You are on page 1of 14

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/297137713

The history and mystery of the: Neher-McGrath formula

Article · July 2014

CITATIONS READS
2 41

1 author:

Tim Linenbrink

3 PUBLICATIONS   2 CITATIONS   

SEE PROFILE

All content following this page was uploaded by Tim Linenbrink on 03 May 2020.

The user has requested enhancement of the downloaded file.


The history and mystery of the
Neher-McGrath formula
In its abbreviated form, the Neher-McGrath formula appears straightforward.
For electrical engineers, it must be carefully reviewed and used.
BY TIMOTHY L. LINENBRINK, PE, North American Engineering PLLC, Richland, Wash.

A
while back, a col-
Learning league brought up
objectives a problem that had
 Learn the scientific basis occurred on a $100 mil-
for the Neher-McGrath for- lion industrial project. A few
mula. months after plant turnover,
 Understand how to use the several feeder breakers start-
Neher-McGrath formula to ed tripping. There are few
find conductor ampacity.
things that will cost an electri-
 Appreciate factors that can
cal engineer more sleep than
impact accuracy.
random breaker operations.
After considerable consterna- Figure 2: The NEC allows, under engineering
tion, the under-slab feeders supervision, the use of an abbreviated form of
were pulled out and found to Neher and McGrath’s cable rating equation. But
have heat damaged insulation. correctly applying it requires understanding the
The cables had been sized per assumptions and simplifications that are buried
National Elec- within. Courtesy: North American Engineering
trical Code
(NEC) Article 310 and the temperature above the insulation’s
there wasn’t any over- thermal limit. Overheating degraded
loading, so what was the the insulation; add a little moisture, and
problem? breakers start opening.
The cause is an object In hindsight the cable ratings should
lesson we could all learn have been adjusted for the actual tem-
from: You can not blindly perature in which they were expected to
apply the NEC tables to operate. Article 310.15 of the NEC points
determine cable ampacity. this out and provides the Neher-McGrath
The installation and oper- formula as the engineered solution for
ating conditions must be doing so.
considered. In its abbreviated form, Neher-McGrath
Figure 1: The NEC tables cannot always be used to Located above the slab appears straightforward (see Figure 2).
determine cable ampacity. The number of ducts, their were a series of industrial The devil is in the details. Tc is readily
proximity, and site-specific conditions combine to make furnaces. The cables’ available from product data; an Rdc value
2
ampacity calculations a complex undertaking. Here we Joule (I R) losses, com- (not necessarily the correct value) is avail-
see a duct bank being installed free of interfering heat bined with heat radiating able from published sources, but what
sources. Courtesy: Stokes Electric of Central Florida Inc. through the slab, elevated about the other factors? Is the ΔTd sig-

46 Consulting-Specifying Engineer • JULY 2014 www.csemag.com


nificant? What if there is an external heat
source affecting the installation? What if
a complex cable type is used? While the
NEC version of Neher-McGrath appears
straightforward, finding the right param-
eters that accompany it is not.
If a cable of complex construction is
used (see Figure 10), the rating equation
takes the form of Figure 3.
In practice the terms “conductor” and
“cable” are used interchangeably. For
clarity the term “conductor,” as used
here, indicates the current-carrying part
of a cable. The term “cable” refers to a
complete assembly, for example, conduc-
tors, filler, insulation, jacket, armor, serv-
ing, and so on.
Adapted from Ander’s, Rating of Figure 4: The underlying principle of Neher-McGrath is the Kennelly Hypothesis.
Electric Power Cables, ΔT is the temper- Developed by Arthur E. Kennelly in 1893, its premise is that the temperature rise at
ature rise over ambient, Wd the dielectric any point can be predicted based on the heat produced, the heat absorbed, the ambi-
heating, Ri the thermal resistances, n the ent temperature, and the intervening thermal resistivities. Courtesy: North American
number of conductors, Rac the ac resis- Engineering
tance, and λ i the loss ratios for armor
and shielding. History of Neher-McGrath Carnegie Institute and a founding mem-
Labeling Neher and McGrath’s work a John Hutchins Neher (1899–1973) and ber of the AIEE Insulated Conductors
“formula” implies ampacity problems can Martin Hager McGrath (1902–1980) did Committee, was vice president and chief
be solved algorithmically; provide a few not have an epiphany when they wrote engineer of the General Cable Corp. In
inputs, crank through some calculations, their 1957 paper, “The Calculation of the 1962 he arranged for then competitor
and out pops an answer. Their paper does Temperature Rise and Load Capability of Anaconda Wire and Cable (General Cable
not provide an algorithm; ampacity prob- Cable Systems.” They had worked with acquired Anaconda Wire and Cable in
lems cannot be solved by plug-and-chug. a cadre of engineers on the problem for 1999) to develop the first set of ampac-
Neher and McGrath’s work is a body decades; the significance of their con- ity tables on an IBM 650—tables that
of knowledge, some aspects of which tribution was in compiling the collected became AIEE Special Publication S-135,
are applicable to an installation, some knowledge as it stood at the time. Power Cable Ampacities, a forerunner of
are not. Equations like those above must Neher, a 1921 graduate of Princeton, IEEE 835, Standard Power Cable Ampac-
be considered case-specific; as presented was a senior engineer with the Philadel- ity Tables.
they include assumptions and simplifica- phia Electric Co. A Navy Commander Neher and McGrath’s goal was to
tions that affect the validity of any calcu- during WWII, he served in London work- develop something the practicing engi-
lated result. ing on the use of radar to find U-boats. neering could use with the computing
He was made a Fel- tools available at the time. For most
low of the American engineers that was the slide rule. This
Institute of Elec- was no easy task considering that in
trical Engineers their lifetimes published ampacities
(AIEE), now the varied widely. AWG #1/0 copper, for
Figure 3: The rating equation becomes more compli- Institute of Electri- example, had values ranging from 71
cated depending on the installation, the application, and cal and Electronics to 372 amps.
the cable used. When dielectric effects and shield losses Engineers (IEEE), (In full disclosure, Neher and McGrath
are factors, Anders’ equation is more applicable. (Based in 1957 for his con- also dealt with conductors in air, a topic
on the discussion in “Rating of Electric Power Cables: tributions in both which will not be discussed here. Inter-
Ampacity Computations for Transmission, Distribution, cable heating and ested readers are directed to IEEE Std
and Industrial Applications” by George Anders.) Cour- protective relaying. 738-2006, IEEE Standard for Calculating
tesy: North American Engineering McGrath, a 1924 the Current-Temperature of Bare Over-
graduate of the head Conductors.)

www.csemag.com Consulting-Specifying Engineer • JULY 2014 47


Neher-McGrath formula
The Kennelly hypothesis A heat source, +qc, is buried at a depth
Arthur E. Kennelly (1861-1939) was Lb in uniform soil (constant ambient tem-
one contemporary Neher and McGrath perature, constant thermal resistivity,
drew on. His research on ampacity was a etc.). Heat will migrate from the source
de facto standard until 1938 when Samuel to all points of lower temperature by con-
J. Rosch, an employee of Anaconda Wire duction (and by convection and radiation
and Cable working under the auspices of if enclosed in a pipe or conduit containing
National Electrical Manufacturers Asso- air). If the heat sink is treated as a single Figure 6: Fundamental to calculating
ciation (NEMA), produced what became point, -qc, represented by the reflected ampacity is the “buildup” of the cable.
the NEC ampacity tables, which were in image of the source, the temperature rise Typical low-voltage cables are made of
use until the 1980s. can be found from the difference in the stranded conductors covered by a layer
Aside from working on other topics, two heat flows. of insulation with a nylon jacket. Cour-
Kennelly developed the hypothesis that A relatively simple design problem tesy: North American Engineering
underlies Neher-McGrath: The movement (Figure 5) helps to put Neher-McGrath’s
of heat through a cable system’s thermal work in context; the objective being to wide, buried 36 in. below the earth’s sur-
resistances will cause a temperature rise. find the cable system’s current carrying face. Three parallel sets of equally loaded,
Treat the cable system as an infinitely capacity. (This example was not chosen single-core, 75 C, 500 kcmil aluminum
long, cylindrical heat source buried at randomly; NEC Annex B and IEEE 835 cables form the circuit. The ambient tem-
some depth in a uniform medium, and the values are used so the results can be com- perate of the earth is 20 C. The thermal
temperature rise at any point—be it the pared to published values, an important resistivity of the earth, insulation, conduit,
surface of a cable, interior of a duct, or verification and validation step.) and concrete are taken as 120, 450, 600,
any other point—can be predicted. This For underground cable systems, deter- and 55 C-cm/W, respectively.
is shown graphically and mathematically mining ampacity involves five elements: The cable buildup (Figure 6) consists
in Figure 4. of a 0.813 in. diameter, multi-strand alu-
1. Defining the minum conductor with 95 mils (0.095
installation in.) polyethylene (PE) insulation under a
2. Finding the nylon jacket. (Insulation thicknesses are
conductor’s ac from IEEE 835.)
resistance at the Assuming no interfering sources, the
desired operating only heat production is by the conductor’s
temperature I2R losses.
3. Determining the Temperature rise due to dielectric
thermal resis- heating, ΔTp, is ignored. Per the NEC,
tances of each dielectric heating is negligible for volt-
element ages below 44 kV. As shown in Figure 7,
4. Calculating the dielectric losses are less than 0.5 W/ft at
total effective that voltage and virtually zero at 480 V.
thermal resis- A helpful technique is to think of the
tance installation as a thermal circuit (Figure
5. Computing the 8): a rationalization justified by Ohm’s
cable’s ampacity. Law being analogous to Fourier’s Law,
electrical resistance equivalent to thermal
Inputs and design resistance, electrical potential (voltage)
Figure 5: For the design example, a three-way underground parameters being on par with thermal potential (tem-
duct bank consists of 4-in. ducts encased in concrete and This problem is for perature change), and charge flow (cur-
buried 36 in. below the earth’s surface. Neher-McGrath an industrial cable duct rent) being synonymous with heat flow.
uses approximations and empirical factors to reduce such operating at 480 V. It Viewing it in this way emphasizes the
systems to a cylindrical equivalent heat source and cylindri- consists of three poly- elements included and allows the prob-
cal layers of thermal resistances. The calculations require vinyl chloride (PVC) lem to be solved using circuit analysis
understanding the factors and approximations that are valid conduits embedded in techniques. (This approach was advanced
and those that are not for each installation. Courtesy: North a concrete duct bank by Donald L. Simons in the 1920s, and is
American Engineering 11.5 in. tall and 27 in. detailed Anders’ texts.)

48 Consulting-Specifying Engineer • JULY 2014 www.csemag.com


Figure 8: Reducing the physical installation to an equivalent
thermal circuit is the accepted practice for modeling under-
ground cable installations. If the only heat is from the Joule
losses, the thermal circuit reduces to a single source and with
the thermal resistivities connected in series. Elementary circuit
analysis techniques can then be used to find the temperature
rise, ΔT. Courtesy: North American Engineering

Figure 7: For voltages less than 44 kV, the NEC ignores heating
due to dielectric effects, justifiable since below this voltage the
losses are less than 0.5 W/ft. The dielectric effect was modeled
in MathCAD Prime 3.0 and the heat loss plotted for voltages
from 0.480 to 110 kV. Courtesy: North American Engineering

ac conductor resistance The work of


A precursor to finding the cable’s (2.1) through Figure 9: The literature uses differing forms for many of the equa-
ampacity is finding the conductor’s ac (2.10) can be tions. Here an example from Neher and McGrath on the left is
resistance at the operating temperature, avoided by substi- contrasted with the equivalent metric form used by Anders on the
or more precisely the factors that con- tuting published right. The two are equivalent, but not obviously so without some
stitute ac resistance, that is, skin and ac resistance unit and exponential manipulation. Courtesy: North American
proximity effects. values, but care Engineering
From Southerwire’s Power Cable must be exercised
Manual, the dc resistance of stranded to ensure that the right temperature- A complication in these calculations is
aluminum, at a base temperature of 25 corrected resistance, skin, and proxim- the mix of British Imperial (U.S. Cus-
C, is 35.4 μΩ/ft. The corresponding 25 ity factors are used. tomary) and metric units: Lengths and
C temperature coefficient of aluminum thicknesses are in inches, temperatures
is 0.00395 K -1. Referring to the accom- Thermal resistances in Celsius and Kelvin, thermal resistances
panying worksheet—related equations Thermal resistance, the inverse of in Thermal-Ohm-Feet, and so on. As a
indicated parenthetically—the dc resis- thermal conductivity, is the property demonstration consider (Figure 9) two
tance by (2.1) at an operating tempera- that restricts heat flow and causes tem- seemingly different, albeit similar, equa-
ture of 75 C is 42.392 μΩ/ft. perature rise. In this example, there are tions are in fact—once conversion factors
ac resistance is found by adjusting five layers of material (cable insulation, and identities applied—the same.
the dc value by the skin effect at 60 Hz the medium in the duct, the duct wall, Cable insulation thermal resistance:
and the proximity effect of all cables. the concrete duct bank, and the soil) that In this example, using three single con-
Using an approximation for the Bessel impede heat transfer. (The concrete and ductor cables in a triangular configura-
function, the skin effect coefficient is soil get treated concurrently in one com- tion, the thermal resistance of the insu-
found by (2.2) and (2.3). Similarly, the plex sequence.) lation is found by (3.2). Equation (3.2)
proximity effect coefficient is found In 1951 John Neher observed in his applies to single conductor cables only.
from (2.5) through (2.8). k s and k p paper “The Determination of Tempera- If there were multiple insulation layers,
are factors related to the arrangement ture Transients in Cable Systems by (3.2) would have to be repeated for each
and coating of the conductor strands; Means of Analogue Computer”: layer. If the cable was armored, or includ-
their values are available in published “The [ampacity] calculation … is a ed skid wires, or used metal tape (Figure
sources. relatively simple matter; the only diffi- 10), the additional thermal resistances
Equation (2.9) is the Y c factor in culty experienced is that of determining would have to be accounted for and the
Neher-McGrath. The total ac resistance the proper thermal constants for the com- additional heating due to ferromagnetic
is found from (2.10). ponents of the thermal circuit.” effects would have to be included.

www.csemag.com Consulting-Specifying Engineer • JULY 2014 49


Neher-McGrath formula
Neher, working in conjunction with a and iterate until a constant air tempera-
General Electric engineer named Fran- ture is achieved. Experience shows a
cis H. Buller, published a separate paper, solution converges in less than 10 itera-
“The Thermal Resistances Between tions if 60 C is used (Figure 11), or 15
Cables and a Surrounding Pipe or Duct iterations if starting at the ambient soil
Wall,” in 1949 that developed an approxi- temperature, Ta.
mation approach to this part of the prob- Duct wall thermal resistance: PVC
lem. Rearranging the terms in the com- conduit is the duct running through the
pound differential equation and plotting bank. Its thermal resistance is factored
Figure 10: Ampacity calculations for the experimental data on duct tempera- in even though the contribution is on the
complex cables like this one are more tures, they observed that the temperature order of 1.0 Thermal-Ohm-Feet. If steel
involved due to the layers of material, rise (of the medium in the duct) against conduit was used, a value of 1.0 could
filler materials, and the additional heat- the thermal resistivity approximates a be assumed.
ing caused by the magnetic induction of straight line. Treating it as such, they In (3.6) Neher and McGrath substitute
the skid wires. Courtesy: ABB extracted A, B, and C from the slopes a “thin-wall approximation” to find Rd; n
and intercepts. has the same value as before. (The thin-
Two- and three-core cables require The values of A, B, and C, along with wall approximation exploits linearity of
equations with a geometric factor, Gb. For the multiplier to find D’s, vary with the logarithms for numbers on the order of
a three-core cable (3.2) would be: cable type and lay. For the various con- 1.001, 1.002, etc. In cases where the ratio
figurations, see Table VII in Neher and on the duct’s outer-to-inner diameter is
McGrath’s paper. of this order, the approximation works.)
n in this equation is not always n; Neher Concrete duct bank and earth ther-
Equations for Gb were developed by and McGrath use the variable in differ- mal resistance: Neher and McGrath’s
Mie, Russell, and Simons but are fairly ent contexts and their intent is not always way of finding the thermal resistance
involved; Anders provides a thorough clear. Sometimes n is the total number
discussion of their makeup. Curves for of ducts, sometimes the total number of
finding Gb are given in Simons’ Calcu- cables in the system, and sometimes the
lation of the Electrical Problems of the total number of cables in the ducts. When
Transmission by Underground Cables, this ambiguity arises, clarity can be found
Standard Underground Cable Company, by asking, “How many heat-producing
Pittsburgh, 1925; in British Standard cables are in this particular part of the
BS7769-2-2.1; and in Anders. problem?” The answer is the value of
Lastly, the nylon jacket is ignored. Nylon n. Here it is 3, the number of energized
has a thermal resistivity up to 830 C-cm/W. cables within a single duct.
Low-voltage cables have very thin jackets, Per Kennelly, heat sources are reduced
on the order of 10 to 50 mils, contributing to a single, equivalent, cylindrical source.
less than 0.5 Thermal-Ohm-Foot. Within the duct this cylinder has the
Duct medium: The thermal resistance diameter circumscribing all three cables,
of the air in the duct is not insignificant, hence the geometric adjustment D’s. A Figure 11: When an iterative approach
amounting to 25% of the thermal circuit. similar situation arises in making the rect- is used to find the temperature in the
Ignoring it overstates ampacity by 16%. angular duct bank appear round. duct, the temperature converges to a
The equation for determining it (3.5) Tm is the assumed mean temperature of steady-state solution in less than 10
is unlike any of the others, and needs the air in the duct; Neher and McGrath iterations if a starting temperature of 60
explaining. use 60 C. This may come from the IEEE C is used, or 15 iterations starting with
Within the duct itself, heat transfer is forerunner, the American Institute of the ambient temperature. A program was
by conduction, convection, and radia- Electrical Engineers AIEE standardiza- developed to calculate cable ampacity by
tion. Assuming everything is linear and tion rules for the maximum allowed tem- Neher-McGrath, determine the resulting
the principle of superposition applies, we perature of natural rubber insulation, but Joule losses, and then the temperature
could just add the three together. How- that is an inference from the literature. in the duct (Tm). The process was then
ever, combining all three heat transfer An alternate approach is to assume any repeated, substituting the new value of
equations crates a differential equation air temperature, solve for the current in Tm, until it became constant. Courtesy:
without an analytic solution. the cables, determine the heat produced, North American Engineering

50 Consulting-Specifying Engineer • JULY 2014 www.csemag.com


of the concrete duct bank and the soil Neher and McGrath derive another
is complicated. Included are load cycle geometric factor, also called Gb, based on
effects, a cylindrical equivalent duct the burial depth and radius of an equiva-
bank, and the mutual heating of all lent round duct bank. This is derived
cables. They find the thermal resistance from (3.10) through (3.12), where x1 is
by assuming concrete throughout then always the short dimension and y1 the
adjusting for the soil, a counterintuitive long dimension. The equivalency is valid
approach compared to the other elements. for y 1/x 1 being less than 3. (For other
ratios, El-Kady and Horruck developed
Load cycles an alternate form of Gb and published it in
The load cycle effects show up twice: “Extended Values for Geometric Factors
in the form of a fictitious diameter, Dx, of External Thermal Resistance of Cables Figure 12: The mutual heating of all the
and a corresponding loss factor, LF. in Duct Banks.”) cables is a dominant factor in ampacity
The fictitious diameter, D x , is an A very critical element is F, the mutual calculations. The mutual heating effect
assumed distance at which the effect of heating factor, a multiplier directly relat- factor, F, for a three-duct system is found
load changes commences. For continuous ed to the Kennelly Hypothesis. The effect from the ratios of the distances between
loading, a fictitious diameter must still of the hottest duct and its reflected image each duct and its reflected image. Cour-
be found. is taken as 1. The effect of the other ducts tesy: North American Engineering
Dx is dependent on α, the soil thermal and their images are proportional to their
diffusivity and the duration of the load locations relative to the hottest duct. arguing over soil resistivity. As Conrad
cycle. Equation (3.7) is the modern form Because of the former, (3.17) will always Bauer, a Commonwealth Edison cable
for α and not the one used by Neher and have 1 term less than the number of ducts. engineer, observed about calculations
McGrath. For continuous loading, the The ratios are the distance between the and thermal resistivities at an AIEE con-
load cycle is 24 hours. hottest duct’s center and a duct’s reflected ference, “To date we have not been able
Due to an odd choice of symbols and image, divided by the axial separation of to match field tests […], without first
an unconventional choice of terms, the the hottest duct and the corresponding introducing known answers and work-
loss-load factors can be a source of con- neighbor (Figure 12). ing backwards…”
fusion. The loss factor LF is dependent In simple situations the hottest duct can
on a daily loss cycle having a load factor be identified by inspection; in complex con- Soil considerations
lf. The latter is a percentage of the cable figurations it may be necessary to calculate For insulation and conduit the ther-
systems daily loading. LF is found from: all possible values of F and use the largest mal resistivity of the material is readily
found. F increases geometrically with the available from the manufacturer or pub-
number of ducts; it tends to dominate (3.18). lished sources. However, NEC Annex B
Referring back to n is not always n, it provides some rather specific values for
For continuous loading lf would be 1.0, is worth pointing out that n here is the earth and concrete without elaboration.
making LF 1.0 as well. (“Load factor” number of cables in one duct and N is the The 1920s and ‘30s produced a lot of
should not be confused with “demand number of ducts in the cable system. De is field research on soil resistivity. Taken as
factor”; both are related to cyclical load- the diameter in contact with the concrete a whole, one can conclude: 1) ρe varies
ing but their application is different.) (if a duct bank is used) or the earth (if a lot between locations, and 2) ρe drifts
Historically, the loss-load factors were direct buried conduit or cable is used). If over time. The latter is especially true
included for underground utility systems. there is no duct bank (3.19) reduces to: where heat causes moisture migration
Utilities optimize their underground plant or in sandy soils that undergo cycles of
by trying to operate near the thermal lim- saturation and drying.
its. In 1953 the AIEE sponsored a sym- In the example, underestimating the
posium on how to include load cycles in value of ρe by 25% will overstate cable
cable ampacity, but no real consensus with ρe substituted for ρc, and n becoming ampacity by 9%.
emerged. Four years later Neher and n’ = n×N. Similarly, the American Concrete
McGrath published their paper and this Institute’s Guide to Thermal Proper-
equation became de facto. It would be an omission not to pause ties of Concrete and Masonry Systems
The use of any loss factor is a matter and discuss ρ; primarily, ρe and ρc. observes that the thermal resistivity of
of engineering judgment and unique to As long as cable engineers have been concrete depends on the aggregate used.
each installation. arguing over ampacity, they have been Values of ρc can range from 8 C-cm/W

www.csemag.com Consulting-Specifying Engineer • JULY 2014 51


Neher-McGrath formula
to 111 C-cm/W. However, the effect on
ampacity is small; the worst-case error
changing ampacity by 3%.
The simple fact is ρ can only be found
by measuring it.
A basis for the ambient soil temperature
also is needed. This can be done by long-
term measurement (a luxury rarely avail-
able) or by consulting research sources.
For example, Pacific Northwest National
Laboratories has extensive soil tempera-
ture measurement histories (Table 3.15). Figure 13: A modern approach to find cable ampacity is the finite element method.
It may be reasonable to infer soil tempera- Here, the finite element mesh for a two-duct system was developed in Lisa 8.0 (for
ture from the frost depth, a data source clarity, the mesh representing the air in the duct was omitted). The next step is to
being the National Oceanic and Atmo- solve the heat transfer equations at each node in the mesh using a known set of
spheric Administration database. Below boundary conditions. The boundary conditions are a constant temperature at the sur-
the frost depth, soil temperature is unlikely face of each conductor and an ambient temperature at the outside edges. Courtesy:
to ever be less than 0 C. North American Engineering
For the sample installation, a 10 C error
in soil temperature will affect ampacity by would be needed to find the temperature been toward numerical methods using a
±10%. rise, ΔTinf. combination of boundary element methods
Returning to the thermal circuit, the An interfering temperature of 15 C (BEM) and finite element analysis (FEA).
effective thermal resistance is found by would drop cable ampacity to 223 amps. These techniques allow multi-physics sim-
summing the components: ulations where both the thermal and electric
Keeping the equation in mind effects can be modeled concurrently. Figure
There is clearly more to Neher-McGrath 13 shows the finite element mesh of a sim-
than meets the eye. The calculations have ple duct installation used for such analysis.
reducing the circuit to a Thevenin’s equiv- ample opportunities for error and misap- This approach models the physical sys-
alent from which the cable ampacity is plication. Proper use requires keeping in tem at a granular level with each element
found. mind that: defined by its material properties. Using
 Neher-McGrath is a body of knowl- known boundary conditions, the underlying
Cable ampacity edge; engineering judgment is energy conservation laws can be solved to
Once the dc resistance (at the desired required. find temperature rise and ampacity values.
operating temperature) is known—along  Buried within its equations are The finite element approach makes pos-
with the skin and proximity effect factors, assumptions, simplifications, and sible representing the entire installation;
the total effective thermal resistance, and constraints that may or may not transitions and terminations need not be
the design temperature—finding cable always be applicable. ignored.
ampacity is, as Neher observed, rather  Neher-McGrath does not consider Today, the Neher-McGrath formula is
anticlimactic. harmonics, and assumes balanced the code-accepted practice for determining
Including rounding and truncation phase loading, two omissions that cable ampacity. Certainly no one wants to
errors, the current carrying capacity of can significantly impact operating revisit the acrimony that accompanied its
the cables is found to be 250 amps, cor- temperature, ampacity, and cable incorporation into the NEC, but modern
responding to the published value. protection. techniques have a great deal to offer over a
If the soil thermal resistivity turned out method intended for the slide rule.
to be 77 C-cm/W instead of 120 C-cm/W, Computational methods and machines
ampacity would increase by 13%; making have progressed considerably in the 45 Timothy L. Linenbrink is a principal
a 250 amp circuit a 300 amp one. years since Neher and McGrath stopped engineer at North American Engineer-
It was assumed there are no external working on the problem. (Neher retired ing. A 1986 graduate of the University
(i.e., interfering) heat sources. If the from Philadelphia Electric in 1966, of Wyoming, he is a licensed professional
ducts ran adjacent to steam lines, as in McGrath from General Cable in 1969.) engineer in 14 states. He has performed
the case of some facilities, the additional Today it is doubtful the problem would be engineering, consulting, and project
heat source’s effect would have to be approached in a pseudo-algorithmic fash- management services for the Naval Sea
included. A calculation similar to (3.2) ion. Since the early 1980s the trend has Systems Command, Bechtel, and others.

52 Consulting-Specifying Engineer • JULY 2014 www.csemag.com


 

ERRATA 

The original article omitted the following calculations. The parenthetical equation referenced starting on 
page 49 correlate to the equations in the following sheets. 

### 
1. INPUTS AND PARAMETERS

CABLE BUILD UP AND PARAMETERS


Conductor base resistance temperature: TBase ≔ 25.00 [ºC]
Conductor dc resistance at 25°C: Rdc_25C ≔ 35.4 [ μΩ /foot]
Conductor diameter: dc ≔ 0.813 [inches]
Conductor proximity effect coefficient: kp ≔ 1.00 [unitless]
Conductor skin effect coefficient: ks ≔ 1.00 [unitless]
Conductor temperature coefficient at 25°C: α25C ≔ 0.00395 [1/K] (aluminum)
Cable insulation thickness: ti ≔ 0.095 [inches]
Cable diameter: Dc ≔ dc + 2 ⋅ ti [inches]
Cable (operating) temperature: Tc ≔ 75.00 [ºC]
Cable system interference temperature: ΔTinf ≔ 0.00 [ºC]

DUCT BANK DIMENSIONS AND PARAMETERS


Duct bank burial depth: Lb ≔ 36.0 [inches]
Duct bank cable group axial separation: s1 ≔ 7.5 [inches]
Duct bank dimensions: x1 ≔ 11.5 by y1 ≔ 27.0 [inches]
Duct inside diameter: Did ≔ 4.026 [inches]
Duct mean temperature: Tm ≔ 73.2083 [ºC] (by iteration)
Duct outside diameter: Dod ≔ 4.5 [inches]
Duct wall thickness: dt ≔ ⎛⎝Dod − Did⎞⎠ ÷ 2 = 0.237 [inches]
Diameter of the surface in contact with the earth: De ≔ Dod = 4.5 [inches]
Number of conductors per cable group: n≔3
Number of cable groups: N≔3

MISCELLANEOUS FACTORS
Earth ambient temperature: Ta ≔ 20.00 [ºC]
Load factor: lf ≔ 100%

THERMAL RESISTIVITY VALUES


Thermal resistivity of the earth: ρe ≔ 120.00 [ºC-cm/W]
Thermal resistivity of the cable insulation: ρi ≔ 450.00 [ºC-cm/W]
Thermal resistivity of the duct: ρd ≔ 600.00 [ºC-cm/W]
Thermal resistivity of the concrete duct bank: ρc ≔ 55.00 [ºC-cm/W]
2. CALCULATE THE CONDUCTOR AC RESISTANCE AT THE TARGET TEMPERATURE

(2.1) DC resistance adjusted to the operating temperature: Rdc_75C ≔ Rdc_25C ⋅ ⎛⎝1 + α25C ⋅ ⎛⎝Tc − TBase⎠⎞⎠⎞ = 42.392

Skin effect equations...


Rdc_75C
(2.2) Input to the Bessel approximation ( ks is an empirical factor): x ≔ ――― = 42.392
ks

11
(2.3) Bessel function approximation: Fsp (x) ≔ ――――― 2
⎛ 4 2.56 ⎞
⎜x + ―
x
− ――2 ⎟
⎝ x ⎠

(2.4) Skin effect coefficient: Ys ≔ Fsp (x) = 0.0061

Proximity effect equations...


Rdc_75C
(2.5) Input to the Bessel approximation ( kp is an empirical factor): x ≔ ――― = 42.392
kp

11
(2.6) Bessel function approximation: Fxp (x) ≔ ――――― 2
⎛ 4 2.56 ⎞
⎜x + ―
x
− ――2 ⎟
⎝ x ⎠

(2.7) By convention S is used instead of Dc : S ≔ Dc = 1.003

2 2
⎛ dc ⎞ ⎛ 1.18 ⎛ dc ⎞ ⎞
(2.8) Proximity effect coefficeint: Ycp ≔ Fxp x ⋅ ⎜―⎟ ⋅ ⎜――――+ 0.312 ⋅ ⎜―⎟ ⎟ = 0.0179
( )
⎝ S ⎠ ⎝ Fxp (x) + 0.27 ⎝S⎠ ⎠

(2.9) Combined skin and proximity effects: Yc ≔ Ys + Ycp = 0.02403

(2.10) AC resistance, at the operating temperature, in μΩ per foot: Rac_75C ≔ Rdc_75C ⋅ ⎛⎝1 + Ys + Ycp⎞⎠ = 43.41
3. CALCULATE THE THERMAL RESISTANCES

Thermal resistance of the cable insulation...

(3.1) Diameter over the insulation: Di ≔ dc + 2 ⋅ ti

⎛ Di ⎞
(3.2) Thermal resistance of the insulation: Ri ≔ 0.012 ⋅ ρi ⋅ log ⎜―⎟ = 0.4925
⎝ dc ⎠
Thermal resistance between the surface of the cable and the inner duct wall...

(3.3) Empirical factors [N-M, Table VII]: A ≔ 17 B ≔ 2.3 C ≔ 0.024

(3.4) Diameter circumscribing the 3 cables: D's ≔ 2.15 ⋅ Dc

Thermal resistance of the duct wall...


⎛ n⋅A ⎞
(3.5) Thermal resistance in the duct: Rsd ≔ ⎜――――――⎟ = 5.231
⎝ 1 + ⎛⎝B + C ⋅ Tm⎞⎠ ⋅ D's ⎠

dt
(3.6) Thermal resistance of the duct wall: Rd ≔ 0.0104 ⋅ ρd ⋅ n ⋅ ――― = 1.0407
Dod − dt
Thermal resistance of the earth and the concrete duct bank...
104
(3.7) Thermal diffusivity of the soil: α ≔ ―― 0.8
= 2.258
⎛⎝ρe⎞⎠

(3.8) Fictitious diameter where lf commence: Dx ≔ 1.02 ⋅ ‾‾‾‾


α ⋅ 24 = 7.508

2
(3.9) Loss factor: LF ≔ 0.3 ⋅ lf + 0.7 ⋅ lf = 1

⎛ ⎛ 2
⎛ 1 ⎞ ⎛ x1 ⎞ ⎛ 4 x1 ⎞ y1 ⎞ ⎛ x1 ⎞⎞
⎜ ― ⋅ ⎜―⎟ ⋅ ⎜―− ―⎟ ⋅ log ⎜1 + ―― ⎟ + log ⎜―⎟⎟
⎜⎝ 2 ⎟⎠ ⎝ y ⎠ ⎝ π y ⎠ 2 ⎝ 2 ⎠⎠
⎝ 1 1 ⎝ x1 ⎠
(3.10) Duct bank equivalent radius: rb ≔ 10 = 8.063

Lb
(3.11) Aspect ratio: u ≔ ―= 4.465
rb
⎛ ‾‾‾‾‾
2 ⎞
(3.12) Duct bank geometric factor: Gb ≔ log ⎝u + u − 1 ⎠ = 0.945

(3.13) Distance from duct 2 (hottest) to duct 1: d21 ≔ s1

(3.14) Distance from duct 2 (hottest) to duct 3: d23 ≔ s1


‾‾‾‾‾‾‾‾‾‾‾‾
2 2
(3.15) Distance from duct 2 to duct 1's image: d'21 ≔ ⎛⎝2 ⋅ Lb⎞⎠ + ⎛⎝s1⎞⎠

(3.16) Distance from duct 2 to duct 3's image: d'23 ≔ d'21

⎛ d'21 ⎞ ⎛ d'23 ⎞
(3.17) Mutual heating effect factor: F ≔ ⎜―― ⎟ ⋅ ⎜―― ⎟ = 93.16
⎝ d21 ⎠ ⎝ d23 ⎠
(3.18) Thermal resistivity of the earth, taking into account the concrete duct.

⎛ ⎛ Dx ⎞ ⎛ 4 ⋅ Lb ⎞⎞
R'e ≔ 0.012 ⋅ ρc ⋅ n ⋅ ⎜log ⎜―⎟ + LF ⋅ log ⎜――⋅ F⎟⎟ + ⎛⎝0.012 ⋅ ⎛⎝ρe − ρc⎞⎠ ⋅ n ⋅ N ⋅ LF ⋅ Gb⎞⎠ = 13.5151
⎝ ⎝ De ⎠ ⎝ Dx ⎠⎠
4. EQUIVALENT THERMAL RESISTANCE OF THE CIRCUIT

(4.1) Total effective thermal resistance: Rca ≔ Ri + Rsd + Rd + R'e = 20.28

5. CALCULATE CABLE AMPACITY


‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾
Tc − ⎛⎝Ta + ΔTinf⎞⎠ 3
(5.1) Cable ampacity: I≔ ―――――― ⋅ 10 A = 250 A
3 Rdc_75C ⋅ ⎝1 + Yc⎠ ⋅ Rca
⎛ ⎞
( 10 comes from the resistance being in μΩ per foot)

The rated current from NEC Table B.310.15(B)(2)(7) aluminum Detail 2 column LF = 100%, ρe = 120, IRated = 250A.

View publication stats

You might also like