Final Load Shedding 1

See discussions, stats, and author profiles for this publication at: https://www.researchgate.
net/publication/349573478
The art of load shedding and online applications in a power system under an
emergency state
Technical Report · February 2021
CITATIONS READS
0 9
1 author:
Mohamed EL-Shimy
Ain Shams University
228 PUBLICATIONS 930 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
Economics of Variable Renewable Sources for Electric Power Production - a new textbook View project
Reliability, availability, maintainability, and adequacy View project
All content following this page was uploaded by Mohamed EL-Shimy on 26 February 2021.
The user has requested enhancement of the downloaded file.

M. EL-Shimy, “The art of load shedding and online applications in a power system under an emergency
state”, Electrical Engineering Portal (EEP), Feb. 24, 2021. Available at: https://electrical-engineering-
portal.com/load-shedding-online-applications-power-system-emergency-state
The art of load shedding and online applications in a power

system under an emergency state
By
Prof. Dr. Mohamed EL-Shimy
1. Introduction
Load shedding is a controllable reduction of a predetermined amount of the load
power consumption according to specific shedding criteria. The predetermined amount
of load to be shed is traditionally determined according to an analysis of the dynamic
security for a set of contingencies. In this case, look up tables are prepared and the
implementation of the load shedding is performed according to them. A shedding look-
up table is dependent on the operating conditions of the systems, its available topology,
the available reserve, and the contingency. In addition, a look-up table must be updated
for changes in the system. A look-up table determined for a given operating conditions
and a specific set of single contingencies might be not valid in the situation of multiple
contingencies such as cascaded outage of lines1.
The main objective of the load shedding is correcting an abnormal system state
(see section 2) to either the normal state or the alert state. For example, the balance of
the generation and the load during abnormal operating conditions is a way for keeping
the system stability. The load shedding can be applied manually or automatically. In
addition, the automatic load shedding has many philosophies. For example, the
automatic reactive load shedding depends on the value of the absolute frequency as a
shedding criterion while with the proactive load shedding not only the absolute
frequency is used as a shedding criterion but also the rate of change of frequency or
even the acceleration of frequency.
Based on its definition, the load shedding is a corrective action by which system
overloads can be relieved. The system overloads can be classified according to its root
causes as,
1. Active power shortage: in this case the available power production sources or
generators are insufficient to meet the system demand. In this case, the main
1
Hafiz, H. M., & Wong, W. K. (2004, November). Static and dynamic under-frequency load shedding: a comparison. In
Power System Technology, 2004. PowerCon 2004. 2004 International Conference on (Vol. 1, pp. 941-945). IEEE.
Page 1 of 14
symptom of the system overload is a drop in the average system frequency. The
shedding criterion in this case is the value of the system frequency. In the
frequency drops below a pre-determined value, the load shedding is activated.
This is called Under Frequency Load Shedding (UFLS).
2. Reactive power shortage: in this case the available capacity is capable to meet the
load active power demand while the voltage at specific locations in the network
is too low for appropriate power flow and load requirements. The shedding
criterion in this case is the violation of the bus voltage magnitudes. For this
purpose, it is called Under Voltage Load Shedding (UVLS).
The UFLS and UVLS schemes can be considered as a special protection or wide
area protection systems that attempt to minimize the impact of disturbances and prevent
either brownouts or blackouts in power systems. Both types of load shedding should
have shedding algorithms. A load shedding algorithm defines the method(s) by which
several loads will be switched-off (and switched-on again) automatically to keep the
power consumption below a defined secure level.
Based on their time-frames and magnitudes of the associated changes in the
operating conditions, system overloads may be also classified as,
1. Sudden and large changes: in this case a disturbance causes to the system to
rapidly move from the initial stable operation conditions and the time needed to
take appropriate corrective actions before a possible system collapse is very
small. In this case, fully automated, fast, and appropriate corrective actions must
be taken.
2. Slow and small changes: in this case a disturbance causes slow changes with
small amplitudes. Consequently, manual corrective actions are possible.
Based on this classification, the appropriate load shedding scheme can be

determined.
The UVLS schemes are integrated into utility electrical systems to operate as a last
resort for the controllable shedding of specific amounts of loads at specific locations in
the grid. This action can prevent the loss of a large amount of the load or the entire load
due to uncontrolled cascading events. In contrast, the UFLS is designed for use in either
the emergency or extreme states (see section 2). The main objective is to stabilize the
Page 2 of 14
balance between the available generation and load before or after an electrical island
has been formed i.e. the UFLS drops enough load to allow the frequency to stabilize.
Therefore, the UFLS helps to prevent the complete blackout, and allows faster system
restoration in case of islanding. Typically, an UVLS responds directly to voltage
conditions in a local area. The goal of a UVLS scheme is to shed load to restore reactive
power relative to demand, to prevent voltage collapse and to contain a voltage problem
within a local area rather than allowing it to spread in geography and magnitude. Load
shedding is generally applied in steps. In each step, a specific amount of the load is
dropped. If the first load-shedding step does not allow the system to rebalance, and
voltage continues to deteriorate, then the next block of the load is dropped. The UFLS
operates in the same manner.
2. Power System Security and System States

Power system security2 may be defined as the continuous ability of the power
system to keep all the system limits not violated with minimum interruption to the
supplied loads. The main target of the power system security is to keep the system intact
under normal and disturbed conditions. Therefore, the successful security system
should minimize the impact of disturbances on the operation, economics, and power
quality of power systems. In addition, an acceptable system security level guarantees
the immunity of power system to disturbances and makes the system defensive.
Therefore, secure operation of power systems requires the integration of all practices
designed for keeping acceptable system operation when components fail.
Power system security covers both static and dynamic phenomena. Therefore, the
security analysis is usually categorized to static (or adequacy) and dynamic security3.
The static security considers the impact of static or slow changes in the system limits
while the dynamic security considers the impact of disturbances (or contingencies) on
the system. The core definition of the dynamic security and stability is the same, but the
security is a wider term than stability. The stability is defined as4 “the ability of an
2
Wood AJ, Wollenberg BF. Power generation, operation, and control. Canada: John Wiley & Sons; 2012.
3
Pavella M, Ernst D, Ruiz-Vega D. Transient stability of power systems: a unified approach to assessment and control:
Springer Science & Business Media; 2000.
4
Kundur P, Paserba J, Ajjarapu V, Andersson G, Bose A, Canizares C, et al. Definition and classification of power system
stability IEEE/CIGRE joint task force on stability terms and definitions. Power Systems, IEEE Transactions on.
2004;19(3):1387-401.
Page 3 of 14
electric power system, for a given initial operating condition, to regain a state of
operating equilibrium after being subjected to a physical disturbance, with most system
variables bounded so that practically the entire system remains intact”; however,
“Security not only includes stability, but also encompasses the integrity of a power
system and assessment of the equilibrium state from the point of view of overloads,
under- or over- voltages and underfrequency”5.
The system limits define the normal operation of power systems. These limits or
constraints can be classified into two categories; the equality and inequality constraints.
In addition, the system limits may be classified according to their origin into intrinsic
limits and operating range limits. The equality constraints basically represent the load
flow equations while the inequality constraints represent the allowable range of
acceptable operation of various components in the system. In fact, the intrinsic and
operating range limits elaborates the inequality constraints associated with a specific
component. The intrinsic limits of an equipment are determined basically from the
design and characteristics of the equipment. The operating range limits are generally
less than the intrinsic limits and they are limited by the fulfilment of the overall
operational requirements of the system.
For example, consider a simple hypothetical system where an off-grid generating
plant supplies a load center via a short transmission line with negligible impedance. The
generator is capable of producing a voltage magnitude at its terminal in the range 85%
- 115% while the load requires a voltage magnitude in the range 95% - 105%. In this
case, the generator voltage limits present the intrinsic limits of the generator and they
are mostly related to its design. Successful operation requires that the voltage magnitude
at the load bus should not be violated. Therefore, the operating range limits of the
generator bus-voltage magnitude becomes equal to the load requirements (i.e. 95% -
105%). It is worthy to be mentioned that the 95% - 105% voltage limits present an
intrinsic limit as viewed from the load perspective. It is also important to know that the
operating range limits should not violate the intrinsic limits of any component within a
system. Otherwise, the system will be incapable of fulfilling the operational
requirements. Both intrinsic and operating range limits are not absolute constants. The
intrinsic capability limits usually decline with time due to degradation of the equipment.
5
Machowski J, Bialek J, Bumby J. Power system dynamics: stability and control: John Wiley & Sons; 2011.
Page 4 of 14
For example, the annual output degradation rate of PV systems is about 0.7%6. The
degradation may be attributed to the aging, operational stresses, and maintenance
quality. The operating range limits are also variable. For example, the ampacity (or
ampere capacity or current limits) of a cable are highly dependent on the temperature of
its surroundings. The ampacity limits are usually increased during the winter and
decreased during the summer. This is for avoiding over-temperature of the cable
insulation.
Recalling that in the normal operation of a power system, all the inequality and
equality constraints of the system are satisfied. In addition, the system security requires
a minimum available, reserve margin7. Power system security may also be defined as
the ability of the system to withstand credible contingencies without violating the normal
operation limits. A system operating under normal conditions is also said to operate in
the normal state. The security strength of the system is usually defined by the maximum
number of time-independent, and simultaneous disconnection of major system
components (such as generators, transformers, and line) without affecting the normal
operation of the system. Defining N as the minimum number of components required to
supply the system peak load. A system with an N-k security criterion is a system in
which k components may be simultaneously disconnected and the system will be able
to fulfil the normal state requirements in the post-contingencies time. Due to investment
constraints, power systems are usually designed according to the N-1 security criterion8.
The normal state is a secure state and a system operating in the normal state is said to
be intact.
Deviations from the normal state requirements cause the system operation to
move to insecure operating states. These deviations are mainly caused by contingencies
which are stochastic and unexpected events; however, the rate of contingencies may be
reduced for example by proper maintenance of components. Four insecure operating
states can be realized9. These states are the alert, emergency, extreme (or collapse), and
restoration states. Fig. 1 illustrates the main operational characteristics of these states,
and the interrelations between them. This figure is usually called the state transition
6
EL-Shimy M. Analysis of Levelized Cost of Energy (LCOE) and grid parity for utility-scale photovoltaic generation
systems. 15th International Middle East Power Systems Conference (MEPCON’12), Dec. 23-25, 2012, Alexandria, Egypt,
pp. 1- 7.
7
Wood AJ, Wollenberg BF. Power generation, operation, and control. Canada: John Wiley & Sons; 2012.
8
Eremia M, and Shahidehpour M, (eds). Handbook of Electrical Power System Dynamics: Modeling, Stability, and
Control, John Wiley & Sons, Inc., Hoboken, NJ, USA. 2013.
9
Fink LH, Carlsen K. Operating under stress and strain. IEEE Spectrum;(United States). 1978;15(3).
Page 5 of 14
diagram. Table 1 summarizes the characteristics of various states, some causes of state
transitions, and examples of the corrective actions for each state. The nomenclature used
in the table is illustrated in Fig. 1.
Fig. 1: State transition diagram
It is worthy to be mentioned here that an intact system is capable of providing

power balance. If the power balance could not be achieved, then the system becomes
not intact. Consequently, the synchronization of generators upsets. Therefore, the
system frequency protection devices will split the system into parts or islands; the
situation is called islanding and it is within the extreme state. The frequency and power
balance conditions in each island are different and abnormal. Therefore, system
blackout or unintentional brownout is usually detected. Delay in activating the possible
corrective actions while the system is in the emergency state may be the main cause of
the transition to the extreme state.
Generally, delayed or unsuccessful corrective actions during the operation in any
state may lead to severe consequences. Therefore, any security programme includes a
contingency analysis block. The contingency analysis is an investigative simulation of
hypothesized contingency for evaluating their impact on the system security. On the
other hand, the corrective action analysis is the process of figuring the possible actions
Page 6 of 14
that may be taken for overcoming the consequences of security upsetting contingencies.
The corrective action analysis works in two distinct modes. The first mode operates for
solving the problems found by the contingency analysis. Therefore, this mode is offline
while the second mode operates in real time operation for securing the system during its
real-time operation. The contingency analysis and the corrective action analysis require
the simulation of the system. Therefore, an accurate system model should be available.
In addition, the results obtained from these analyses are highly dependent on the
accuracy of the system model. Real-time models of a power system require centralized
real-time data collection available from local measuring and monitoring devices at each
system component. Therefore, telemetry is required for communication within the
system and for estimating its state.
Table 1: Summary of operating states and state transitions

E I N-1 N Intact Causes of transition Corrective actions
State system from normal state
Normal     Yes - -
Alert   x  Yes Constraints are near Preventive control.
their limits. Examples, startup of non-
Examples, reduction in spinning reserve or switching
the reserve margin or on compensators
bus voltage close to the respectively.
limits.
Emergency  x x x Yes Severe disturbances. Emergency control actions
Example, short-circuit (heroic measures).
faults or cascaded Example: fast fault isolation
outages. or operation of reclosers.
Extreme x x x x No Delayed or Heroic and remedial actions

unsuccessful such as load shedding,
emergency control generator trip, or intentional
actions. Severe power islanding for keeping power
imbalance. balance.
Restoration x  x x No Attempt of restoring Manual or automated
the system to the reinsertion of generators and
normal state or at least loads. The inequality
to the alert state. constraints should be kept
satisfied during the entire
restoration process.
Page 7 of 14
3. UVLS schemes
As a result of a disturbance, the voltage magnitude may drop to a preselected
level for a predetermined time. In such cases, the UVLS sheds a selected amount of
loads at selected locations in the system. The objective is then to prevent the widespread
of voltage problems or voltage collapse. System planning engineers perform numerous
studies (see Fig. 2) using the PV curves as well as other analytical methods for the proper
determination of the minimum amount of load that must be shed for securing the system
to retain voltage stability under credible contingencies.
Various voltage control equipments provide a defense against voltage collapse;
however, in the situations the system is subjected to severe disturbances or cascaded
outages, various voltage control and voltage restoration equipment may fail to restore
the normal state voltage range. In these situations, the load shedding provides an
effective corrective action for preventing either the voltage collapse or the islanding of
the system. This is illustrated in Fig. 3. The shown voltage problems are attributed to
the insufficiency of the reactive power sources needed for restoring an acceptable
voltage level.
Fig. 2: Voltage security assement algorithms
Page 8 of 14
It should be noted that system operators usually shed load as the last resort. Under
Voltage Load Shedding (UVLS) schemes, drops a load when the voltage gets too low.
The dropping of load will alleviate the system by eliminating the current flowing to the
dropped load. UVLS usually triggers distribution feeders to open when voltage of the
bulk electric system is around 90%. Definite time relays usually act when all three
phases show low voltage for around 10 seconds at 90% voltage magnitude, this would
be after some of the ULTC transformers have been acted. Certain critical customers
cannot be dropped from load despite the help it may present to the system. Critical
customers include hospitals or customers that would loose lots of revenue from being
dropped.
Fig. 3: Long-term (slow dynamic) voltage security and corrective actions
The order of various corrective events following a contingency is shown in Fig. 3.

The events started as a response to a contingency such as a forced outage of a generator
or a line. Erroneous operation or vandalism may also cause disturbances of the similar
impact.
Page 9 of 14
There are two main UVLS schemes; the decentralized (also called distributed UVLS)
and the centralized schemes. With a decentralized scheme, protection relays are
installed at the loads that are candidate to be shed upon severe voltage problems. As
voltage conditions at these locations begin to collapse, load assigned to that relay is
automatically shed. On the other hand, a centralized scheme has undervoltage relays
installed at key system buses within system areas. The trip information is transmitted to
shed loads at various locations. Both schemes require high-speed and reliable
communication for proper operation. Fig. 4 illustrates a typical distribution substation
with integrated UVLS and UFLS special protections.
Fig. 4: A typical distribution substation with integrated UVLS and UFLS special
protections
In this system, the UV relay is installed on the HV side of the transformer for correct
detection of the grid voltage. This is because the voltage on the secondary side is not a
real indicative of the grid voltage due to the actions of the ULTC transformer (or any
other load side voltage controllers). The UF relay is installed on the LV side of the
transformer because the transformer does not affect the frequency while the LVPT is
more economical in comparison with the HVPT. The main complication associated with
the setting UVLS is the inaccuracies associated with the PT and the UV relay. Therefore,
proper and secure setting should be chosen for considering the probable inaccuracies.
Page 10 of 14
4. UFLS – AN OVERVIEW
In this section an overview of the UFLS schemes and technologies will be presented.
Generally, the system frequency is a good indicator of the power balance and overload
conditions in a power system. The UFLS is the last resort for the treatment of serious
frequency declines in power systems when subjected to large disturbances. Under the
emergency state or the extreme state, the ability to maintain the power balance and
stabilize the frequency is directly related to the effectiveness of the employed UFLS
Strategy. An effective UFLS strategy should be capable of:
(1) Restrain the frequency decline,

(2) Restore the normal frequency,
(3) Minimize the load shedding,
(4) Minimize of the frequency recovery time,
(5) Minimize of the frequency fluctuations, and
(6) Provide the desired protection functions as economical as possible.
Typically, an UFLS scheme sheds the loads in several stages. In each stage a pre-
defined amount of the load is disconnected and the shedding of the load is continued till
the normal frequency (i.e. the power balance) is restored. This is illustrated in Fig. 5. In
50 Hz systems, the common practices of most utilities use 49.3 Hz (i.e. 1.4% drop in
the frequency) as the first frequency step, and between 48.5 and 48.9 Hz for the last
step10. In proper dynamic control of power systems, a sufficient time delay must be left
between the shedding of each load block. This is of major importance for monitoring
the correct impact of disconnecting a load block on the system frequency and for
avoiding excessive, unnecessary load shedding as well as for avoiding subjecting the
system to over-frequency conditions due to the over - shedding of loads. The proper
timing of shedding each load block should not only depend on the frequency but also
the rate of change of the frequency. This results in an adequate time separation between
the shedding of blocks. Using small load shedding blocks in conjunction with shedding
10
Hafiz, H. M., & Wong, W. K. (2004, November). Static and dynamic under-frequency load shedding: a comparison. In
Power System Technology, 2004. PowerCon 2004. 2004 International Conference on (Vol. 1, pp. 941-945). IEEE.
Page 11 of 14
timing based on the rate of change of frequency can be an effective way for prevention
of over-shedding.
The blocks of load shedding shown in Fig. 5 can be selected based on two criteria;
the static criterion and the dynamic criterion. Fig. 6 shows flowcharts describing the
logic of each criterion. In the static criterion, fixed load blocks are disconnected in each
load shedding stage. This criterion may reduce the impact and the effectiveness of the
load shedding, especially in large disturbance conditions that are associated with a steep
decline in the frequency. The dynamic load shedding is constructed for solving this
problem. In the dynamic load shedding the amount of load to be disconnected at each
shedding stage is dynamically selected based on the system frequency, the rate of
change in the frequency, the voltage, and the severity of the disturbance(s). In other
words, the amount of a load shedding block is a function of the magnitude of the power
imbalance.
Fig. 5: UFLS conceptual operation – blocks of load shedding in k-stages scheme
There are three main methods for the implementation of UFLS strategies. These
methods are
Page 12 of 14
(a) (b)
Fig. 6: UFLS criteria; (a) Static (semi-adaptive) UFLS; (b) Dynamic (adaptive)
UFLS
1. The traditional method; when the frequency is lower than the first setting value,
the first level of load shedding will be implemented. If the frequency continues
to decline, it is clear that the first load shed amount is insufficient. When the
Page 13 of 14
frequency is lower than the second setting value, the second stage of load
shedding is then implemented. If the frequency continues to decline, the further
load shed stages are activated until the normal frequency value is restored. The
traditional method follows the static shedding criteria and the amount of load
shedding per each shedding stage is determined based on the analysis of the worst
possible expected events. Therefore, for less severe events, the first stage of
shedding may result in an over-shedding and may also cause over-frequency
problems
2. The semi-adaptive method; to some extent, this method is similar to the
traditional method; however, the specific amount of load to be shed is determined
in terms of the measuring value of the rate of change of frequency.
3. The self-adaptive method follows the dynamic shedding criterion for more
accurate estimation of the proper amounts of the load to be shed in each stage and
the timing of each stage.
Page 14 of 14
View publication stats

Final Load Shedding 1

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

Final Load Shedding 1

Uploaded by

Copyright:

Available Formats

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

Technical Report · February 2021

Reliability, availability, maintainability, and adequacy View project

The user has requested enhancement of the downloaded file.

The art of load shedding and online applications in a power

Based on this classification, the appropriate load shedding scheme can be

2. Power System Security and System States

Fig. 1: State transition diagram

It is worthy to be mentioned here that an intact system is capable of providing

Table 1: Summary of operating states and state transitions

Extreme x x x x No Delayed or Heroic and remedial actions

Fig. 2: Voltage security assement algorithms

Fig. 3: Long-term (slow dynamic) voltage security and corrective actions

The order of various corrective events following a contingency is shown in Fig. 3.

(1) Restrain the frequency decline,

Fig. 5: UFLS conceptual operation – blocks of load shedding in k-stages scheme

View publication stats

You might also like