An Intelligent Load Shedding (ILS) System

Application in a Large Industrial Facility

Farrokh Shokooh1, Fellow IEEE, J J Dai1, M. IEEE, Shervin Shokooh1, M. IEEE,

Jacques Tastet , M. IEEE, Hugo Castro1, M. IEEE, Tanuj Khandelwal1, M. IEEE, Gary Donner1, Fellow IEEE
2

Abstract ― Conventional methods of system load frequency. A brief discussion of these two dynamic loops is
shedding are too slow and do not effectively calculate the given below.
correct amount of load to be shed. This results in either
excessive or insufficient load reduction. In recent years, A. Excitation / Generator – Reactive Power – Voltage
load shedding systems have been repackaged using
conventional under-frequency relay and/or breaker During a fault condition, one of the direct effects of a fault
interlocks schemes integrated with Programmable Logic current is the drainage of reactive power from the system.
Controllers to give a new look to an antiquated load This reactive power is essential for the transfer of mechanical
preservation methodology. A truly modern and intelligent energy to electrical energy (and vice versa) in the rotating
load shedding system with a computerized power machines (generators and motors). After the fault clearance,
management system should provide fast and optimal load system is faced with partially collapsed flux energy in the
management by utilizing system topology and actual rotating machines and has to balance its generation and load
operating conditions tempered with knowledge of past levels while rebuilding its magnetic energy. During this time,
system disturbances. This paper demonstrates the need depending on the motor residual back emf, the system is also
for a modern load shedding scheme and introduces the faced with an additional reactive power demand from the
new technology of intelligent load shedding. Comparisons motor loads under reacceleration conditions.
of intelligent load shedding with conventional load
shedding methods are made from perspectives of system The voltage regulation and operating voltage of the overall
design, system engineering, project implementation, and system will directly depend on the amount of reactive power
system operation. A case study of the application of an that the generators could deliver to the system. On severe
intelligent load shedding scheme in a large industrial disturbances, the generators may automatically call upon its
facility is provided. over-excitation capability (ceiling voltage), which help in
recovering the system stability.
Index Terms ― Load Shedding, Intelligent Load
Shedding, Power System Monitoring and Simulation, B. Prime Mover / Generator – Real Power - Frequency
Frequency Relay, PLC, ILS.
Turbine governors and the type of prime movers also have
I. INTRODUCTION AND BACKGROUND a dramatic impact on the performance of the power system
during major disturbances.
In general, load shedding can be defined as the amount of
load that must almost instantly be removed from a power The frequency conditions of the overall system directly
system to keep the remaining portion of the system depend on the amount of real power that the generator prime
operational. This load reduction is in response to a system movers can deliver to the system. Also, the mechanical
disturbance (and consequent possible additional disturbances) energy available to help the generators prime mover ride
that results in a generation deficiency condition. Common through a fault or other disturbances plays an important role
disturbances that can cause this condition to occur include on the system behavior. This stored energy varies
faults, loss of generation, switching errors, lightning strikes, dramatically between that of a gas turbine, steam turbine, and
etc. hydro units. As a consequence, the performance of power
systems supplied by different types of prime movers and
When a power system is exposed to a disturbance, its governors will behave very differently under both steady-
dynamics and transient responses are mainly controlled state and transient conditions.
through two major dynamic loops. One is the excitation
(including AVR) loop that will control the generator reactive In addition to system upsets caused by faults, there are
power and system voltage. Another is the prime-mover loop, disturbances caused by switching surges or lightning strikes.
which will control the generator active power and system As an example, some switching disturbances can result in a
loss of generation or cause a system to separate from the
utility grid (system islanding condition). This condition can
1
Operation Technology, Inc., 17 Goodyear, Suite 100, Irvine, CA 92618, USA, (949) 462-0100, [email protected]
2
Technip, P.O. Box 1, 92973 Paris La Defense, France, + 33 (0) 1 4778 3664, [email protected]

IAS 2005 417 0-7803-9208-6/05/$20.00 © 2005 IEEE

cause the power system to collapse and will be adversely
impacted by inappropriate load reduction caused by an
improper load shedding scheme. For some switching
disturbances (that results in a loss of generation or system
islanding condition), the cascading effects may be of the
primary concern if the load shedding action is not set
correctly and/or timed properly.

Moreover, the type of disturbance impacts the dynamic

response of the prime mover. For instance, a short circuit at
the power station busbar may result in acceleration of the
generator prime mover. When this occurs the speed regulator
will then initiate closing of the fuel or gas inlet valve. After
the fault has been cleared, the turbines face the impact of the
load still connected. At this time their fuel or gas inlet valves
are closed resulting in difficult reacceleration conditions.

II. CONVENTIONAL IMPLEMENTATION OF

LOAD SHEDDING Fig. 1. Breaker Interlock Load Shedding Scheme

This section is a review of a number of load shedding In addition, the breaker interlock scheme has other inherent
techniques that have been previously devised. Each system drawbacks:
has its own set of applications and drawbacks. • Difficult to change load priority since the actions for
load shedding are hardwired and amount of load
A. Breaker Interlock Scheme shedding is calculated for the worst-case scenario.
• Only one stage of load shedding is available.
This is the simplest method of carrying out load shedding. • More loads are shed than necessary.
For example a source breaker would be interlocked via • The operation of this type of load shedding system
hardwired or remote signals to a set of load breakers that will most likely shut the entire industrial facility down
have been pre-selected to trip. When a generator breaker or a in a non-orderly way. This unplanned outage may
grid connection is lost for any reason, signals are result in processing equipment damage, reduced
automatically sent to load breakers to open. This system is equipment lifetime, or worse.
very fast since there is no processing required and all • Plant restarting may be delayed because of the
decisions about the amount of load to be shed were made requirement to shut down and then restart other remote
long before the fault occurred. facilities that have been affected by the loss of the
main facility, before the main facility can be started.
In Fig. 1, the load is supplied by a combination of a
generator and a power grid. A disturbance outside the facility B. Under Frequency Relay (ANSI Device 81) Scheme
causes the main breaker to operate and open. This would
isolate the system from the power grid causing the system Frequency relays do not detect disturbances but react to the
load to be supplied solely by the local generator (STG1). The disturbances. They detect either a rapid change in frequency
opening of the main breaker (MainBreaker) would signal the or gradual frequency deterioration and initiate staged
interlocked load breakers (LoadCB_1…n) to trip without any operation of interlocked breakers. When the first stage is
intentional time delay. This pre-selected breaker interlock list reached, the relay waits a predetermined amount of time, to
is typically determined without any knowledge of system avoid nuisance tripping, and then trips one or more load
transient response and is often too conservative, resulting in breakers. This is done to allow the frequency to recover. If
unnecessary load shedding. the frequency continues to decay, the relay will wait for the
next stage to be reached and after an additional time delay,
opens other load breakers. For the system shown in Fig. 2, the
frequency relay (FreqRelay) detects the first load shedding
stage and the interlocked load circuit breakers (LoadCB_1 to
LoadCB_i) are tripped accordingly, which will reduce the
real and reactive power demand on the generator. If the
frequency continues to decay then subsequent load shedding

IAS 2005 418 0-7803-9208-6/05/$20.00 © 2005 IEEE

stages will be reached and additional load breakers
(LoadCB_j to LoadCB_k) will be tripped until frequency
returns to normal.

Fig. 3 System Frequency Response Depicting Stage 1

Frequency Setting

During this period of time, the generators are faced with a

continuous overload condition. The fault causes the system to
be drained of reactive power before load shedding can relieve
the strain on the system and prevent further instability. Table
I outlines the settings for the various frequency relay stages
and the corresponding amount of load to be shed for each
stage.
TABLE I
LOAD SHED TABLE FOR CASE 2

Stage Frequency Delay MW Shed

(seconds)
Fig. 2. Under Frequency Relay (81) Load Shedding Scheme Stage 1 58.5 Hz 0.25 10
Stage 2 57.5 Hz 2.00 30
The load shedding schemes that use only conventional
frequency relays are primarily used for static control of If the amount of load shed in the first stage is insufficient
system loads. The frequency relay based load shedding and the frequency continues to decay, the system frequency
scheme has a number of inherent drawbacks as listed below. will reach the next set point and additional load shedding will
be initiated. The next block of load is shed with additional
1) Slow Response Time of Frequency Relays time delay. Hence each additional stage introduces delay in
the load shedding process.
Frequency relays must be slow to avoid nuisance trips. In
addition to the time it takes for the frequency to reach relay 2) Incorrect Load may be Dropped Causing Undesirable
settings, there is an intentional time delay setting to prevent Blackouts
nuisance tripping during frequency spikes and transient
deviations. Due to the fact that disturbances like three-phase The settings of a frequency relay are usually determined by
faults that prevent flow of real power, this time delay may be the most severe disturbance conditions and the minimum
further prolonged due to the overfrequency condition that can amount of local generation. This type of relay setting will
occur during the fault. result in excessive load shedding for other faults that are not
as severe. Low system frequency signal to the frequency
As shown in Fig. 3, when a three-phase fault occurs in this relay does not provide other pertinent information such as the
system, frequency initially increases. The fault is cleared type and location of the disturbance. In addition, the electrical
when the MainBreaker trips and opens. System separation distance between generators and loads are unknown.
eliminates utility real power support and as frequency decays,
underfrequency relay set point is reached at 0.15 seconds (9 In response to a frequency mandated operation, frequency
cycles). An intentional time delay of 0.15 seconds is relays operate a set of fixed circuit breaker, independent of
introduced in the frequency relay to avoid any nuisance their actual operating load. The operating load on the
tripping. The first set of breakers trip in 30 cycles after the breakers may be different than the loading used to determine
onset of the fault, resulting in a total load shedding time of frequency relay settings. Additionally, the sequence of
0.583 seconds. operation of the breakers may not be correct and/or optimal.
Modification cost is high since it may require field changing
of hardware.

IAS 2005 419 0-7803-9208-6/05/$20.00 © 2005 IEEE

2) Analysis Knowledge is Lost
The state-of-the-art load shedding system uses real-time
Conventional load shedding systems that rely solely on system-wide data acquisition that continually updates a
frequency measuring systems cannot be programmed with the computer based real-time system model. This system
knowledge gained by the power system designers. The produces the optimum solution for system preservation by
system engineer must perform numerous system studies that shedding only the necessary amount of load and is called
include all of the conceivable system operating conditions Intelligent Load Shedding.
and configurations to correctly design the power system.
Unfortunately, the engineer’s knowledge of the system, III. INTELLIGENT LOAD SHEDDING (ILS)
which is gained through the studies are not utilized fully.
Additionally, most data and study results are simply lost. This A. The Need for ILS
unavailability of information for future changes and
enhancement of the system will significantly reduce the Due to the inherent drawbacks of existing load shedding
protection system performance. methods, an intelligent load shedding system is necessary to
improve the response time, accurately predict the system
C. Programmable Logic Controller-Based Load Shedding frequency decay, and make a fast, optimum, and reliable load
shedding decision. This system must have the following
The use of Programmable Logic Controllers (PLCs) for capabilities:
automatic sequencing of load has become an important part • Able to map a very complex and nonlinear power
of substation automation in recent years. The application of system with a limited number of data collection points
PLCs in industrial load management and curtailment schemes to a finite space.
started in the early 1980s. However, it wasn’t until power • Automatically remember the system configuration,
management systems were combined with microprocessor operation conditions as load is added or removed, and
based PLCs that distributed fast load shedding systems the system response to disturbances with all of the
became a reality. system configurations.
• Recognize different system patterns in order to predict
With a common type of PLC-based load shedding scheme, system response for different disturbances.
load shedding is initiated based on the system frequency • Utilize a built-in knowledge base trainable by user-
deviations and/or other triggers. The circuit breaker tripping defined cases.
can be programmed based on the system loading, available • Adaptive self-learning and automatic training of
generation, and other specific logics. Each subsystem is system knowledge base due to system changes.
equipped with a PLC that is programmed to shed a preset
• Make fast, correct, and reliable decisions on load
sequence of loads. This static sequence is continued until the
shedding priority based on the actual loading status of
frequency returns to a normal condition. Modification of the
each breaker.
logic requires changing of the latter-logics that are
• Shed the minimum amount of load to maintain system
programmed in the PLCs.
stability and nominal frequency.
PLC-based load shedding scheme offers many advantages • Shed the optimal combinations of load breakers with
over the frequency-based scheme since they have access to complete knowledge of system dependencies.
information about the actual operating status of the power
system. However monitoring of the power system is limited In addition to having the above list of capabilities, ILS
to the sections of the system that are connected to the data system must have a dynamic knowledge base. For the
acquisition system. This drawback is further compounded by knowledge base to be affective, it must be able to capture the
the implementation of pre-defined load priority tables in the key system parameters that have a direct impact on the
PLC. These load reduction tables are executed sequentially to system frequency response following disturbances. These
curtail blocks of load until a preset load shedding level is parameters include:
achieved. This process may be independent of the dynamic • Power exchanged between the system and the grid
changes in the system loading, generation, or operating both pre and post disturbance.
configuration. The system-wide operating conditions are • Generation available before and after disturbances.
often missing from the PLC’s decision-making process • On-site generator dynamics.
resulting in insufficient or excessive load shedding. In • Updated status and actual loading of each sheddable
addition, the load shedding systems response time (time load.
period for which the load shedding trigger is detected by the • The dynamic characteristics of the system loads. This
PLC or relay up to the time when the trip signal is received includes rotating machines, constant impedance loads,
by the circuit breaker) during transient disturbances is often constant current loads, constant power loads,
too long requiring for even more load to be dropped. frequency dependent loads, or other types of loads.

IAS 2005 420 0-7803-9208-6/05/$20.00 © 2005 IEEE

B. Additional Requirements for ILS System load shedding tables in turn are downloaded to the
Distributed Controls that are located close to each sheddable
Some additional requirements must be met during the load. When a disturbance occurs, fast load shedding action
designing and tuning of an ILS scheme: can be taken.
• Carefully selected and configured knowledge base
cases. D. Implementation Configuration of ILS
• Ability to prepare and generate sufficient training
cases for the system knowledge base to insure ILS knowledge base and computation engine reside in an
accuracy and completeness. ILS server computer. The server interfaces with an advanced
• Ability to insure that the system knowledge base is real-time power system monitoring and simulation system
complete, correct, and tested. that continuously acquires real-time system data. Based on
• Ability to add user-defined logics. ILS calculations, the server dynamically updates the load
• Ability to add system dependencies. shedding tables and downloads that information to the
• To have an online monitoring system that is able to distributed PLCs. Upon detection of any disturbance by the
coherently acquire real-time system data. PLCs, load shedding is initiated. The load circuit breakers
will be tripped based on the pre-generated optimal load
• The ability to run in a preventive and predictive mode
shedding tables. This is shown in Fig. 5.
so that it can generate a dynamic load shedding table
that corresponds to the system configuration changes
and pre-specified disturbances (triggering).
• A centralized distributed local control system for the
power system that ILS system supervises.

C. Function Block Diagram of the ILS

In Fig. 4, the system knowledge base is pre-trained by

using carefully selected input and output databases from
offline system studies and simulations. System dynamic
responses, including frequency variation, are among the
outputs of the knowledge base.
Fig. 5. ILS Implementation Diagram

E. Comparison with Conventional ILS Schemes

A comparison of ILS system response time with that of

frequency relay load shedding is illustrated in Fig. 6. As
shown, frequency relay load shedding will be delayed until
the system frequency drops below the relay set point (Stage
1). Additional load shedding will be needed if the system
frequency does not recover to normal (Stage 2). Thus the
total response time for the frequency relay based load
shedding is much longer than ILS system.

Fig. 4. Function Block Diagram of the ILS Scheme

The trained knowledge base runs in the background of an

advanced monitoring system, which constantly monitoring all
of the system operating conditions. The network models and
the knowledge base provide power system topology,
connection information, and electric properties of the system
component for ILS. The disturbance list is prepared for all
pre-specified system disturbances (triggers). Based on the
input data and system updates, the knowledge base Fig. 6. ILS vs. Frequency Relay Load Shedding
periodically sends requests to the ILS computation engine to
update the load shedding tables, thus ensuring that the ILS requires only one load shedding stage and has a much
optimum load will be shed should a disturbance occur. The faster response time (less than 100 ms in most cases).

IAS 2005 421 0-7803-9208-6/05/$20.00 © 2005 IEEE

 A comparison of ILS system response time with that of
PLC-based load shedding is illustrated in Fig. 7. The PLC TABLE II
LOAD SHED TABLE
based load shedding will take longer time to respond to the
fault due to lack overall system topology, calculation time,
and time delays associated with frequency relays. These Stage Frequency Delay (ms) MW Shed
examples demonstrate the advantages of ILS over the Stage 1 48 Hz 1000 28
conventional load shedding schemes. Stage 2 47.2 Hz 250 28

B. Types of Disturbances

The main disturbances considered for load shedding are:

a. Loss of generation due to electrical faults
b. Loss of generation due to boiler trip

1) One Generator Trips

Fig. 7. ILS vs. PLC-Based Load Shedding The loss of a generator has tremendous effects on the
system process since the electrical demand is approximately
92% of the generating capacity of the steam plant. For
IV. ILS CASE STUDY: INSTALLATION AND example, losing one STG unit reduces the generating capacity
IMPLEMENTATION IN A REAL INDUSTRY FACILITY by about 20%. As a result of this condition, the two mill
motors are shed. Operators then have to manually start up the
A. System Description diesel generators in order to restart the mill motors.

A working example of ILS system was recently installed at 2) Two Generators Trip
PT Newmont Batu Hijau, a mining plant in Indonesia. An
overall one-line diagram is shown in Fig. 8. This islanded When a fault occurs on the transmission line, two
system draws power from four 34 MW steam turbine driven generators (about 45% of the generating capacity) could be
generators (STG) and nine 5.1 MW diesel engine driven tripped by the transformer protection relays. The loss of two
generators (DG). The 11 kV generation plant supplies system STG units can potentially escalade very quickly to a total
load through two 150 kV transmission lines stretching 15 system shutdown if the proper amount of load is not shed
kilometers (10 miles). The voltage is stepped down to supply before the remaining system becomes unstable.
the distribution system at 33 kV. The entire system operates
at 50Hz. 3) Pulverizer and Boiler Trips

Under normal operation all STGs are online and The powerhouse steam boilers are fired with pulverized
maintaining an average operating load of 110 MW. The coal. Each boiler has two pulverizes that are capable of
average load per generator is between 25 and 30 MW. supplying only half of the steam capacity from each
Spinning reserve is provided by two DGs. When one STG pulverizer. One or both pulverizes can shut down when a
unit goes offline, about five to seven DGs are manually mechanical problem occurs, or the supply of coal is lost.
brought online to carry approximately 18 to 20 MW of plant When a pulverizer is shut down for any reason, the steam
load, respectively. generating capacity of the boiler is reduced with a reduction
in electric power generating capacity. The generating
Based on historical disturbances, an electrical fault on the capacity will not immediately fall to zero if both pulverizers
150 kV transmission lines would cause the generator units to are lost because of the residual capacity of the boiler. The
trip offline, which in turn triggered the existing frequency remaining generating capacity of each boiler during upset
relay load shedding scheme to operate. Due to the inherently conditions must be known so appropriate settings of the load
slow speed of this scheme, too much load was often dropped shedding system can be made.
resulting in significant impact on production with losses
averaging USD 200,000 per day.

The system utilized a multi-stage frequency based load

shedding scheme with a load shedding sequence as shown in
Table II.

IAS 2005 422 0-7803-9208-6/05/$20.00 © 2005 IEEE

 Fig. 8. Simplified One-Line Diagram for PT Newmont Nusa Tenggara

Fig. 9. ILS Communication System Architecture for PT Newmont Nusa Tenggara

IAS 2005 423 0-7803-9208-6/05/$20.00 © 2005 IEEE

C. Data Communication Architecture detected, the response time was measure around 20 ms. In
order to distinguish between isolated case disturbance and
As previously explained, the speed of operation and subsequent contingencies, an intentional delay of 50 ms
continually updated load shedding tables are what sets the was introduced prior to sending the trip signals to the load
ILS system apart from other types of automatic load breakers. This is illustrated in Fig. 10.
reduction systems. The electric demand for this mining
operation is fairly constant except for the small motor load
that can quickly fluctuate between 3 MW to 25 MW. For
the ILS to work correctly, this load swing must be included
in the continual load tabulation, so a data collection server
was located in the substation that supplied these loads.
This allowed the ILS to continually monitor these loads.

The ILS server is installed in the power plant control

room. To bring the data from the Data Collection Server
(located at the load substations), fiber optic cable was
chosen to be the most effective way to accomplish this
task. Radio communication and other methods were
evaluated, but discarded since running 15 kilometers of Fig. 10. ILS Response Time with Intentional Delay
fiber was feasible.
2) Optimal Load Shedding
Ethernet equipped smart meters or Intelligent Electronic
Devices (IEDs) were used for data acquisition. Online The shedabble loads are two 13 MW motors and four
system data combined with circuit breaker status and other 7.5 MW motors. The first stage of the existing frequency
pertinent information are passed to ILS server for based load shedding system would have shed the one 13
processing and calculating the following: MW motor and two 7.5 MW motors when the system
frequency fell to 47.2 Hz. If the underfrequency condition
• Total generation persisted below 48 Hz for an additional 750 ms, the
• Total load to shed for each triggering event remaining 13 MW motor and two 7.5 MW motors would
• Generation capacity be shed. This means a total response time of about 1000
• Total spin reserve ms after the main breaker tripped. This scheme ensured
• Minimum load to be shed for each triggering event that sufficient load would be shed to maintain system
• Optimal combination of circuit breakers stability. Less critical loads, such as the mine loads are not
considered as sheddable loads since their loading can vary
Using the above information, ILS performs all of the between 3 to 25 MW during normal mining operation.
calculations necessary to determine the optimum load
shedding tables. These tables are then downloaded to the The original frequency relay scheme for a specific
local PLCs every 500 ms. The load shedding trigger is disturbance is achieved by dropping Load 1 through Load
hardwired directly to the PLCs. This configuration 4 as shown in Table III. For the same disturbance, ILS
produces a total response time of less than 70 ms, which is monitors pre-disturbance generation level of about 25 MW
a significant improvement over the original system and spinning reserve of about 19 MW. Based on these
response time of 300 ms. In addition to an improved inputs and additional system data, ILS calculates the
response time; the optimal load is now shed. required load to be shed equal to 7.0 MW, thereby,
selecting Load 2 as the optimal load to be dropped.
D. System Improvements Achieved with ILS
TABLE III
LOAD SHED TABLE FOR FREQUENCY RELAY SCHEME
1) Reduced Response Time
Load Group Quantity Operating
ILS will significantly reduce the system load shedding
ID Priory MW
time with its master/local PLC configuration. The local
Load 1 1 1 13
PLC is able to detect the operation of a STG breaker
almost instantly because the trip signal is hardwired to the Load 2 1 2 7.5
PLC. The local PLC then distributes this signal to all of the Load 3 2 1 13
secondary/remote PLCs. For this system where Load 4 2 2 7.5
independent triggers (isolated case disturbance) were

IAS 2005 424 0-7803-9208-6/05/$20.00 © 2005 IEEE

 A comparison between ILS and the frequency relay ACKNOWLEDGMENT
scheme for one generating unit trip is shown in Table IV.
ILS will shed an optimal (minimum) load of 7.5 MW in 75 The authors greatly appreciate and acknowledge the
ms vs. 28 MW in 300 ms when using the original engineering staff and management of PT Newmont Nusa
frequency relay load shedding scheme for one generator Tenggara, Indonesia for their contribution and support in
unit tripping. The comparisons show a significant installation and testing of the ILS system at Batu Hijau
improvement in both load shedding response time and mine.
reduction in the amount of load shedding when utilizing
ILS technology. REFERENCES
TABLE IV
[1] Edward Kimbark, Power System Stability, Wiley-
COMPARISON BETWEEN ILS AND FREQUENCY RELAY
SCHEMES (1 Unit Trip) IEEE, February 1, 1995.
[2] Warren C. New, Load Shedding, Load Restoration and
MW Shed Time (ms) Generator Protection using Solid State and
ILS 7.5 75 Electromechanical Underfrequency Relays, General
Frequency Relay 28 250 Electric.
[3] Tom Wilson, PLC Based Substation Automation and
In case of two generating units tripping at the same time, SCADA Systems; Selecting a Control System
ILS monitors pre-disturbance generation level of about 50 Integrator, Western Electric Power Institute, March
MW and spinning reserve of about 16 MW. Based on these 1999.
inputs and additional system data, ILS calculates the [4] Kevin Warwick, Arthur O., Ekwue and Raj Aggarwal,
required load to be shed equal to 34 MW. A comparison Artificial Intelligence Techniques in Power Systems,
between ILS and the frequency relay scheme for two The IEE, 1997.
generating units tripping is shown in Table V.
TABLE V
COMPARISON BETWEEN ILS AND FREQUENCY RELAY
SCHEMES (2 Unit Trip)

MW Shed Time (ms)

ILS 34 75
Frequency Relay 56 1000

V. CONCLUSION

Load shedding in industrial power systems serves as the

ultimate guard that protects the system from an overload
induced collapse. This critical load preservation is
normally done with the use of circuit breaker interlocks,
under frequency relaying, and PLC-based schemes.
Common drawbacks of these schemes include lack of
detailed pre- and post-disturbance data, real-time system
configuration, type and duration of the disturbances, as
well as other important information. This paper has
introduced an intelligent optimal and fast load shedding
technology referred to as ILS. ILS combines system online
data, equipment ratings, user-defined control parameters, a
knowledge base obtained from offline system simulations,
system dependencies, and continually updated dynamic
load shed tables. This system can perform load shedding in
less than 100 milliseconds from the initial occurrence of a
disturbance. ILS technology has been successfully
installed and operational at industrial facilities.

IAS 2005 425 0-7803-9208-6/05/$20.00 © 2005 IEEE