steady state, where dT/dt is zero, this
can be solved for current, as shown in Equation 6.4-2.
6.4-1
Where:
Q
gen
= Heat input by ohmic losses, I
2
R [watts/m],
a function of current and resistance.
Q
sun
= Heat input by solar [watts/m].
This can be directly measured or calculated.
Q
rad
= Heat loss by radiation [watts/m],
a function of temperature rise, diameter, and
emissivity.
Q
conv
= Heat loss by convection [watts/m],
a function of temperature rise, diameter, and
heat transfer coefficient (wind speed).
MCp*dT/dt=
Heat storage term [watts/m]. This is zero in
the steady state.
6.4-2
The weather-based model is very accurate if the weather
stations are positioned appropriately to measure the
weather actually seen by the line conductors. Multiple
weather stations may be required if the weather is
expected to vary along the line route.
The weather-based model, which uses standard weather
instruments, is usually the simplest method of dynamic
line rating to implement. No instruments need to be
mounted on the line itself, and therefore they do not
need to survive in a high electromagnetic stress environ-
ment. The measurement of wind speed and direction by
an anemometer is not dependent on the lines electrical
loading, so the method works equally well under pre-
and post-contingency loading.
Conductor Temperature-Based Dynamic Line Ratings
This system is capable of calculating dynamic line rat-
ings based on direct conductor temperature measure-
ments in combination with the line current, air
temperature and solar heating. To calculate line ratings,
the real-time conductor temperature is converted to an
equivalent wind speed perpendicular to the line. Then
the wind speed is used in combination with the other
weather data to calculate the dynamic line rating.
The advantage of temperature-based ratings is that the
user has a direct measurement of conductor tempera-
ture. If the line rating is intended to limit the loss of con-
ductor strength at high temperature, then this direct
measurement of the primary parameter makes sense.
The disadvantage of using conductor temperature mon-
itors is that these instruments must function in the very
high electrical stress immediately around the energized
conductor on which they are mounted. The measured
conductor temperature must be communicated to a
ground station some distance away, and the conductor
temperature may or may not be a good estimate of the
average conductor temperature along the line, which is
related to the line sag and tension. Additionally, it is
possible for the process of temperature measurement to
Figure 6.4-1 Temperature response to step changes in
current load.
*
p
dT
Qgen Qsun Qrad Qconv mC
dt
+ = + +
AC_conductor
R
conv rad sun
rating
Q Q Q
I
+
=
6-7
actually interfere with the temperature being measured.
That is, if the measurement device introduces significant
mass and/or shielding in the measurement area, the tem-
perature being recorded might not be representative of
conductor temperatures a few meters away.
Sag/Tension-Based Dynamic Line Ratings
The fundamental parameter of interest on most over-
head lines is the sag or clearance at maximum allowable
conductor temperature. Thus real-time measurement of
sag or tension provides a direct measurement of the lim-
iting parameter. Also, sag-tension monitors respond to
the weather conditions along an entire line section being
monitored rather than to weather conditions at a single
point along the line. Therefore, ratings based on a single
sag-tension monitor are equivalent to several weather
stations along a line section. Of course, most lines con-
sist of multiple line sections, so that one sag-tension
monitor does not indicate the rating of the whole line.
DTCR is capable of calculating dynamic line ratings for
commercial conductor tension monitors such as the
widely used CAT-1 line tension monitor. The CAT-1
uses load cells placed in series with the insulators at a
strain structure. Air temperature and solar heating are
measured at the same structure. The monitors are linked
by radio or cellular telephone to a PC running DTCR
or to the utility SCADA/EMS system. Tension monitors
can be installed with the line in service but are normally
installed with the line de-energized.
As with the use of temperature monitors, the real-time
tension is converted to an equivalent wind speed. This is
done in two steps. First the tension is converted to an
average conductor temperature along the line section
based on field calibration data. Second, the average con-
ductor temperature (in combination with the line cur-
rent, air temperature, and solar heating) is converted to
an effective average wind speed along the line section.
The line rating is then calculated using the weather-
based heat balance algorithm. DTCR allows for a simi-
lar process based on a real-time sag monitor, but no
such commercial device presently exists.
6.4.3 Power Transformers
The dynamic rating algorithms used to rate power trans-
formers are usually based on the IEEE Top Oil model
(IEEE 1996), the IEEE Bottom Oil model (IEEE 1996)
or the IEC model (IEC 1991). Each of the models
requires slightly different transformer test data (the
IEEE Bottom Oil model requiring considerably more
detailed test data than the other two), and the more
complex Bottom Oil model does a better job modeling
short time loads.
Transformer ratings are primarily a function of air tem-
perature and electrical load. Solar heating of equipment
in unshaded areas and wind effects on transformers that
do not have forced air cooling are secondary but occa-
sionally important parameters. Since wind is less impor-
tant to power transformers, the calculation of dynamic
ratings and the prediction of ratings are simpler than for
overhead lines.
However, in comparison to overhead lines and under-
ground cable, real-time monitoring and dynamic load-
i ng of power transformers present some speci al
difficulties:
The critical transformer temperature is the winding
hot spot. Normally, it is impossible to directly mea-
sure the hottest spot winding temperature in older
transformers without a fiber optic hot spot monitor,
so there is a level of uncertainty as to the difference
between the hot spot and the average winding tem-
perature.
All of the transformer dynamic loading calculation
models require laboratory test data for hot spot rise
over top oil and oil rise over air temperature at rated
electrical load. This data may be difficult or impossi-
ble to find for a 40 year-old unit.
In addition to maximum winding current limits, the
dynamic thermal rating of power transformers may
depend on loss of insulation life, maximum oil tem-
perature, maximum hot spot temperature, or the for-
mation of bubbles in winding insulation. The present
state of insulation may be difficult to determine.
Auxiliary equipment such as bushings, conductor
leads, tap-changers, and associated protection limits
may limit transformer dynamic loading. This can be
dealt with either by replacing such limiting elements
with ones of higher capacity or by dynamically rating
the auxiliary equipment as well as the transformer
itself.
Protection settings must also be considered.
Unwanted tripping could occur in overload situa-
tions.
Older units may be in poor condition, requiring that
limits on winding temperature and insulation loss-of-
life be reduced below those of newer units.
Rating Constraints for Power Transformers
Dynamic (and static) thermal ratings for overhead lines
and underground cables are typically calculated for a
single rating constraint (conductor temperature). How-
ever, the dynamic rating of power transformers is typi-
cally determined by simultaneous constraints on
maximum winding hot spot temperature, maximum
bulk top oil temperature, maximum % loss of insulation
6-8
life per loading event, and avoidance of gas bubble for-
mation in the winding insulation (Figure 6.4-2). Thus
while the thermal model itself may not be more compli-
cated than that of overhead lines or underground cables,
the multiple constraints often complicate the rating cal-
culation.
Also, while dynamic and static thermal ratings for over-
head lines may utilize higher temperatures for shorter
rating durations (e.g., 75C for continuous operation
and 100C for emergencies), power transformers are
typically rated using multiple rating limits as shown in
Table 6.4-1.
These limitations are used to calculate thermal ratings
with either the Top Oil or Bottom Oil thermal
models and apply to both static and dynamic rating cal-
culations. The possibility of gas bubble formation in
transformer windings is also considered but only for
short time emergency loadings where the sudden change
in electrical loading may lead to the formation of gas
bubbles in regions of high electrical stress near the
transformer windings.
Transformer Top Oil Thermal Model
In this thermal rating model, the transformer is repre-
sented with only two temperatures, the bulk oil tempera-
ture at the top of the tank and the winding hot spot
temperature. The Top Oil model is described in IEEE
Standard C57.91-1995, section 7-2 (IEEE 1996).
Given a constant measured load, the ultimate top oil
temperature rise over air is:
6.4-3
Where:
TO,U
= Ultimate top oil temperature reached
with measured load.
TO,R
= Top oil temperature at rated load.
K
u
= Ratio of measured load to rated load,
I
meas
/I
R.
R = Ratio of load loss at rated load to no-
load loss.
N = Empirical constant, typically between
0.8 and 1.
The hot spot temperature rise over the top oil tempera-
ture for a constant load is:
6.4-4
Where:
H,U
= Ultimate hot spot temperature reached
with measured load.
H,R
= Hot spot temperature at rated load.
K
u
= Ratio of measured load to rated load,
I
meas
/I
R.
m = Empirical constant, typically between
0.8 and 1.
After the winding and oil thermal time constants have
been entered, the calculation of dynamic thermal ratings
for the transformer involves tracking the hot spot wind-
ing temperature and bulk top oil temperatures in real-
time, as the electrical load and air temperature change,
and based on the present thermal state of the trans-
former, calculating its thermal response to predicted air
temperature and electrical loadings. For any type of
dynamic rating (continuous, long-time emergency, or
short-time emergency), the maximum allowable electri-
cal load is calculated such that one or more of the rating
constraints (hot spot or top oil temperatures, loss of life,
bubbles) is reached but none are exceeded.
Daily 24-Hour Dynamic Rating Example
In the power transformer dynamic rating example
shown in Figure 6.4-3, the tracking of the top oil tem-
perature over the preceding 24 hours is shown to the
left. The predicted hourly average air temperatures over
Table 6.4-1 Temperature Limits for Ratings From IEEE
C57.91-1995
Rating
Constraint
[
C
] Continuous
Long-Time
Emergency
Short-Time
Emergency
Insulated
Winding
Temperature [
C
]
120 140 180
Top Bulk Oil
Temperature [
C
]
105 110 110
Loss-of-Insulation
Life [%]
0.0133%
[per day]
0.1%
[per event]
1% [per event]
n
U
R TO U TO
R
R K
+
+
=
1
1
2
, ,
Figure 6.4-2 Dynamically rated 93 MVA power
transformer at Con Ed.
m
U R H U H
K
2
, ,
* =
6-9
the next 24 hours are shown. The predicted electrical
load is also shown. As part of the dynamic rating calcu-
lation process, the top oil temperature that results from
the predicted air temperatures and loads is calculated
for the shown pre-load.
In this example, the transformer has a nameplate rating
of 1000 A (HV winding), and the oil temperature only
reaches a peak value of a bit more than 70C about 18
hours from NOW.
Given the thermal model, the historical values of air
temperature and electrical loading for the last 24 hours,
and the predicted air temperature and load shape (not
absolute magnitude), the maximum electrical load for
the next 24 hours can be calculated. Since we are
assumed to be using real-time air temperature and elec-
trical loading, this is a dynamic rating calculation.
With reference to Figure 6.4-4, it may be seen that,
although the nameplate rating of this transformer is
Figure 6.4-3 Example of 24-hour daily dynamic rating calculation considering only top oil temperature limits.
Figure 6.4-4 Dynamic rating calculation for the next 24 hours.
6-10
1000 A (HV winding), the peak loading predicted to
occur 18 hours from NOW can reach 1200 A without
exceeding the 100
C
limit on top oil temperature. This
rather busy figure also shows that the allowable electri-
cal load is a function of the predicted load shape being
only 1120 A with a flat load.
Transformer Bottom Oil Thermal Model
The Bottom Oil model is a newer and more complex
thermal model. It is more of a physical model than the
empirical Top Oil model. Bottom oil and duct oil
temperatures, as well as the top oil temperature, repre-
sent the thermal state of the transformer fluid. Varia-
tions of viscosity and winding resistance with
temperature are included. The Bottom Oil model is
described in Annex G of IEEE Standard 57.91-1995
(IEEE 1996).
When applied to dynamic rating calculations, however,
the results of using the bottom oil model in place of the
Top Oil model, are small except during certain short-
time emergency loading events.
6.4.4 Underground Cables
The dynamic thermal rating module for cables is based
upon the 1957 paper prepared by J.H. Neher and M.H.
McGrath (Neher and McGrath 1957), and includes
numerical methods from IEC-287 and IEC-853 (IEC
1982, 1989). The cable rating algorithms in DTCR are
based on EPRI's ACE (Alternative Cable Evaluation)
program from UTWorkstation. The approach essen-
tially solves an analogous equivalent thermal circuit
(Temperature = Heat x Thermal Resistance, see Figure
6.4-5) to Ohms Law (Voltage = Current x Resistance)
for electrical circuits.
Heat is transferred away from the conductor by thermal
conduction through the thermal resistances in the insu-
lation, jacket, duct (if present), and earth out to ambient
soil (for a solid dielectric cable). The thermal resistances
can all be calculated based on knowledge of the cable
construction, details of the installation configuration,
and assumptions about the earth thermal parameters
(thermal resistivity and ambient temperature).
The equivalent thermal circuit from Figure 6.4-5 is
described in Equation 6.4-5 (thermal quantities are
shown with an overline).
Since the system voltage is generally held to be a con-
stant, the dielectric heating is constant, and the temper-
ature rise caused by dielectric heating is fixed. Then, if a
maximum conductor temperature (limited by insulation
material) is selected, the conductor current can be found
by solving Equation 6.4-5 for current as shown in Equa-
tion 6.4-6.
Figure 6.4-5 Cable equivalent thermal circuit (extruded dielectric cable).
6.4-5
6.4-6
( )
( )
( )
Ambient Earth Duct Jacket Insulation Dielectric
Earth Duct Jacket
Earth Duct Jacket Insulation Conductor
T R R R R R W
R R R R W
R R R R R W T
+ + + + + +
+ + + +
+ + + + =
Duct J acket to 2
1
Duct J acket to ath Shield/She
Duct J acket to Conductor

Earth and Cable AC
R R
T T T
I
Ambient Dielectric Conductor

=
6-11
The thermal capacitance of each cable component and
the earth determine the temperature response of the
cable to changing load conditions, both during emer-
gency rating calculations and normal load cycling. The
time constant for buried cables is typically 24 hours or
more. As a result, the change in conductor temperature
as a function of time, in response to a sudden change in
current, is quite slow as shown in Figure 6.4-6.
As a consequence of the long thermal time constants
associated with underground cable, and the stable envi-
ronment (soil temperature and thermal resistivity) in
which it is immersed, the dynamic thermal rating of
underground cable is primarily a function of the load
shape over considerable time periods. Thus the reduced
loading of an underground cable circuit over the week-
end may yield a significantly higher load capability on
Monday, whereas the same phenomenon has no impact
on the rating of an associated overhead line.
Early in the research into dynamic thermal ratings, sub-
station terminal equipment (bus, switches, CTs, line
traps, bushings, LTCs, circuit breakers, etc. ) were either
not considered or represented by only the simplest of
thermal models. Almost all of the researchers attention
was focused on overhead lines, underground cables, and
power transformers. This benign neglect of substation
terminal equipment seemed justified because the ther-
mal model data requirements were considerable, the
condition of old equipment was uncertain, and the eco-
nomic benefit of dynamically rating relatively inexpen-
sive substation terminal equipment was questionable.
The replacement or physical modification of overhead
or underground transmission lines is very expensive and
can be extremely difficult to schedule, given the need for
reliable power service. Failure of lines and cables occurs
outside of the restricted access of a substation and can
result in legal and safety issues. Because of this, real-
time monitoring and dynamic rating of lines and cables
are relatively easy to justify, as is the use of sometimes
expensive and complex monitoring systems.
Similarly, power transformers are usually the primary
cost component of substations. Their replacement typi-
cally involves large capital outlays and extended service
outages. Failure of transformers can occur in a number
of ways, and cooling equipment can be complex.
Because of this, as with lines and cables, real-time moni-
toring of transformers and the development of dynamic
thermal rating methods is also easily justified.
On the other hand, after dynamically rating lines,
cables, and transformers, the useful increase in circuit
rating is often found to be limited by substation termi-
nal equipment, and the replacement of switches, line
traps, etc. can be both difficult operationally and expen-
sive in terms of required outages. Also the dynamic ther-
mal rating of substation terminal equipment can often
be accomplished with the same real-time weather and
electrical load data used for other more capital-intensive
power equipment, and the dynamic rating of lines and
power transformers is usually highly coordinated with
the dynamic rating of terminal equipment in circuit rat-
ing for little additional complexity and investment.
Thermal Models for Substation Terminal Equipment
Substation terminal equipment falls into one of four
categories:
1. Conductors, which connect current-carrying (and
non-current-carrying) equipment (strain bus, jump-
ers, rigid tubular bus, bolted and welded connectors)
2. Air-insulated terminal equipment (line disconnects,
free-standing current transformers, series reactors,
and PLC line traps)
3. Oil-insulated equipment associated with a power
transformer (load tap changers, transformer bushings)
4. Oil circuit breakers and associated bushings.
None of these types of terminal equipment has associ-
ated forced cooling equipment (fans, circulating pumps
for oil, etc.). All have much simpler failure modes than
cables, lines, and power transformers. All are consider-
ably less expensive to replace. As a result, as Rodney
Dangerfield might have said, they dont get the same
respect.
There are several ways in which the dynamic rating of
substation terminal equipment is quite different than for
lines, cables, and transformers. Even a large substation
Figure 6.4-6 Tracking underground cable temperature.
6-12
may have only a few power transformers and three or
four lines connected to it, yet it may have tens or hun-
dreds of line disconnects, bus segments, connectors, etc.
Even a moderately short overhead line has miles of con-
ductor, which the public can stand under or ride under,
whereas all of the substation equipment at a location is
enclosed by a fence and warning signs. Imminent fail-
ures in power transformer windings, underground
cables, and overhead conductor splices are not directly
visible or measurable in most cases, whereas an over-
heated line disconnect is readily apparent with a simple
infrared scan.
ANSI, IEEE, and IEC standards (ANSI 1978, 1979a,
1979b, 1981) for substation equipment usually allow the
nameplate rating of the equipment to be adjusted for air
temperatures below 40C. Numerous technical publica-
tions suggest that the equipment thermal rating can be
further adjusted for heat storage capacity and that the
attainment of equipment temperatures higher than the
continuous limits is acceptable for short periods of time
(Cronin 1972, Conway et al. 1979, Massey 1971).
In (Coneybeer 1992), basic thermodynamic methods are
applied to switches, bus, and wave traps. The experi-
ments and resulting thermal algorithms account for
solar heating and forced convection (wind) cooling. In
these models, the terminal substation equipment is mod-
eled by multiple, thermodynamically coupled compo-
nents. While shown to be accurate, the dynamic rating
algorithms based on this work required a great deal of
detailed weather data, and equip-ment parameters and
dimensions not readily available to the utility engineer.
The choice of a practical dynamic thermal model for the
various types, sizes, and designs of substation terminal
equipment is a matter of maximizing accuracy without
requiring excessive complexity in monitoring or charac-
terizing the equipment. The complexity of dynamic
thermal models and monitoring methods must be bal-
anced against the cost and complexity of implementa-
tion, the practicality of maintenance procedures, the
consequences of equipment failure.
Monitoring and Communications for Substation Terminal
Equipment
Communications is an essential part of dynamic ther-
mal monitoring and rating of any power circuit compo-
nents. Dynamic rating of overhead lines, may require
communication of measured data from multiple remote
locations along the line route to a nearby substation
where the data is collated and communicated to an
operations center by means of RTU channels. This pro-
cess can be complex and require frequent maintenance
visits to unprotected sites.
Dynamic rating of power transformers and terminal
substation equipment is much simpler since any equip-
ment or weather monitoring equipment is kept within
the secure boundaries of the substation, and the com-
munications link to the utility operations center is near
at hand.
Given the number of switches and other terminal sub-
station equipment, the use of equipment monitors (e.g.,
a switch contact temperature monitor) is impractical
and almost certainly uneconomical. On the other hand,
a single weather station located in or near the substation
is probably sufficient to dynamically rate all equipment
in the station.
Maintenance and Inspection Procedures Substation
Terminal Equipment
An initial inspection and periodic inspection visits are
crucial to reliable operation of dynamically rated termi-
nal substation equipment since it is not economic to
monitor the equipment in real-time. In contrast to over-
head lines, the inspection of most terminal equipment
can be performed quickly and easily with infrared imag-
ing equipment and a trip to the single substation loca-
tion. Clearly, algorithms for the dynamic rating assume
that the equipment is operating in excellent condition.
It may be very difficult to detect imminent failures of
overhead lines (particularly full tension splices) or
underground cables, but thermal problems in substation
terminal equipment can usually be spotted before an
unexpected outage can occur.
Reliability and Consequences of Failure
Substations are designed to be reliable with alternate
configurations available if certain terminal equipment
should fail. Thus the consequences of failure of a single
substation component may be less than for a critical line
or cable.
Failures may occur as the result of metallic deteriora-
tion (e.g., switch contact plating), annealing (e.g., strain
bus), or insulation aging (e.g., wave traps or free-stand-
ing CTs). The mechanism of failure depends on the type
of terminal equipment.
Overhead lines and underground cables are built in cor-
ridors that are not secured against public access. If either
fails, the public or property may be harmed. Substation
equipment is enclosed by fencing designed to limit
access, so neither the public nor nonutility property is
likely to be directly involved in the event of a failure.
On the basis of these observations, dynamic rating cal-
culation methods that yield higher thermal ratings in
6-13
exchange for an increased (but low) probability of
exceeding temperature limits may be justified when
appl i ed to substati on termi nal equi pment. Such
approaches can seldom be justified for overhead lines
where the public safety may be directly involved.
6.5 EPRI'S DTCR TECHNOLOGY
In the 1990s, EPRI developed a software product called
Dynamic Thermal Circuit Rating (DTCR) for the pur-
pose of rating transmission circuits in real time. DTCR
calculates dynamic thermal ratings of power equipment
based on actual load and weather conditions that are
generally accessed through the utilitys SCADA/EMS
system. Given series or parallel combinations of equip-
ment, the software determines dynamic circuit ratings by
evaluating all equipment ratings on a circuit and finding
the most limiting ampacity for each rating scenario.
This section describes how the software is used.
6.5.1 Power Circuit Modeling
Circuits may be modeled using series-only or parallel
circuit branches. Series-only circuits are analyzed by
evaluating ratings of all circuit elements (assumed to be
connected in series) and then picking the lowest rating
of any of the elements (circuit rating). Parallel circuit
branches are modeled assuming each branch is com-
prised of one or more series elements, but the software
accounts for the effects of one of the parallel branches
being out of service and impacting the loading on the
other branches. Three types of data are entered into the
system:
Data Channels This input specifies the source of
monitored parameters whether from real-time
sources or from override files. The data may
include load, air and soil temperatures, wind speed
and direction, solar radiation, and line tension/sag.
Circuit Data This data specifies overall circuit
parameters such as emergency rating periods and the
rating that is evaluated for time-to-overload. The
data channel to use for incoming load data is selected
from one of the data channels.
Circuit Element Data This data is specific to the
equipment being modeled conductor sizes, maxi-
mum allowable temperatures, and installation condi-
tions. The data channels associated with weather
parameters or other monitored physical data are also
specified.
The main input screen for the program is shown in Fig-
ure 6.5-1. Double-clicking on a cell brings up data
screens specific to that item.
6.5.2 DTCR Output
The software tracks critical equipment temperatures
and calculates up to four dynamic thermal ratings for
each power circuit element. These are saved to the out-
put file and displayed on the screen (see Figure 6.5-2).
Typical dynamic ratings might include the following:
24 hour continuous rating
1 to 24 hour long-time emergency rating
1 to 60 minute short-time emergency rating
1 to 24 hour maintenance starting at some time in
the next 24 hours
Emergency ratings up to 300 hours can be specified for
underground cables to calculate ampacities based on the
typical repair times of these systems at some utilities.
The system also determines the limiting circuit element
for each circuit and displays that elements dynamic rat-
ing as that of the circuit. Finally, if the electrical load
current exceeds one of the calculated ratings (selected by
the user), the software estimates the time until the
equipment temperature will begin to exceed a critical
Figure 6.5-1 Setting up two circuits in DTCR.
Figure 6.5-2 DTCR output screen.
6-14
temperature for the equipment this is called Time to
Temperature Overload.
6.5.3 DTCR is a Calculation Engine for SCADA
While the software provides a simple spreadsheet-like
output on a PC monitor, it is intended that each utility
user will integrate the output within the standard
SCADA operational displays. In most setups, the
dynamic rating programs will run on a PC that is net-
worked to SCADA.
Figure 6.5-3 illustrates the methods of obtaining real-
time data and supplying output to SCADA. Data can
either be supplied to the system by SCADA itself or by
non-SCADA monitors. A file-based method of input
and output is used, so that any monitor that can write a
file to the PC hard drive or network can be used for real-
time data. SCADA can also retrieve the output calcula-
tions from the DTCR output files.
6.5.4 Modeling Complex InterfacesCalifornia
Path 15
In a recent EPRI research project, a dynamic thermal
software model was developed and tested for the Cali-
fornia Path 15 power transmission interface. This
complex dynamic thermal model incorporated real-time
rating calculations for four 230-kV lines and three sin-
gle-phase 500/230-kV autotransformers, which supply
the 230-kV lines at the Gates substation. Either the
autotransformer or one of the 230-kV lines can deter-
mine the Path rating depending upon system operating
conditions.
As shown in Figure 6.5-4, Path 15 is a complex interface
whose total pre-contingency power flow rating can
determine maximum power flow from southern to
northern California under certain system operating con-
ditions. One of the most important power flow limita-
ti ons through thi s i nterface concerns the post-
contingency power flow on the four 230-kV lines if both
of the 500-kV lines are suddenly taken out of service.
Independent of any dynamic thermal rating procedures,
if the two 500-kV lines are suddenly taken out of service,
the power flow on the 230-kV lines is reduced through a
process of automatic load shedding of interruptible
loads and a manual generation reduction at the Diablo
Canyon nuclear plant. The 230-kV lines that are most
heavily loaded after the loss of the two 500-kV lines are
those between Gates and Panoche substations. These
lines were dynamically rated in order to reduce the need
for post-contingency load reduction and/or to increase
the pre-contingency power flows on the 500-kV lines.
Figure 6.5-3 Real-time data transfer flowchart for DTCR.
6-15
The Path 15 rating (maximum pre-contingency loading)
is calculated as follows (refer to Figure 6.5-4):
1. Tension monitors and weather stations were installed
on the two 230-kV lines between Gates and Panoche
substations. Real-time air temperature from the
weather station near Gates is used for the trans-
former rating calculations. Electrical loads are avail-
able for all components, as is the list of loads that
may be shed if the 500-kV double-line outage occurs
(remedial actions or RAS).
2. The post-contingency electrical loads on the Gates-
Panoche (GP) line (singular again) and on the Gates
Power Transformer (GPT) are calculated through the
use of effectiveness factors relating the pre-contin-
gency loads on the 500-kV lines and the other moni-
tored components to the post-contingency loads.
Effectiveness factors are decimal fractions (derived
by studying numerous load flow solutions) that equal
the fraction of load shed at a particular location that
reduces the load on Path 15.
3. The calculated post-contingency loads are reduced
according to the list of nearly instantaneous remedial
actions (load reductions and generation reductions)
and the ramp load reduction undertaken manually at
Diablo Canyon. The post-contingency loads for the
GP 230-kV lines and the Gates Power Transformers
are calculated as shown in Figure 6.5-5.
Figure 6.5-4 Circuit diagram for Californias Path15.
Temperature Curve
Conductor Loading (Amps)
Loading Curve
Conductor Temp (C)
2 4 6 8 10 12 14 16
Preloading
PostLoading
Peak Loading
0
Pre-outage temp
Max Temp. 100C
Time (min)
Los Banos South DLO
Diablo Canyon
Ramp Down
Figure 6.5-5 Post-contingency load shape for Gates-Panoche
230-kV line and Gates 500/230-kV autotransformers with the
temperature response of the 230-kV line.
6-16
4. The temperature of the GP lines and the oil/winding
temperatures of the GPT are limited in the dynamic
rating calculation.
5. The loss-of-life and the possibility of gas bubble for-
mation due to the post-contingency power flow in the
GPT windings are limited.
6. The pre-contingency Path 15 power flow is adjusted
until the post-contingency power flow through both
the GPT and the GP 230-kV lines is less than their
dynamic ratings. This pre-contingency Path 15 rating
does not incorporate other system limiting condi-
tions such as voltage and stability limits.
6.5.5 Conclusions about the Dynamic Rating of
Complex Interfaces
The Path 15 dynamic thermal rating is calculated such
that, in the event of a double-line outage (DLO),
where both 500-kV lines are suddenly taken out of ser-
vice, the post-contingency power flow in the two Gates-
Panoche 230-kV lines and in the Gates 500/230-kV
transformers does not exceed the dynamic rating of the
equipment.
Path 15 is probably the most complex transmission
interface in the California power system. The successful
development and field-testing of this dynamic model of
Path 15 strongly suggests that similar dynamic models
can be developed for other less complex transmission
interfaces. It appears that any such dynamic model
development requires:
System analysis to define the power flow relation-
ships (distribution factors) and the impact of load
and generation reduction events.
Installations of line and power transformer monitors
and communication links from remote locations to a
server through SCADA.
Development of a real-time software shell, which
allows the calculation of pre-contingency and post-
contingency loads on critical interface components
and includes iterative calculation of interface rating.
Inclusion of previously developed and tested thermal
model algorithm objects such as the EPRI DTCR
library.
Inclusion of off-line power flow limits reflecting con-
cerns about voltage and stability limits.
Economic trade-off analysis.
The Path 15 dynamic modeling project serves as a pro-
totype for managing complex transmission interface
power flow with state-of-the-art technology.
6.6 OPERATING WITH DYNAMIC THERMAL
RATINGS
The implementation of dynamic thermal ratings usually
requires certain changes in operating rules and proce-
dures. The biggest change is simply incorporating the
variability of dynamic ratings into operating rules and
calculations. Historically, ratings for power equipment
have been defined for worst-case conditions and do
not vary with time and weather or system conditions.
The increased capacity attainable with dynamic rating
methods requires a major shift in thinking for opera-
tions personnel.
6.6.1 Traditional Rating Definitions
Traditionally, transmission circuits have one or more
assigned limits on power flow that the operator must
heed. These limits may be set in order to avoid excessive
voltage drop or electrical phase shift, to maintain sys-
tem reliability under certain contingencies, or to avoid
overheating equipment in the circuit. This discussion
concerns only thermal limits (i.e., ratings) specified in
order to avoid damaging power equipment by causing
excessively high temperatures in power equipment.
A good example of a fairly complex definition of tradi-
tional thermal circuit ratings is as follows:
Continuous (Normal) rating Power flows at this
level may continue indefinitely without infringing
minimum electrical clearances of bare energized con-
ductors in lines and substations, without exceeding
agreed-upon maximum (Normal) temperatures for
conductors in all types of power equipment, and, for
up to 7665 hours over the life of such equipment,
without reducing the 40-year assumed life of such
power equipment or reducing the strength of line
conductors by more than 10%.
Long-time Emergency (LTE) rating Power flows
may continue at this level for infrequent nonconsecu-
tive four-hour periods without infringing minimum
electrical clearances of bare energized conductors in
lines and substations, without exceeding agreed-upon
maximum (LTE) temperatures for conductors in all
types of power equipment, and, for up to 300 hours
over the life of such equipment, without reducing the
40-year assumed life of such power equipment or
reducing the strength of line conductors by more
than 10%.
Short-time Emergency (STE) rating Power flow
may continue at this level for very infrequent post-
contingency periods of up to 15 minutes without
infringing minimum electrical clearances of bare
energized conductors in lines and substations, with-
out exceeding agreed-upon maximum (STE) temper-
6-17
atures for conductors in all types of power
equipment, and, for up to 12.5 hours over the life of
such equipment, without reducing the 40-year
assumed life of such power equipment or reducing
the strength of line conductors by more than 10%.
The agreed-upon maximum temperatures, and loss
of life per event, are usually allowed to increase in
going from Normal to LTE to STE ratings. For
example, with an overhead line having ACSR con-
ductors, if the Normal maximum temperature is
95C, the LTE maximum might be 115C, and the
STE maximum temperature, 125C. Similarly, for
power transformers, if the loss of insulation life per
event for Normal ratings is 0.0133% (per 24 hours),
then the loss of life during an LTE rating could be
increased to 0.25%.
The shorter duration emergency ratings, heat storage
should be considered. This may be a factor is the STE
rating of lines and the STE and LTE rating of trans-
formers.
For example, an overhead transmission line with 1590
kcmil 45/7 ACSR (Lapwing) might have a Normal Sum-
mer rating of 1794 amps (100%), an LTE rating of 1924
amps (107%), and an STE rating of 2247 amps (125%),
all calculated for the same weather conditions.
In many transmission systems, the available thermal rat-
ing categories are often simpler. For example, the system
operator may be given a single rating for lines (and
transformers) with the understanding that normal loads
will be no more than 60% of the rating and that emer-
gency loads not exceed it. In this case, the duration of
the emergency rating is typically indefinite and equip-
ment damage is avoided by defining modest temperature
limits on equipment.
6.6.2 Traditional Operating Rules
Given a rating hierarchy, the transmission system opera-
tor is also normally given rules that govern his/her
actions in keeping circuit loads below these limits. The
goal of these operating rules is that the power system
should remain secure. That is, all equipment is operat-
ing below normal ratings and, contingency analysis
programs determine that, even with a major generating
station or bulk transmission circuit out of service, no
equipment exceeds its long-time emergency rating.
Should the system be found not to be secure, then the
operator must take action to restore system security.
The type of action and the speed with which it is to be
implemented depend on the severity of the equipment
overloads.
For example, given the Normal LTE and STE ratings
for a circuit, the system operator is instructed to apply
them in this fashion:
1. As long as the circuit load is below the Normal rat-
ing, no action is required.
2. If the circuit load exceeds the Normal rating but is
less than the LTE rating, the circuit load must be
reduced until it is below the Normal rating within
four hours.
3. If the circuit load exceeds the LTE rating but is less
than the STE rating, the circuit load must be reduced
until it is below the Normal rating within 15 minutes.
4. If the circuit load exceeds the STE rating, the circuit
load must be reduced until it is less than the Normal
rating immediately.
Thermal ratings for other durations can also be defined
if useful with the operating rules functioning in a similar
fashion.
In addition to this desire to keep the power system oper-
ating in a secure mode, there is considerable economic
motivation to maximize the utilization of low cost gen-
eration to meet load demands. To accomplish this,
short-term and longer-term economic dispatch calcula-
tions are performed.
6.6.3 Operating with Dynamic Ratings
When circuit flows are limited by thermal concerns
(rather than stability limits), dynamic thermal ratings
offer advantages (and challenges) compared to tradi-
tional static thermal ratings. Since dynamic thermal rat-
ings are normally higher than static ratings, the
advantages to their use are:
When dynamic normal ratings are available, the oper-
ator may determine that normal load levels that
exceed the static rating do not require action since the
load does not exceed the dynamic rating.
When dynamic emergency ratings are available, low-
probability, post-contingency loads that exceed static
emergency ratings may be less than dynamic emer-
gency ratings, thus avoiding the need for operator
intervention to reduce load.
Dynamic ratings may allow for considerable cost sav-
ings during generation dispatch calculations because
they are higher than static ratings, though this may
require accurate prediction of dynamic rating values
up to a day ahead.
6-18
A number of challenges may arise with the use of
dynamic ratings, which results in their not being more
widely integrated into power system operations. These
challenges include:
SCADA/EMS flexibility and capability
Instrumentation reliability
Availability and reliability of communication links to
SCADA
Rating variability
Engineering acceptance
The purpose of this discussion is not be exhaustive but
rather to identify obvious barriers to implementation
and suggest solutions that might overcome them.
SCADA/EMS Flexibility and Capability
SCADA/EMS systems are constantly being upgraded,
and some older systems may not possess the capability
to incorporate dynamic ratings. This limitation is much
less common today than it was 5 to 10 years ago. It is
likely that any new operational software will be capable
of updating circuit ratings if provided by dynamic rating
calculation programs.
Instrumentation Reliability
Instruments may be used to monitor power transformer
oil and winding hot spot temperatures, soil tempera-
tures, weather conditions, and tension or sag of over-
head lines. Instruments within the substation perimeter
are generally safe from vandalism and protected to some
extent from lightning and extreme weather conditions.
Instruments mounted on overhead line structures and in
underground cable manholes are more vulnerable.
Verification of proper instrument function needs to be
considered as part of any dynamic rating calculation
software. Ideally, indications of malfunction would be
reported to system operations in conjunction with
dynamic rating values to help in assessing the accuracy
of dynamic ratings.
Availability and Reliability of Communication Links to
SCADA
In many power systems, the number and capability of
communication links connecting remote substations to
SCADA may be a limitation in implementing dynamic
ratings.
For overhead lines, where multiple monitors may be
required to ensure that worst-case locations are
included, the number of real-time data values coming
into SCADA may be 10 or more. In this case, placement
of computation in a substation location might be prefer-
able to placement at a central location, since only the
line rating need be communicated to SCADA/EMS. No
similar advantage exists for substation equipment.
Communication of monitor data from remote locations
to a substation has been successfully accomplished by
means of spread spectrum radio and by Cellular Digital
Packet Data (CDPD). Solar power sources are usually
adequate for such low power transmitters and avoid all
the reliability issues inherent in supplying distribution
power in remote locations.
Rating Variability
To many system operators, one of the most annoying
aspects of dynamic ratings involves their variability. It is
difficult to establish system security if ratings are chang-
ing rapidly.
There are several ways to minimize the variability of
dynamic ratings. The simplest method is to be sure that
the rating variability is not based on monitor inaccuracy
or communication problems.
Another method of limiting variability is through a
combination of averaging and limiting the range of
dynamic rating values. Expressing dynamic ratings as
the moving average of values calculated over the last
several time intervals can reduce variability with little or
no loss of accuracy. Limiting dynamic rating values to
no more than 30% above static can be a simple method
of limiting variability.
Figure 6.6-1 illustrates how the use of a moving average
calculation can reduce the variability of dynamic ratings
for an overhead line.
Engineering Acceptance
Engineering personnel, both in asset management and
operations, typically question the accuracy of important
1500
1700
1900
2100
2300
2500
2700
0:00 2:24 4:48 7:12 9:36 12:00 14:24 16:48
Time-of-Day
R
a
t
i
n
g

&

L
o
a
d

-

a
m
p
e
r
e
s
LTE-4Hr 6 per. Mov. Avg. (LTE-4Hr)
Figure 6.6-1 Dynamic ratings with moving average to
reduce variability.
6-19
calculated equipment parameters such as ratings. Cer-
tainly, fundamentally, dynamic ratings need to be based
on the same equipment limitations as are used for static
ratings. For example, with an overhead line, conven-
tional static ratings are calculated for a maximum con-
ductor temperature. That same maximum conductor
temperature would also be used in performing dynamic
rating calculations.
One way to improve engineering acceptance of dynamic
ratings is by using the same calculation method for
dynamic ratings as is used for static ratings. For exam-
ple, the EPRI DTCR software models power transform-
ers with the same Top Oil and Bottom Oil models
that are used in the IEEE Loading Guide C57.92-1995
and in the engineering program PTLOAD. The provi-
sion of the PTLOAD software for static rating calcula-
tions has been very helpful in gaining acceptance for the
DTCR dynamic rating calculation method and eases the
set-up and results checking necessary to verify the
dynamic rating results.
6.7 FIELD STUDIES OF DYNAMIC RATINGS
Dynamic ratings have been employed to allow increased
loading of lines, power transformers, and underground
cables. More recently, substation equipment has been
included in the list of equipment for which dynamic rat-
ings may be applicable.
Real-time thermal algorithms have been developed for
overhead lines, underground cables, power transform-
ers, and substation switches, bus, and line traps. Most of
the research has centered on verifying these algorithms
with field test data.
6.7.1 Overhead Lines
The potential increase in thermal rating of overhead
lines through the use of dynamic rating methods is
larger than for other power equipment. As a result,
there have been many field studies of lines.
A variety of overhead line thermal model algorithms
have been fi el d tested. (Exampl es are the EPRI
DYNAMP algorithm [Black and Byrd 1983], the
IEEE P738 thermal model, and a slightly different
model proposed by CIGRE Study Committee 22
[CIGRE 1997].) Each algorithm is capable of tracking
the present conductor temperature based on real-time
weather station data and on line current or on sag or
tension monitor data.
A study by Schmidt (Schmidt 1997) indicates that any
difference between these algorithms and the IEEE Stan-
dard 738-1993 (IEEE 1993) is quite small, particularly
when compared to differences produced by measure-
ment methods. That is, given a certain series of chrono-
l ogi cal weather and l oad val ues, each real -ti me
algorithm will yield very similar calculated temperatures
(and consequently, dynamic ratings).
Conclusions based on these field measurements include
the following:
1. For a line section (multiple suspension spans having
nearly the same tension), limited by electrical clear-
ance, dynamic ratings based on a single sag or ten-
sion monitor are more accurate than those based on
a single weather station or temperature monitor.
2. For lines consisting of many line sections, multiple
measurement locations coupled by radio communi-
cation links are necessary for accurate dynamic rating
calculations of the whole line.
3. At low wind speeds, when the dynamic rating is low-
est and most sensitive to weather, the prediction of
wind direction and persistence is nearly impossible
due to limitations on equipment and the turbulent
nature of low speed winds.
In particular, Figure 6.7-1 illustrates the need to
monitor multiple line sections for long lines. It shows
the dynamic rating of each of four different line sec-
tions (from the same line) over a 24- hour period.
Note that the limiting line section (the section with
the lowest rating) varies during the day. Thus all
would need to be monitored to ensure a safe dynamic
rating for the whole line.
4. At current levels less than 0.5 A per kcmil, dynamic
thermal ratings determined by sag or tension moni-
tors are not accurate.
Figure 6.7-1 Variation in limiting line section for a
monitored 230-kV line.
6-20
6.7.2 Power Transformers
Three distinct dynamic loading algorithms are available
in the DTCR software to model power transformers.
These are based on the IEEE Top Oil model (IEEE
1996), the IEEE Bottom Oil model (IEEE 1996) and the
IEC model (IEC 1991). As is the case for overhead
lines, except for very short time loads, the calculated
dynamic loadings using the three models for power
transformers are relatively similar. Each requires slightly
different transformer test data (the IEEE Bottom Oil
model requiring considerably more detailed test data
than the other two). The more complex Bottom Oil
model does a better job modeling short-time loads.
In comparison to overhead lines and underground
cable, real-time monitoring and dynamic loading of
power transformers offer some special difficulties:
The critical transformer temperature is the winding
hot spot. Normally, it is impossible to directly mea-
sure the hottest spot winding temperature in older
transformers without a fiber optic hot spot monitor.
All the transformer dynamic loading calculation
models require laboratory test data for hot spot rise
over top oil, which may be difficult or impossible to
find for a 40 year-old unit.
In addition to maximum winding current limits, the
dynamic thermal rating of power transformers may
depend on loss of insulation life, maximum oil tem-
perature, maximum hot spot temperature, or the for-
mation of bubbles in winding insulation.
Auxiliary equipment such as bushings, conductor
leads, tap changers, and associated protection limits
may limit transformer dynamic loading.
Protection settings must also be considered.
Unwanted tripping could occur in overload situa-
tions.
Older units may be in poor condition requiring that
limits on winding temperature and insulation loss-of-
life be reduced below those of newer units.
6.7.3 Underground Cables
Many field tests of dynamic rating for underground
cable circuits have been undertaken. By and large this
application of dynamic rating methods is the most suc-
cessful. The dynamic rating of underground cable is
determined by the loading of the cable (which varies)
and the thermal resistivity of the surrounding soil, which
varies between locations but not very much with time.
The thermal time constants associated with under-
ground cables are much larger than those of other ele-
ments, even power transformers. As a result, short-time
emergency ratings for underground cable are normally
much higher than for other types of power equipment.
Chapter 3 in this IPF Guidebook discusses the thermal
rating of underground cable in considerable detail.
Substation terminal equipment consists of many differ-
ent types and designs of power equipment. Included in
this classification are line traps, oil circuit breakers, SF
6
circuit breakers, rigid tubular bus, line disconnects, cur-
rent transformers, bolted connectors, and insulated
bushings.
The dynamic thermal rating of substation terminal
equipment centers on practical, rather simple methods
that avoid the use of monitoring instruments and dedi-
cated communication links. In dynamically rating
equipment such as switches, bus, line traps, breakers,
and power transformer auxiliary equipment, it must be
kept in mind that replacement is generally less expensive
than for lines, cables, and transformers, and therefore
the more elaborate methods of monitoring are difficult
or even impossible to justify economically. Also,
because of the large number of switches, circuit break-
ers, etc., in any power system, and the variety of designs,
thermal models and weather monitoring must be kept
simple.
6.7.5 Power Circuits
Power systems are operated with the goal of maintain-
ing a secure system wherein normal circuit loads are less
than circuit normal ratings and post-contingency loads
(mostly for the loss of a single major system compo-
nent) are less than circuit emergency ratings. Most cir-
cuits consist of multiple series circuit elements. Figure
6.7-2 shows a simple circuit consisting of a switch and
an overhead line segment. The static rating of the switch
is 1200 A, and the line, consisting of Bunting ACSR
conductor, is rated at 1310 A at 90
C. The static circuit

rating is therefore 1200 A, limited by the switch.

Switch
LineSegment
LineCurrent
Figure 6.7-2 Circuit diagram for a simple transmission
circuit.
6-21
Using DTCR3 to model this circuit, we find that the line
segment limits the circuit for short-time emergency rat-
ings, and either element can determine the LTE rating,
depending on the wind speed (see Figure 6.7-3).
6.7.6 Communications and Monitoring
In early field tests, the DTCR software was installed on
computers placed in the substation environment. This
generally yielded unsatisfactory results, which included
hardware thefts, travel to remote locations to investigate
problems, and unreliable communications. In conse-
quence, the software was modified in its most recent ver-
sion to obtain real-time input from and send calculated
output to, volatile ASCII files on the PC. These files are
read and written by simple SCADA-based software pro-
grams.
Real-time monitors are interfaced with the SCADA sys-
tem not with the dynamic loading computer, and calcu-
lated dynamic thermal ratings are read into the SCADA
database from which they can be displayed according to
the wishes of each utility. The advantage of this arrange-
ment is primarily one of simplicity and reliability. Other
benefits include centralizing the DTCR application
onto one computer, and having this computer near com-
puter maintenance personnel.
The communication between remote monitors and the
utility SCADA database is handled by separate com-
mercial software and hardware, which has nothing to do
with the dynamic rating calculation software. Utilities
likely have existing interface methods in use for other
applications. They may choose the most appropriate
interface based on maintenance and ease of use criteria.
This arrangement also allows the utility operations peo-
ple to display the calculation results in the fashion they
prefer rather than being forced to use the output screen
provided. The information may be dynamically linked
to transmission/generation scheduling or generation
shedding systems. The calculated and measured infor-
mation is available for data storage by a component of
an Energy Management System (EMS), making histori-
cal information available to dispatchers.
The advantage of this arrangement is primarily one of
simplicity and reliability. The communication between
remote monitors and the utility SCADA database is
handled by separate commercial software and hardware,
which has nothing to do with the dynamic rating calcu-
lation software. This arrangement also allows the utility
operations people to display the calculation results in
the fashion they prefer rather than being forced to use
the output screen provided.
6.8 CONCLUSIONS
Dynamic thermal rating methods apply to all power
equipment, indeed to all power circuits. By monitoring
the thermal environment and perhaps the equipment
itself in real-time, the utilization of thermal capability
may be safely increased.
Dynamic methods are capable of yielding an increase of
5 to 15% in the effective thermal rating of most power
equipment, which can result in both more economic
operation and in postponement of equipment capital
investment in low growth areas.
Certain specific observations about dynamic ratings can
be made:
1. Real-time monitoring of power equipment may be
used in evaluating dynamic loading methods, but
monitoring alone does not predict the future conse-
quences of present loading. Dynamic loading meth-
ods allow time for remedial action.
2. Dynamic rating methods require the cooperation of
transmission planners, operators, and engineers in
order to gain maximum benefit.
3. Wind is the most crucial factor in rating overhead
lines correctly. Because of the variability of wind
along the line, multiple monitoring locations are
required, especially for long lines.
4. Power transformers are the most complex equipment
to rate dynamically since such methods involve limits
on insulation life, oil and winding temperature, gas
bubbles, and auxiliary equipment.
5. The most attractive candidates for dynamic thermal
rating are those requiring the most capital to replace.
Figure 6.7-3 Short-time and long-time emergency
dynamic thermal ratings of 1200 A switch and Bunting
line segment.
6-22
6. Dynamic thermal rating methods and the accurate
prediction of post-contingency loadings are most
useful in combination with fast-reaction load con-
trols.
7. Real-time data should be obtained from direct
SCADA links rather than by direct telephone or
radio communication with remote monitors.
8. Probabilistic ratings offer an attractive alternative to
dynamic rating methods for certain types of trans-
mission equipment, provided the thermal models and
field data are accurate.
REFERENCES
ANSI. 1978. ANSI C57.13-1978. Requirements for
Instrument Transformers.
ANSI. 1979a. ANSI C37.37-1979. Loading Guide for
AC High-Voltage Air Switches (in excess of 1000 volts).
ANSI. 1979b. ANSI C37.010-1979. Application Guide
for AC High Voltage Circuit Breakers rated on a Sym-
metrical Current Basis.
ANSI. 1981. ANSI C93.3, 1981. Requirements for
Power-Line Carrier Line Traps.
Black, W. Z. and W. R. Byrd. 1983. Real Time Ampac-
ity Model for Overhead Lines. IEEE Transactions. Vol.
PAS-102. No. 7. July. pp. 22892293.
CIGRE. 1997. CIGRE WG12-22. Thermal State of
Overhead Line Conductors. Electra. No. 121. pp. 51-67.
Coneybeer, Robert T. 1992. Transient Thermal Models
for Substation Transmission Components. Master's
Thesis. School of Mechanical Engineering, Georgia
Institute of Technology.
Conway, B. J. et al. 1979. Loading of Substation Elec-
trical Equipment with Emphasis on Thermal Capabil-
ity. IEEE Transactions on Power Apparatus and
Systems. Vol. PAS-98. No. 4. July/August. pp. 1394
1419.
Cronin, John. 1972. Rate Substation Equipment for
Short-time Overloads. Electrical World Magazine.
April 15.
Douglass, D. A. and A. Edris. 1996. Real-time Moni-
toring and Dynamic Thermal Rating of Power Trans-
mission Circuits. IEEE Transactions on Power Delivery.
Vol. 11. No. 3. July.
Douglass, D. A. and A. Edris. 1999. Field Studies of
Edris, A. 2000. FACTS Technology Development: An
Update. IEEE Power Engineering Review. Vol. 20. No.
3. March.
IEEE. 1993. IEEE Standard 738-1993. Standard for
Calculating the Current-Temperature Relationship of
Bare Overhead Conductors.
IEEE. 1996. IEEE Standard C57.91-1995. IEEE
Guide for Loading Mineral-Oil-Immersed Transform-
ers. April 25.
287. Calculation of the Continuous Current Rating of
Cables (100% Load Factor).
853-2. Calculation of the Cyclic and Emergency Cur-
rent Rating of Cables. 1st Edition.
International Electrotechnical Commission. 1991. IEC
International Standard 354. 2nd Edition, Loading
Guide for Oil-immersed Power Transformers.
Massey, D. E. et al. 1971 Determination of Discon-
necting Switch Ratings for the Pennsylania-New Jersey
Maryland Interconnection. IEEE Symposium on High
Power Testing. Portland, OR. July.
Neher, J. H. and M. H. McGrath. 1957. The Calcula-
tion of the Temperature Rise and Load Capability of
Cable Systems. Paper 57-660. AIEE Insulated Conduc-
tors Committee. June.
Schmidt, N. 1997. Comparison between IEEE and
CIGRE Ampacity Standards. IEEE PE-749-PWRD-0-
06-1997. Berlin, Germany. July.
Seppa, T. O. et al. 1998. Use of On-Line Tension Mon-
itoring Systems for Real Time Ratings, Ice Loads and
Other Environmental Effects. CIGRE Report 102-22.
September. Paris, France.
G-1
Glossary
AAAC. All Aluminum Alloy Conductor. (Chapter 2)
AAC. All Aluminum Conductor. (Chapter 2)
ACAR. Aluminum Conductor Alloy Reinforced.
(Chapter 2)
ACCR. Aluminum Conductor Composite Reinforced.
(Chapter 2)
ACSR. Aluminum Conductor Steel Reinforced.
(Chapter 2)
ACSS. Aluminum Conductor Steel Supported.
(Chapter 2)
AEIC. Association of Edison Illuminating Companies.
(Chapter 3)
Aeolian Vibration. Most common type of wind-induced
motion of conductors. The vibration is caused by the
alternate shedding of wind-induced vortices from the
top and bottom sides of the conductor. This action cre-
ates an alternating pressure unbalance, inducing the
conductor to move up and down at right angles to the
direction of the air flow. This vibration occurs at rela-
tively low wind speeds, and conductor fatigue damage is
cumulative in nature. (Chapter 2)
Aluminum Conductor Steel Supported (ACSS). Conduc-
tor design consisting of fully annealed strands of alumi-
num (1350-H0) stranded around stranded steel core.
The steel core wires may be aluminized, galvanized, or
aluminum clad, and are normally high strength, hav-
ing a tensile strength about 10% greater than standard
steel core wire. (Chapter 2)
Ampacity. The ampacity of a conductor is that maxi-
mum constant current that will meet the design, secu-
rity, and safety criteria of a particular line on which the
conductor is used. In this Guidebook, ampacity has the
same meaning as steady-state thermal rating. (Chap-
ter 2)
Annealing. Process wherein the tensile strength of cop-
per or aluminium wires is reduced at sustained high
temperatures. (Chapter 2)
ASTM. American Society for Testing and Materials.
(Chapter 2)
Bottom Oil Model. One of several methods of modeling
the winding temperature of power transformers. Tem-
perature calculation methods presented in the 1995 revi-
sion of IEEE C57.91 as Annex G. See also Top Oil
Model and IEC Model 354-1991. (Chapter 4)
CGIT. See Compressed Gas Insulated Transmission.
(Chapter 3)
Compressed Gas Insulated Transmission (CGIT) Cables.
A system of SF
6
gas and epoxy insulators used to insu-
late a hollow, rigid aluminum conductor from a tubular
aluminum enclosure. The most common applications
for CGIT lines are situations where very high ampaci-
ties are required (i.e., > 2000 A), usually to connect with
overhead lines entering a station or as a high-capacity
bus within a station. (Chapter 3)
Conductor Hardware. Noncurrent carrying devices
attached directly to the conductor. Conductor hardware
includes components such as suspension clamps (with
or without armor rods), dampers, repair sleeves and
splices, spacers and spacer dampers, shackles, pins, etc.
(Chapter 2)
Continuous (Normal) Rating. Current at this level (MVA
or amperes) may continue indefinitely without without
exceeding agreed-upon maximum (Normal) tempera-
tures for conductors or critical components in all types
of power equipment. (Chapter 6)
Core-Form Construction. Power transformer design con-
structed with windings that are in the general form of
concentric cylinders. (Chapter 4)
Glossary Increased Power Flow Guidebook
G-2
Creep Elongation. Creep constitutes an irreversible,
plastic elongation occurring in the aluminum strands of
bare overhead conductor, which occurs as a result of
tension over extended periods of time. Higher rates of
creep occur when a conductor is operated for extended
periods of time at operating temperatures in excess of
approximately 50C. (Chapter 2)
Cross-Linked Polyethylene (XLPE) Cables. Most com-
mon on modern XD cable systems with applications up
to 500 kV. The insulation is cross-linked (vulcanized),
forming long polymer chains that are joined to one
another at intermediate carbon atoms. (Chapter 3)
Deterministic Methods. Method of uprating overhead
transmission lines (without reconductoring) using tradi-
t i onal cal cul at i on met hods, such as t he EPRI
DYNAMP program, with fixed, worst-case weather
condition assumptions. (Chapter 2)
Direct Flow Design. Design of power transformer
employing forced oil cooling. The transformer internal
assembly is designed with oil manifolds that direct the
incoming cool oil to the lower part of the core and
windings. A directed flow design is normally used with
larger sizes of transformers equipped with heat exchang-
ers rather than radiators. (Chapter 4)
DTCR. See Dynamic Thermal Circuit Rating. (Chapter 3)
Dynamic Rating. Limits on the level and duration of
power carried by power equipment based on actual
weather conditions. (Chapter 2)
Dynamic Thermal Circuit Rating (DTCR). EPRI software
developed to take real-time data from off the shelf
monitoring hardware, and determine optimal ratings
(not worst-case) for the conditions at the time the rat-
ings were performed. (Chapter 3)
Emergency Rating. Conductor rating that specifies how
much current can flow through power equipment under
emergency conditions for a specified amount of time
e.g., 30 minutes. (Chapter 2)
EPR. See Ethylene-Propylene-Rubber. (Chapter 3)
Ethylene-Propylene-Rubber (EPR) Cables. Insulation
type often considered for distribution cables and trans-
mission cables up to 138 kV. The insulation is very
lossy as compared to XLPE insulation, resulting in
high dielectric losses and charging current. (Chapter 3)
Exceedance Level. Amount of time a conductor exceeds
the design temperature expressed as a percentage of
total time. (Chapter 2)
Extruded Dielectric (XD) Cables. Cables are so named
because the insulation is extruded onto the conductor
core, as compared to paper-insulated cables (HPFF or
SCFF), where the insulation is a laminar application of
paper tapes. Three types of XD cables include: cross-
linked polyethylene (XLPE) cables, ethylene-propylene-
rubber (EPR) cables, and linear low- or medium-density
polyethylene (LLPE, MDPE) cables. (Chapter 3)
Factory Heat Run. A direct measurement of the thermal
performance of a transformer at a particular bench-
mark, the nameplate rating. Two methods are used to
perform factory heat runs: the short-circuit method
and the loading-back method. (Chapter 4)
Flexible AC Transmission System (FACTS). A power
electronic-based technology for enhancing controllabil-
ity and increasing power transfer capability of transmis-
sion circuits. FACTS controllers provide the system
operator with the means of rapidly controlling loads on
particular circuits in order to maximize power transfer
capability of transmission corridors. (Chapter 6)
Fluidized Thermal Backfill (FTB). Fill used in trenches for
underground cable that helps ensure good heat transfer
away from the cable pipes. (Chapter 3)
Gapped ACSR (GTACSR). Type of high-temperature con-
ductor invented in Japan, which consists of a conven-
tional steel core surrounded by layers of trapezoidal
zirconium aluminum wires, and the gap filled with grease.
The zirconium aluminum does not anneal until reaching
temperatures in excess of 200
o
C. Through the use of spe-
cial terminations and suspension clamps and by preload-
i ng the steel core, the thermal el ongation of the
conductor is less than that of conventional ACSR, while
maintaining the full strength of a conventional ACSR
conductor under heavy ice conditions. (Chapter 2)
High-Pressure, Fluid-Filled (HPFF) Pipe-Type Cables.
One of two kinds of pipe-type cables (the other being
HPGF). HPFF cables are installed in cable pipes where
the pipe is filled with very clean, very low moisture
dielectric fluid. Older HPFF cable systems (before 1970)
typically used mineral oil for the pipe filling dielectric
fluid. HPFF cable systems installed after 1970 have used
alkyl benzene or polybutene dielectric fluid. (Chapter 3)
G-3
Increased Power Flow Guidebook Glossary
High-Pressure, Gas-Filled (HPGF) Pipe-Type Cables. One
of two kinds of pipe-type cables (the other being
HPFF). HPGF cables use pressurized dry nitrogen gas
inside the cable pipe. (Chapter 3)
High-Temperature, Low-Sag (HTLS) Conductors. Con-
ductors capable of continuous operation at tempera-
tures above 100
o
C with stable tensile strength and creep
elongation properties. Such conductors are commer-
cially available or under development. (Chapter 2)
HPFF. See High-Pressure, Fluid-Filled. (Chapter 3)
HPGF. See High-Pressure, Gas-Filled. (Chapter 3)
HTLS. See High-Temperature, Low-Sag Conductors.
(Chapter 2)
ICEA. Insulated Cable Engineers Association.
(Chapter 3)
Ice Galloping. Wind-induced conductor motion occur-
ring with both single and bundled conductors, and
requiring high winds and ice on the conductors. (Chap-
ter 2)
IEC. International Electrotechnical Commission (Chap-
ter 3) or International Engineering Consortium (Chap-
ter 2)
IEC Model 354-1991. One of several methods of model-
ing the winding temperature of power transformers. See
also Top Oil Model and Bottom Oil Model. (Chapter 4)
IEEE. The Institute of Electrical and Electronics Engi-
neers. (Chapter 3)
Invar Steel. A type of steel core wire used in transmis-
sion conductors, which has a high nickel content. It has
a 15-20% lower tensile strength than conventional gal-
vanized steel wire and a much lower coefficient of ther-
mal expansion than regular high-strength steel wire.
(Chapter 2)
Knee-point Temperature. The conductor temperature
above which the aluminium strands of an ACSR con-
ductor have no tension or go into compression. (Chap-
ter 2)
Linear Low- or Medium-Density Polyethylene (LLPE,
MDPE) Cables. Insulation type less common for new
installations, although there are several installations,
predominantly in France. As compared to XLPE insula-
tion, LLPE and MDPE were first used at the higher
voltage levels because the extrudate could be raised to
higher temperatures without forming cross-linking
agents present. (Chapter 3)
LLPE. See Linear Low-Density Polyethylene. (Chapter 3)
Load Losses. Losses generated by transformers that
vary wi th l oad current but not wi th exci tati on.
(Chapter 4)
Long-Term Emergency (LTE) Rating. A rating where the
thermal heat storage capacity of the equipment does not
greatly impact the rating. For power transformers, the
bulk oil time constant makes this definition a bit unclear
given the cyclic variation in load and air temperature.
Sometimes defined as a rating greater than 2 to 4 hours
in duration. (Chapter 4)
Mass Impregnated (MI) Cables. Cables sometimes used
up to 69 kV for ac systems, although they are not that
common at this voltage. These cables have paper tapes
that are impregnated with a high-viscosity dielectric
fluid. MI cables are used for ac applications, but are
more common for HVDC submarine applications where
there may be a significant change in elevation along the
cable route that would otherwise be complicated by
hydrostatic head pressures. (Chapter 3)
MDPE. Medium-Density Polyethylene. (Chapter 3)
MI. See Mass Impregnated. (Chapter 3)
NESC. National Electric Safety Code. (Chapter 2)
No-Load Losses. Losses generated by transformers that
do not vary with load current but rather vary with exci-
tation or voltage. (Chapter 4)
Normal Rating. Power equipment thermal rating that
specifies how much current may flow in the circuit on a
continuous basis. (Chapter 4)
ODAF (Directed FOA). One of four cooling configurations
used by oil-immersed power transformers. Pumps are
used to circulate the oil. Fans are used to force air over
the radiators or heat exchangers. The forced circulation
of the oil increases the convective heat transfer from the
windings to the oil. The forced air increases the convec-
tive heat transfer from the oil to the air. With ODAF,
ducts are added to direct the oil over the winding. This
forces a significant portion of the forced oil to flow
upward through the vertical winding ducts. (Chapter 4)
OFAF (Non-directed FOA). One of four cooling configu-
rations used by oil-immersed power transformers.
Pumps are used to circulate the oil. Fans are used to
Glossary Increased Power Flow Guidebook
G-4
force air over the radiators. The forced circulation of the
oil increases the convective heat transfer from the wind-
ings to the oil. The forced air increases the convective
heat transfer from the oil to the air. With OFAF, there
are no ducts to direct the oil over the winding. In gen-
eral, the bulk of the forced oil flow passes upward
between the winding and the tank, bypassing the wind-
ings. (Chapter 4)
ONAF (FA). One of four cooling configurations used by
oil-immersed power transformers. No pumps are used
to circulate the oil. Fans are used to force air over the
radiators to increase heat transfer from the bulk oil to
the surrounding air. As with ONAN, oil circulates
upward through the windings and down through the
radiators by natural thermosiphon flow. (Chapter 4)
ONAN (OA). One of four cooling configurations used by
oil-immersed power transformers. Also referred to as
self-cooled, no pumps are used to circulate oil, and no
fans are used to increase airflow over the radiators. Oil
circulates upward through the windings and down
through the radiators by natural thermosiphon flow.
(Chapter 4)
Paper Insulated Lead Covered (PILC) Cables. Cables
sometimes used up to 69 kV for ac systems, although
they are not that common at this voltage. Uprating
approaches would be somewhat similar to those of
extruded or self-contained cables. These cables have
paper tapes that are impregnated with a high-viscosity
dielectric fluid. (Chapter 3)
PILC. See Paper Insulated Lead Covered. (Chapter 3)
Probabilistic Line Rating Methods. Methods that use the
actual weather data and conditions prevailing on a con-
ductor to determine the likelihood or probability of a
certain condition occurring. Probabilistic methods
include the Absolute Method, the Exceedance Method,
the Modified Exceedance Method, and the Safety
Method. (Chapter 2)
Quasi-Dynamic (Real-Time) Ratings. Quasi-dynamic
ratings are applied by monitoring load and tempera-
tures for a period of time and then calculating what the
conductor temperature might be as a result of that load.
From this, the temperature of the cable conductor at
rated temperature can be extrapolated for rating pur-
poses. (Chapter 3)
Rated Breaking Strength (RBS). Breaking strength of a
bare overhead conductor as calculated by the methods
described in appropriate ASTM or IEC manufacturing
standards. (Chapter 2)
Ruling (Effective) Span. Hypothetical level span length
wherein the variation of tension with conductor temper-
ature is the same as in a series of suspension spans.
(Chapter 2)
Sagging Line Mitigator (SLiM). A new class of line hard-
ware that uses a shape-memory alloy actuator, activated
by increased temperature, to reduce excessive sag in con-
ductors. (Chapter 2)
Sag-Tension Calculations. Calculations performed
using numerical programs in order to determine the sag
and the tension of a conductor catenary as a function of
ice and wind loads, conductor temperature, and time.
(Chapter 2)
Self-Contained Liquid-Filled (SCLF) Cables. Cables uti-
lizing the dielectric liquid impregnated laminated paper
insulation similar to pipe-type cables, but with three sep-
arate cables installed for the three phases. The cable is
called fluid-filled because there is a hollow fluid chan-
nel in the center of the conductor that allows dielectric
liquid to move through the cable with thermal expan-
sion and contraction. Also known as self-contained oil-
filled (SCOF) or low-pressure oil-filled (LPOF). (Chap-
ter 3)
Self-Contained Oil-Filled (SCOF) Cables. Cables utiliz-
ing the dielectric liquid impregnated laminated paper
insulation similar to pipe-type cables, but with three sep-
arate cables installed for the three phases. Also known
as self-contained liquid filled (SCLF). (Chapter 3)
Shell-Form Construction. Power transformer design in
which windings are initially assembled flat with insula-
tion and cooling ducts between sections. Complete
phase assemblies are then clamped and oriented in a
vertical direction so that the plane of the individual sec-
tions are upright. (Chapter 4)
Short-Term Emergency (STE) Rating. Short-term rat-
ings are usually defined as extremely short duration rat-
ings that take advantage of the thermal capacity of the
equipment. These ratings range from 5 to 30 minutes in
duration. (Chapter 4)
SIL. See Surge Impedance Loading. (Chapter 3)
Simulated Winding Temperature Indicator (WTI). The
most common device for measuring winding tempera-
tures in power transformers. These devices simply mea-
sure the temperature of a specially calibrated heating
element that is immersed in the top bulk oil near the
tank wall. (Chapter 4)
G-5
Increased Power Flow Guidebook Glossary
SLiM. See Sagging Line Mitigator. (Chapter 2)
Static Ratings. Limits on the level and duration of
power transferred over a line based on specified worst-
case weather conditions. (Chapter 2)
Stray Flux Heating. Heating of non-current carrying
metal components by the leakage flux of the windings
and leads. The leakage flux induces eddy current in any
conducting material that it passes through. This
includes the steel clamping structure, tie rods or tie
plates, metal core bands, and the tank wall itself. Since
leakage flux varies proportionally with load current, the
stray flux heating increases roughly with the square of
winding current. For larger power transformers and
GSUs in particular, the problem of stray flux heating
can be substantial. (Chapter 4)
Subconductor Oscillation. Wind-induced conductor
motion occurring with bundled phase conductors when
wind speeds exceed a certain critical velocity. (Chapter 2)
Surge Impedance Loading (SIL) Limits. Limits involving
a greater than allowable phase shift in power frequency
from one end of a transmission system to the other. As a
result, the two ends of the system cannot remain syn-
chronous, resulting in instability and outages. This sys-
tem stability consideration is generally an issue on
overhead transmission lines that are 80-320 km (50-200
miles) in length. (Chapter 3)
TACSR. See Thermal-resistant Aluminum Conductor
Steel Reinforced. (Chapter 2)
Thermal Elongation. Metallurgical phenomenon in con-
ductors where the material increases in length in pro-
portion to an increase in temperature. (Chapter 2)
Thermal Property Analyzer (TPA). Device used for field
and laboratory measurement of thermal resistivity.
(Chapter 3)
Thermal-resistant Aluminum Conductor Steel Rein-
forced (TACSR). Conductor widely used in Japan, with a
special type of aluminum strand capable of operating at
temperatures up to 150
o
C wi thout losing tensile
strength. (Chapter 2)
Time-To-Overload (TTO). Parameter indicating to the
system operator how much time is left until equipment
temperatures exceed safe limits. (Chapter 6)
Top Oil Model. One of several methods of modeling the
winding temperature of power transformers. Tempera-
ture calculation methods presented in Clause 7 of IEEE
C57.91-1995. See also Bottom Oil Model and IEC
Model 354-1991. (Chapter 4)
Trapezoidal Wire (TW). Aluminum trapezoidal wire used
in conductors in place of round wires, which thereby
potentially increases the cross-sectional area of a round
wire conductor of the same diameter by approximately
20%. (Chapter 2)
Upgrading. Using available infrastructure to economi-
cally put in new cables. (Chapter 3)
Uprating. Improving the capacity of existing equipment.
(Chapter 3)
Video Sagometer. EPRI device, based on digital video
technology, for monitoring conductor sag in real time.
(Chapter 2)
Voltage Drop. Limit placed on power flow correspond-
ing to the maximum allowable decrease in voltage mag-
nitude. (Chapter 2)
Worst-case Weather Conditions for Line Rating Calcu-
lation. Weather conditions that yield the maximum or
near-maximum value of conductor temperature for a
given line current. (Chapters 2 and 5)
XD. See Extruded Dielectric. (Chapter 3)
XLPE. See Cross-Linked Polyethylene. (Chapter 3)
ZTACIR. ZTAL aluminium alloy conductor reinforced by
an Invar steel core. (Chapter 2)
ZTAL (Super Thermal-resistant Aluminium). An alu-
minium zirconium alloy that has stable mechanical and
electrical properties after continuous operation at tem-
peratures of up to 210
o
C. (Chapter 2)

I-1
Index
All references are to section or subsection numbers.
A
AAC
Sag-tension calculations, 2.2.3
ACCR, 2.6.7
ACSR
High-temperature sag, 2.2.3
Sag-tension models, 2.4.3
ACSS, 2.6.3
Air-core reactor, 5.3.3
Ampacity, 3.4.2
Ampacity audit, 3.6.5
Calculating ampacity, 3.4.5
Effect of various parameters, 3.4.6
Increasing ampacity (of underground cables), 3.6
Annealing, 2.4.2
B
Bus conductors, 5.3.1
Bushings, 5.3.6
C
Case studies
Overhead, 2.8
Underground, 3.9
CGIT (Compressed Gas-Insulated Transmission) cables,
3.2.4
Compressed gas-insulated transmission cables, 3.2.4
Conductor blowout, 2.2.4
Conductor hardware, 2.4.14
Constraints on uprating overhead transmission lines, 2.2
Constraints on uprating underground cables, 3.5
Continuous (normal) rating, 6.6.1
Creep elongation, 2.4.12
Cross-linked polyethylene, 3.2.2
Current transformers, 5.3.7
D
DTCR (Dynamic Thermal Circuit Rating), 2.7.5, 3.8.2,
3.8.3, 6.5
Dynamic monitoring, 2.7
Real-time monitors, 2.7.5
Video sagometer, 2.7.5
Dynamic ratings (overhead transmission), 2.3.1, 2.7.2
Advantages, 2.7.3, 3.8.3
Calculations, 2.7.6
Disadvantages, 2.7.4
Underground cables, 3.8
Dynamic ratings (underground cables), 3.8, 6.4.4
Benefits, 3.8.3
Dynamic Thermal Circuit Ratings, 3.8.2
Monitoring, 3.8.4
Quasi-dynamic ratings, 3.8.5
Dynamic Thermal Circuit Rating (DTCR), 2.7.5, 3.8.2,
3.8.3, 6.5
DTCR output, 6.5.2
Power circuit modeling, 6.5.1
Dynamic Thermal Ratings
Condition assessment and real-time monitors, 6.3
Costs, 6.2.2
Field studies, 6.7
Models, 6.4
Accounting for heat storage, 6.4.1
Overhead lines, 6.4.2
Power Transformers, 6.4.3
Substation terminal equipment, 6.4.5
Underground cables, 6.4.4
Operating with dynamic thermal ratings, 6.6
E
Electric field, 1.3.5
Electrical clearance, 2.2.5
Elongation
Thermal elongation, 2.4.11
Emergency ratings, 2.4.1, 3.4.7
Environmental limits (for overhead transmission lines),
1.3.5, 2.1.4
EPR (Ethylene-Propylene-Rubber) cables, 3.2.2
Ethylene-Propylene-Rubber cables, 3.2.2
Extruded dielectric, 3.2.2
F
FACTS (Flexible AC Transmission Systems), 1.3.3, 6.2.3
Flexible AC Transmission Systems, 1.3.3, 6.2.3
Index Increased Power Flow Guidebook
I-2
G
Gapped Construction, 2.6.6
H
Heat balance methods, 2.3.5
Convection, 2.3.5
Ohmic losses, 2.3.5
Radiation, 2.3.5
Solar heating, 2.3.5
Steady-state thermal rating, 2.3.5
High-temperature operations (of overhead transmission
lines), 2.4
Connectors at high temperature, 2.4.13
High-pressure fluid-filled cables, 3.2.1
High-pressure gas-filled cables, 3.2.1
High-pressure pipe-type cables, 3.2.1
Hot spots
Identification (underground cables), 3.5.6
Remediation (underground cables), 3.6.6
HPFF (High-Pressure Fluid-Filled) cables, 3.2.1
HPGF (High-Pressure Gas-Filled) cables, 3.2.1
Hybrid (underground and overhead) circuits, 3.3.3
Hydraulic circuit, 3.5.8
I
Ice loading, 2.2.7
Invar steel core conductor, 2.6.5
L
Limiting conditions, 1.3
Circuit power flow limits, 1.3.1
Environmental limits, 1.3.5
Surge impedance loading of line, 1.3.2, 3.3.1
Thermal limits, 1.3.4, 3.3.1
Voltage drop limitations, 1.3.3
Linear low-density polyethylene cables, 3.2.2
Line traps, 5.3.8
LLPE (Linear Low-density Polyethylene) cables, 3.2.2
Long-time emergency rating, 6.6.1
Losses (underground cables), 3.4.3
M
Magnetic field, 1.3.5
Mass impregnated cables, 3.2.4
MDPE (Medium-Density Polyethylene) cables, 3.2.2
Medium-density polyethylene cables, 3.2.2
MI (Mass Impregnated) cables, 3.2.4
Modeling
Complex interfaces, 6.5.5
Power circuits, 6.5.1
Monitoring (underground cables), 3.8.4
N
National Electric Safety Code, 2.2.5
NESC (National Electric Safety Code), 2.2.5
Normal rating, 2.4.1
O
Ohmic losses, 2.3.5
Oil circuit breakers, 5.3.4
Operating with dynamic thermal ratings, 6.6
Overhead transmission lines
Case studies, 2.8
Dynamic monitoring and line rating, 2.7
Disadvantages, 2.7.4
Dynamic rating calculations, 2.7.6
Dynamic ratings versus static ratings, 2.7.2
Real-time monitors, 2.7.5
Dynamic thermal rating models, 6.4.2
Effects of high-temperature operations, 2.4
Annealing of aluminum and copper, 2.4.2
Axial compressive stresses, 2.4.4
Built-in stresses, 2.4.5
Calculation of conductor high-temperature sag
and tension, 2.4.8
Conductor hardware, 2.4.14
Connectors at high temperature, 2.4.13
Sag and tension of inclined spans, 2.4.7
Sag-tension models for ACSR conductors, 2.4.3
Wind speed effects on thermal ratings, 2.4.10
Line thermal ratings, 2.3
Line design effects on line ratings, 2.3.4
Maximum conductor temperature, 2.3.2
Transient thermal ratings, 2.3.7
Weather conditions for rating calculations, 2.3.3,
2.3.6
Reconductoring without structural modifications, 2.6
ACCR conductor, 2.6.7
ACSS and ACSS/TW, 2.6.3
Gapped construction, 2.6.6
High temperature aluminum alloy conductors,
2.6.4
Invar steel core, 2.6.5
TW aluminum wires, 2.6.2
Uprating constraints, 2.2
Constraints on structural loads, 2.2.7
Environmental effects, 2.2.8
High-temperature sag, 2.2.3
Loss of conductor strength, 2.2.6
Wind-induced conductor motion, 2.2.4
Uprating without reconductoring, 2.5
Deterministic methods, 2.5.2
Measure of Safety as a basis for line rating,
2.5.4
Probabilistic methods, 2.5.3
Sagging Line Mitigator, 2.5.6
I-3
Increased Power Flow Guidebook Index
P
Pipe-type cables, 3.2.1
Power flow example, 1.2
Power flow limits, 3.3
Load flow considerations, 3.3.2
Stability limits, 3.3.1
Surge impedance loading limits, 3.3.1
Thermal limits, 3.3.1
Uprating hybrid circuits, 3.3.3
Power system issues, 1.2
Power transformers
Design, 4.2
Cooling types, 4.2.2
Core-form construction, 4.2.1
Directed flow designs, 4.2.1
Factory testing, 4.2.4
Losses, 4.2.3
Shell-form construction, 4.2.1
Examples, 4.8
Modest increases in capacity, 4.7
Risks of increased loading, 4.3
Long-term risks, 4.3.2
Short-term risks, 4.3.1
Thermal modeling, 4.4
Bottom oil model, 4.4.3, 6.4.3
IEC model, 4.4.4, 6.4.3
Mechanisms of heat transfer, 4.4.1
Proposed IEC model, 4.4.5
Top oil model, 4.4.2, 6.4.3
Thermal ratings, 4.5
Ambient air temperature, 4.5.1
Condition-based loading, 4.5.5
Load, 4.5.2
Maintenance considerations, 4.5.6
Rating procedure, 4.5.4
Rating type and duration, 4.5.3
Winding temperature measurement, 4.6
Q
Quasi-dynamic ratings (underground cables), 3.8.5
R
Ratings
Continuous (normal) rating, 6.6.1
Emergency ratings, 2.4.1, 3.4.7
Long-time emergency rating, 6.6.1
Short-time emergency rating, 6.6.1
Traditional rating definitions, 6.6.1
Underground cable ratings, 3.4
Reconductoring
Overhead transmission lines, 2.6
Resistances (underground cables), 3.4.4
Route thermal survey, 3.6.1
S
Sag
High-temperature sag with aluminum conductors,
2.2.3
High-temperature sag with ACSR, 2.2.3
High-temperature sag-tension calculations, 2.4.9
SAG10, 2.2.2, 2.4.8, 2.4.9
Sag and tension of inclined spans, 2.4.7
Sag-tension calculations, 2.2.2, 2.4.6, 2.4.8, 2.5
Tension-elongation diagram, 2.2.2
Sag-tension models, 2.4.3
Sagging Line Mitigator, 2.5.5
SCLF (Self-Contained Liquid-Filled) cables, 3.2.3
Self-contained liquid-filled cables, 3.2.3
SF
6
circuit breakers, 5.3.5
Shield/sheath bonding scheme, 3.6.8
Short-time emergency rating, 6.6.1
SIL (Surge Impedance Loading), 1.3.2, 2.1.1, 3.3.1
SLiM (Sagging Line Mitigator), 2.5.5
Static ratings, 2.3.1
STESS software, 2.2.3
Substation terminal equipment
Equipment types and IPF opportunities, 5.2
Rating parameters, 5.2.1
Thermal models, 5.3
Air-core reactor, 5.3.3
Bus conductors, 5.3.1
Bushings, 5.3.6
Current transformers, 5.3.7
Line traps, 5.3.8
Oil circuit breaker, 5.3.4
SF
6
circuit breaker, 5.3.5
Switch (air disconnect), 5.3.2
Thermal parameters, 5.5
Uprating, 5.4
Maintenance and inspection procedures, 5.4.2
Monitoring and communications, 5.4.1
Reliability and consequences of failure, 5.4.3
Superconducting cables, 3.7.5
Surge impedance loading, 1.3.2, 2.1.1, 3.3.1
Switches (air disconnect), 5.3.2
T
T-Aluminum Conductor Steel Reinforced (TACSR),
2.6.4
Temperature monitoring (underground cables), 3.6.4
Tension-elongation diagram, 2.2.2
Thermal limits, 1.3.4, 2.1.3, 3.3.1
Thermal models
Substation terminal equipment, 5.2.1, 5.3
Index Increased Power Flow Guidebook
I-4
Thermal ratings
Dependence on weather conditions, 2.3.6
Dynamic ratings, 2.3.1
Maximum conductor temperature, 2.3.2
Static ratings, 2.3.1
Transient thermal ratings, 2.3.7
Weather conditions for rating calculation, 2.3.3
Wind speed, 2.4.10
Power transformers, 4.5
Ambient air temperature, 4.5.1
Condition-based loading, 4.5.5
Load, 4.5.2
Maintenance considerations, 4.5.6
Rating procedure, 4.5.4
Rating type and duration, 4.5.3
Transformers. See Power transformers.
TW aluminum wires, 2.6.2
U
Underground cables
Cable system types, 3.2
Compressed gas insulated transmission, 3.2.4
Extruded dielectric, 3.2.2
High-pressure pipe-type, 3.2.1
Mass impregnated, 3.2.4
Paper insulated lead covered, 3.2.4
Self-contained liquid-filled, 3.2.3
Case studies, 3.9
Dynamic ratings, 3.8, 6.4.4
Benefits, 3.8.3
Dynamic Thermal Circuit Ratings, 3.8.2
Monitoring, 3.8.4
Quasi-dynamic ratings, 3.8.5
Increasing ampacity, 3.6
Active uprating, 3.6.7
Ampacity audits, 3.6.5
Evaluation of load patterns, 3.6.3
Remediation of hot spots, 3.6.6
Review circuit plan and profile, 3.6.2
Route thermal survey, 3.6.1
Shield/sheath bonding scheme, 3.6.8
Temperature monitoring, 3.6.4
Power flow limits and system considerations, 3.3
Load flow considerations, 3.3.2
Surge impedance loading limits, 3.3.1
Thermal limits, 3.3.1
Uprating hybrid circuits, 3.3.3
Ratings, 3.4
Ampacity, 3.4.2, 3.4.5, 3.4.6
Emergency ratings, 3.4.7
Equivalent thermal circuit and thermal
resistances, 3.4.3
Inferring conductor temperatures from measured
temperatures, 3.4.8
Losses, 3.4.3
Reconductoring, 3.7
Cupric oxide strand coating, 3.7.3
Larger conductor sizes, 3.7.2
Superconducting cables, 3.7.5
Voltage upgrading, 3.7.4
Uprating and upgrading constraints, 3.5
Accessories, 3.5.7
Direct buried cable systems, 3.5.1
Duct bank installations, 3.5.3
Fluid-filled cable systems, 3.5.2
Hot spot identification, 3.5.6
Hydraulic circuit, 3.5.8
Trenchless installations, 3.5.4
Uprating, active (underground cables), 3.6.7
Uprating case studies
Uprating constraints (overhead transmission lines), 2.2
Constraints on structural loads, 2.2.7
Environmental effects, 2.2.8
Limiting high-temperature sag, 2.2.3
Loss of conductor strength, 2.2.6
Uprating constraints related to wind-induced
conductor motion, 2.2.4
Uprating constraints (underground cables), 3.5
Direct buried cable systems, 3.5.1
Duct bank installations, 3.5.3
Fluid-filled cable systems, 3.5.2
Trenchless installations, 3.5.4
Uprating hybrid (underground and overhead) circuits,
3.3.3
Uprating substation terminal equipment, 5.4
Uprating without reconductoring (overhead
transmission lines), 2.5
Deterministic methods, 2.5.2
Probabilistic methods, 2.5.3
Absolute method, 2.5.3
Exceedance method, 2.5.3
Modified exceedance method, 2.5.3
V
Video sagometer, 2.7.5
Voltage drop, 1.3.3, 2.1.2
Voltage limits, 3.3.1
Voltage upgrading, 3.7.4
W
Wind-induced conductor motion, 2.2.4
Wind loading, 2.2.7
Wind speed, effect on thermal ratings, 2.4.10
I-5
Increased Power Flow Guidebook Index
X
XD (Extruded Dielectric), 3.2.2
XLPE (Cross-Linked Polyethylene), 3.2.2

The Electric Power Research Institute (EPRI), with major
locations in Palo Alto, California; Charlotte, North Carolina; and
Knoxville, Tennessee, was established in 1973 as an independent,
nonprofit center for public interest energy and environmental
research. EPRI brings together members, participants, the Institutes
scientists and engineers, and other leading experts to work
collaboratively on solutions to the challenges of electric power. These
solutions span nearly every area of electricity generation, delivery,
and use, including health, safety, and environment. EPRIs members
represent over 90% of the electricity generated in the United States.
International participation represents nearly 15% of EPRIs total
research, development, and demonstration program.
Together...Shaping the Future of Electricity
Export Control Restrictions
Access to and use of EPRI Intellectual Property is granted with the
specic understanding and requirement that responsibility for ensur-
ing full compliance with all applicable U.S. and foreign export laws
and regulations is being undertaken by you and your company. This
includes an obligation to ensure that any individual receiving access
hereunder who is not a U.S. citizen or permanent U.S. resident is
permitted access under applicable U.S. and foreign export laws and
regulations. In the event you are uncertain whether you or your com-
pany may lawfully obtain access to this EPRI Intellectual Property, you
acknowledge that it is your obligation to consult with your companys
legal counsel to determine whether this access is lawful. Although
EPRI may make available on a case-by-case basis an informal as-
sessment of the applicable U.S. export classication for specic EPRI
Intellectual Property, you and your company acknowledge that this
assessment is solely for informational purposes and not for reliance
purposes. You and your company acknowledge that it is still the ob-
ligation of you and your company to make your own assessment
of the applicable U.S. export classication and ensure compliance
accordingly. You and your company understand and acknowledge
your obligations to make a prompt report to EPRI and the appropriate
authorities regarding any access to or use of EPRI Intellectual Prop-
erty hereunder that may be in violation of applicable U.S. or foreign
export laws or regulations.
Electric Power Research Institute
3420 Hillview Avenue, Palo Alto, California 94304-1338PO Box 10412, Palo Alto, California 94303-0813 USA
800.313.3774650.855.2121askepri@epri.comwww.epri.com
Program:
Increased Transmission Capacity
1010627
2005 Electric Power Research Institute (EPRI), Inc. All rights reserved. Electric Power
Research Institute, EPRI, and TOGETHER...SHAPING THE FUTURE OF ELECTRICITY are
registered service marks of the Electric Power Research Institute, Inc.
Printed on recycled paper in the United States of America

Increased Power Flow Guidebook

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Increased Power Flow Guidebook

Uploaded by

Copyright:

Available Formats

Electric Power Research Institute

cos cos 0.368 sin (2 )

Table 2.3-3 Total Heat Flux Received by a Surface at Sea

Table 3.4-1 Typical Cable Insulation Material Parameters

Chapter 3: Underground Cables Increased Power Flow Guidebook

Chapter 3: Underground Cables Increased Power Flow Guidebook

C. The static circuit

You might also like