Automated Power Substation Using Intelligent Load Shedding
Automated Power Substation Using Intelligent Load Shedding
1
[email protected]
2
[email protected]
Abstract—A Substation automation is provided with the functions of automatically controlling the
substation by using instrumentation and control devices locally as well as remotely. A Supervisory Control
& Data Acquisition (SCADA) System is deployed to control, monitor and protect these devices. Substations
are automated to reduce the human intervention and also to improvise the operating efficiency of the
overall system. Asset optimisation and reduction in the operating costs are the added advantages of the
substation Automation. Intelligent Electronic Devices (IEDs) are used to prevent disasters while supplying
energy and to assure human safety.
Intelligent Load Shedding (ILS) is useful in providing optimal and fast load shedding
techniques based on actual operating conditions that includes the type and the disturbances in the system.
Keywords-Substation Automation, SCADA system, Intelligent Electronic Devices (IEDs), Intelligent
Load Shedding.
I. INTRODUCTION
Substations play a vital role in design and operation of power system. They are provided with various access
points to the working parameters of the system that enables monitoring, controlling and secure operation of the
bulk power system. Automated Substations are deployed with huge number off instrumentations to facilitate the
functions like monitoring, controlling and communication. The parameters of power grid are voltage and current
where as in the data network it is the number of bits transferred. Both the grids are intertwined in operation and
hence improper functioning of one of the components of the homogenous system will affect the corresponding
grid. It is noteworthy that substations exhibit a vital role in the rigorous control of all the normal and abnormal
conditions of the system. Hence, it requires a great deal of attention while developing a fast, reliable and secured
data streaming communication system[1].
Substation automation systems(SAS) have been designed to improve efficiency of the control and the
communication schemes that are installed in the substations. SAS are strategically designed to provide a locally
computerized network at the substation to improvise the response time for any unpredictable event as quick as
possible. Main functions of substation automation consists of several sub-functions that are appropriately
interfaced [2]. The types of equipment in the substations are classified into two types : (i) primary equipment and
(ii) secondary equipment. Primary equipment comprises of transformer, switchgear etc, Secondary equipment
comprises of protection, control and communication equipment. Secondary equipment are further categorized into
3 levels by IEC 61850 standards. They are classified as station level, bay level and the process level equipment.
Fig.1 indicates the above three levels of equipment.
The ideology of the IEC-61850 standard is to support the substation functions by the communication of
Sampled values for current transformers and potential transformers.
I/O data for protection and control.
Configuration and Engineering data.
Supervision and monitoring signals.
Data to control-centre.
Time-synchronization signals, etc.
Many other functions such as metering, condition monitoring and asset management are also supported in IEC-
61850. Several functions may be implemented by a single IED or one single function may be hosted by multiple
IEDs. IEDs communicate with each other by information exchange mechanism of the IEC standards. IEC 61850
is categorised as the most suitable protocol for substation automation from the point of view of compatibility
among various hardware and software components that are supplied by different vendors.
Page 1 of 10
Fig.1: 3-level substation automation layout
Process level function is to extract information from sensors or transducers in the substation and send them to
upper level devices, i.e. bay level devices. Other main task of process level function is to receive the control
command from bay level device and execute it appropriately. Bay Level function is to collect data from same bay
and/or from different bay and act mainly on the primary equipment in its own bay. Station level functions are
classified into two categories. (i) Process related functions that process the data from multiple bays or substation
level. These functions are used to collect the substation data like voltage, current, power factor etc. from the bay
level devices and give the control command to the primary equipment. (ii) Interface related functions that enable
interface of substation automation system to the local station operator HMI (Human Machine Interface), to a
remote control center TCI (Tele Control Interface) or to remote engineering for the monitoring and maintenance
TMI (Tele Monitoring Interface) .
A bay is that part of a substation which contains extra-high (or high) voltage switching devices and the
connections of a power line , power transformer, etc,.to the busbars and also the protection ,control and
measurement devices for the power line, power transformer, etc,. It provides access points to the feeders.
Load shedding techniques that have been devised over a number of years have its own set of applications and
drawbacks. Various conventional load shedding techniques are discussed below.
A. Breaker Interlock Load Shedding
In this scheme, the interdependent circuit breakers are arranged to operate through hardwired trip signals from
an intertie circuit breaker or a generator trip. Even though the execution time of this scheme is fast, it has a number
of inherent drawbacks :
Load shedding is based on worst-case scenario.
Only one stage of load shedding is possible.
Mostly , more load is shed than required.
Any modifications to the system are expensive.
B. Under-Frequency Relay (81) Load Shedding
For setting up a frequency load shedding the guidelines are common to both small and large systems. Design
criteria considers only fixed load reduction at fixed system frequency levels. When the frequency set point is
attained and with the expiration of pre-specified time delay, the frequency relay trips one or more load breakers.
This cycle is thereafter repeated until the system frequency is recovered, e.g. , 10% load reduction for every 0.5%
reduction in frequency. This scheme is considered to be totally independent of system dynamics, hence the total
loss of the system is possible at times. Further drawbacks of this scheme are listed below .
Page 2 of 10
To prevent unnecessary tripping during frequency spikes there is an intentional time delay setting in
addition to the time it requires for the frequency to reach the predefined settings. Time delay might be
prolonged further because of the over-frequency condition that can occur during the fault.
When there is frequency decay and set time delay is attained, the frequency relay activates the first stage
of load shedding. If the load shed was insufficient, the frequency continues to decay which initiates the
next stage of load shedding. Each additional stage introduces more delays in the load shedding process.
The ultimate setting of each frequency relay is based on the most severe disturbance condition and the
most conservative generation and loading levels. This leads to excessive load shedding for most of the
conditions that are not so severe.
In response to a dip or rate-of-change in frequency, frequency relays are designed to operate a set of
fixed circuit breakers, independent of their actual operating load condition. The sequence of operation of
the circuit breakers may might be incorrect or not optimal.
Simulation of hundreds of transient stability studies is required to determine the frequency relay settings.
The main objective of the analysis is to find the minimum fault clearing time and to determine the minimum
required load shedding by the appropriate trial and error methods. The engineer performs the study and
learns the behaviour of the system and can intuitively predict the response of the system under various
operating conditions. The only study result that is utilised by the load shedding system is a set of frequency
relay settings. All other pertinent results of the analysis along with the engineer’s knowledge of the system
are lost.
C. Programmable Logic Controller-Based Load shedding
In this scheme, load shedding is based on total load versus number of generators online and detection of under-
frequency conditions. Each substation PLC is well programmed to initiate a trip signal to the appropriate circuit
breaker of the feeder to shed a pre-set sequence of load. This sequence is continued until the frequency returns to
a normal and stable level.
This scheme offers many advantages such as the use of distributed network based on power management system
and automated means of load shedding. However, monitoring of the power system is limited to a portion of
network with acquisition of scattered data. This drawback is compounded by the implementation of pre-defined
load priority tables at the PLC level that are executed sequentially, to curtail bulks of load regardless of the
dynamic changes in the loading of the system, generation or operating configuration. The system-wide operating
condition is often missing from the decision making process resulting in insufficient or the excessive load
shedding. Further , response time (time taken between the detection of the need for load shedding and action by
circuit breakers ) during the transient disturbance is often too long that requires even more load to be dropped.
An intelligent load shedding approach can provide faster and optimal load relief based on actual operating
conditions and knowledge of past system disturbances. This scheme requires a comprehensive understanding of
power system dynamics and process constraints along with the knowledge of system disturbance.
A. Pre-disturbance operating conditions
Total system load demand
Total system power exchange to the grid
Generation of each o-site unit
Spinning reserve for each on-site unit
Control settings for each running unit
Settings and loading conditions for all major rotating machines
System configurations (tie-line numbers, tie-line status and power transferring, bus-tie status and
flows, transformers and feeder status and loading , loading of each load , especially loading for the
sheddable loads, etc,.)
B. Post-disturbance operating conditions
Page 3 of 10
New system load demand
Remaining generation from on-site generation
Spinning reserve for each remaining unit
Time duration to bring up the spinning reserve
New system configurations
Status, settings and loading conditions of the remaining major rotating machines
Status of each sheddable load
C. Nature and duration of disturbances
Electrical and/or Mechanical faults
Complete or partial loss of power grid connection
Complete or partial loss of on-site generation
Load addition (impact)
Duration of disturbance and its termination ( Self-clearance, fault isolation, protection device tripping,
etc,.)
Subsequent system disturbances
D. System transient response to the disturbance
System frequency response (decay , rate-of-change , final frequency )
System voltage response
Rotor angle stability of each remaining unit
Operation of protective devices
Automated substations are being equipped with the modern data acquisition and monitoring system capable of
detecting and re-ordering online operating data and disturbance condition. Further, power system modelling and
simulation software tools have been remarkably improved to perform various system analysis from a simple load
flow study to more advanced studies like transient stability analysis. In recent studies, modern system analysis
software programs are designed as a component of larger power management system for performing system
analysis using real time data. Furthermore, techniques such as Neural Network(NN) , Genetic Algorithm,
Simulated Annealing (SA), Fuzzy Logic (FL), Expert Systems (ES) etc,. have been introduced in the field of
power system that can offer more effective problem solving techniques, knowledge representation and reasoning,
search, planning and action for some highly non-linear problems that cannot be often solved by using conventional
techniques.
With the introduction of such technological advances in power system, an automated and intelligent load
shedding system can be formulated to meet the following requirements :
Map a complex and highly non-linear , non-parametric load shedding problem to a finite space with a
limited number of data collection points.
Automatic recall of system configuration, the operating condition and system response to any
disturbance.
Pattern recognition capability to predict the system response to disturbances.
System knowledgebase trainable by user defined cases.
Self learning capability to new system changes.
Make prompt decisions regarding which loads to be shed based on the online status of sheddable loads.
Shed the minimum amount of load to maintain system stability.
Intelligent load shedding (ILS)scheme includes several basic functional blocks as shown in Fig.2 and illustrated
below.
Knowledgebase utilizes selected input and output data under different conditions based on off-line system
studies and simulations. It gives the output as system dynamic responses including frequency variation.
Advanced Monitoring continuously surveys the changes in system operating condition and calculates tie-line
flows, on-site generations, transformer and feeder loadings and evaluates status of sheddable loads.
Network models contain system topology, connection information and electrical properties of the system
components.
Trigger List is carefully compiled based on pre-specified system disturbance types.
Load Shedding Optimizer computes the optimal load shedding tables corresponding to system changes.
Distributed Control utilizes PLCs to rapidly execute the load shedding actions by detecting the disturbance
triggers from the system.
Page 4 of 10
Fig.2 : ILS Scheme Functional Block diagram
Substation automation involves the implementation of substation and feeder operating functions and the
applications ranging from SCADA and alarm processing to integrated volt-var control in order to obtain optimal
management of capital assets and improve the operation and maintenance efficiencies with minimal human
intervention. Distribution SCADA exhibit the basic functions like monitoring and control, report generation,
storage of historical data and several special functions are implemented in the substation automation scheme.
A. Integrated protection functions :Conventional approach and IED-based approach
In the conventional approach ,the relays had hardwired inputs from the instrument transformers and the
hardwires carried the trip signals to the circuit breakers. In the IED based approach , the information from
instrument transformers transfers to the relay via LAN and the relays also exchange information via LAN,
with the process bus becoming a reality, the breaker receive a trip signal via generic object-oriented
substation event (GOOSE) message travelling in the process bus as shown in Fig. 4(b)
In case of failure of a breaker, in the conventional protection wiring scheme, the hardwiring will carry the
trip signal to the backup protection scheme. On the other hand ,in the modern protection scheme, the backup
protection is initiated via LAN which reduces the wiring tremendously and it utilises alternate pathways.
Protection functions like automatic reclosing and bus differential schemes can also be implemented and
breaker failure can be dealt effectively.
The advantage is that separate protection relays can be avoided and thus the performance and the
reliability can be improved.
Page 5 of 10
Fig.3 (a) Protection via hard wiring (b) Protection via LAN
Fig. 4 (a) Protection via hard wiring (b) Protection via GOOSE messaging using LAN
B. Automation Functions
Substation automation application includes the following functions.
Page 6 of 10
transformer failure. Fig.5 shows the fault on transformer B. Circuit breakers 1 and 2 will trip and isolate
transformer B and the circuit breaker 3 will be closed to transfer the load on to the transformer A, which
will get overloaded and may have to shed load by opening circuit breaker 4. However, the load can be
transferred by connecting the line to the adjacent substation automatically by closing breaker 5.
3) Adaptive relaying
The special protection function of the substation automation system can play a vital role by altering the
settings of some of the relay IEDs. Adaptive relaying is the process of automatically altering the settings
of protective relay IEDs based on operating conditions of the system. This can be of much help to an
operator when there is a crisis in the system, with main lines and generators tripping.
For example, in the event of tripping of a critical generator, a line may get heavily loaded due to the
diversion of power from other sources. Under normal conditions, it will lead to tripping of this line which
may escalate an emergency situation. The operator at the control centre will detect the trip and can inform
the corresponding SA system of the event. The SA system can switch the appropriate relays to the new
settings to carry the required power until the crisis is over and then can switch back to the original settings.
Page 7 of 10
functions in the power system. Last, there is the utility enterprise where the integration of different control
centres can be performed and utility-wide data sharing and applications can be run.
1) Disturbance analysis
When IEDs are incorporated, disturbance analysis possess the added advantage of inherent capabilities
to record waveform at the fault and it can also time stamp the operational data that is measured. These
values can be used to recreate a disturbance sequence as the time stamping to the level of milliseconds
or less will help the operator and other personnel in the utility to assess the situation and take corrective
action for the next disturbance. For example, in case of an islanding of a section of the power system,
the faults, line tripping, overloading and under loading, and tripping of generators may happen in quick
succession. After the islanding the utility personnel must assess the situation and determine the correct
sequence of events. Time stamping of events, analog, or status changes immensely helps the utility to
recreate an event and do a thorough analysis of the situations that led to the islanding. The waveforms
captured can be used by the maintenance and protection departments to assess the severity of the damage
and take corrective action. Visualization and user interface can be implemented as client-side
application executing visualization application or user interface on client computers. Direct notifications
and customized reports are formatted using all available technology means for particular user groups.
Intelligent alarm processing is of utmost importance in a control room to help the operator from getting
confused by the battery of alarms triggered by an event.
Real-time equipment monitoring is an offshoot of the equipment condition monitoring discussed earlier.
Traditionally, power system equipment is loaded to the rated capacity under normal conditions, if the
equipment is condition monitored, the loading is based on actual conditions, rather than on conservative
assumptions.
For example, a transformer detected with “hot spot” will be loaded to a much lower value, due to the fear
of catastrophe. However, in a condition monitoring environment, transformer can be loaded to a higher
value by monitoring the true winding hot spot temperature.
The loadability can be derived, and it is reported that 5% to 10% additional loading can be achieved by
this process. This way the utility can squeeze more capacity out of the existing equipments and delay the
investment required for a new equipment.
ILS knowledge base and the computation engine resides in ILS server computer. The server interfaces with
the advanced real-time power system monitoring and simulation system that continuously acquires real-time
system data. Based on ILS calculations, the server dynamically updates the load shedding tables and downloads
that information to the distributed PLCs. Upon detection of any disturbance by the PLCs, load shedding is
initiated. The load circuit breaker will be tripped based on the pre-generated optimal load shedding tables. This
is shown in Fig. 5.
Page 8 of 10
Fig.5 Implementation diagram of ILS scheme
The benefits achieved by the implementation of ILS scheme as illustrated above are described below.
ILS will remarkably reduce the time required for load shedding with its master/local PLC configuration.
The local PLC detects the operation of circuit breaker almost instantly as the trip signal is hardwired to the
PLC. The local PLC then distributes this signal to all of the secondary/remote PLCs. In the system where
independent triggers (isolated case disturbance) were detected, the response time was measure around 20
ms. In order to distinguish between isolated case disturbance and subsequent contingencies, an intentional
delay of 50 ms was introduced prior to sending the trip signals to the load breakers. This is illustrated in
Fig. 6.
For instance, the shedabble loads are two 13 MW motors and four 7.5 MW motors. At first stage of the
existing frequency based load shedding system would have shed the one 13 MW motor and two 7.5 MW motors
when the system frequency fell to 47.2 Hz. If the under frequency condition persisted below 48 Hz for an
additional 750 ms, the remaining 13 MW motor and two 7.5 MW motors would be shed. This means a total
response time of about 1000 ms after the main breaker is tripped. This scheme ensured that sufficient load
would be shed to maintain system stability. Less critical loads, such as mine loads are not considered as shed
able loads since their loading can vary between 3 to 25 MW during normal mining operation. The original
frequency relay scheme for a specific disturbance is achieved by dropping Load 1 through Load 4 as shown in
Table I. For the same disturbance, ILS monitors pre-disturbance generation level of about 25 MW and spinning
reserve of about 19 MW. Based on these inputs and additional system data, ILS calculates the required load to
be shed equal to 7.0 MW, thereby, selecting Load 2 as the optimal load to be dropped.
Page 9 of 10
Load Id Group Quantity Operating Mw
Priority
Load 1 1 1 13
Load 2 1 2 7.5
Load 3 2 1 13
Load 4 2 2 7.5
Table II. Comparison between ILS and Frequency Relay Schemes (1 unit Trip)
A comparison between the ILS and frequency-relay scheme for one-generating unit trip is shown in Table 2.
The ILS will shed an optimal (minimum) load of 7.5 MW in 75 ms versus 24.4 MW in 250 ms when using the
original frequency-relay load-shedding scheme for one generator unit tripping. The comparisons show a
significant improvement in both load-shedding response time and reduction in the amount of load shedding
when utilizing the ILS technology even with the introduction of intentional delay.
In case of two generating units tripping at the same time, ILS monitors pre-disturbance generation levels of
about 50 MW and spinning reserve of about 16 MW. Based on these inputs and additional system data, ILS
calculates the required load to be shed equal to 34 MW. A comparison between ILS and the frequency relay
scheme for two generating units tripping is shown in Table III.
Table III. Comparison between ILS and Frequency Relay Schemes (2 unit Trip)
VI. CONCLUSION
Load shedding in automated substation system acts as the ultimate protective system from an overload induced
collapse. Hence the critical load preservation is primarily done with the use of circuit breaker interlock system,
under frequency relaying, and PLC-based schemes. Common drawbacks of these schemes are the insufficient
detailed pre-disturbance and post-disturbance data, real-time system configuration, type and duration of the
disturbances, as well as other crucial information. This paper has explained an intelligent optimal and fast load
shedding technology referred to as ILS in an Automated Substation. ILS uses a combination of 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. ILS system can perform
load shedding in less than 100 milliseconds from the initial occurrence of a disturbance.
VII. REFERENCES
[1] Won-Seok, C., et al. The study of substation automation systems in review. Computer Applications in Power, IEEE Transaction ,
1998.11(2): p.24-30
[2] Matsuda, S., et al. Issues overcome in the design and application of IEC 61850-compliant substation automation systems. In Advanced
Power System Automation and Protection (APAP), 2011 International Conference on. 2011, UK.
[3] Farrokh Shokooh and JJ Dai, “An Intelligent Load Shedding (ILS) System Application in a Large Industrial Facility,” IEEE IAS
Conference Record, 2005
[4] “Intelligent Load Shedding-Need For a Fast and Optimal Solution”, by Shervin Shokooh ,Tanuj Khandelwal,Irvine,CA,USA,IEEE 2005
[5] W. C. New, Load Shedding, Load Restoration and Generator Protection using Solid State and Electromechanical Under frequency Relays.
Philadelphia, PA: General Electric
[6] Load Shedding – An Application Guide, John Berdy, General Electric Company Electric Utility Engineering Operation, Schenectady,
N.Y.
[7] Kirkman, R. Development in Substation Automation Systems. In Intelligent Systems Applications to Power Systems, 2007. ISAP 2007.
International Conference on. 2007, Italy.
Page 10 of 10