Power Network

Location:

India

Posted:

November 12, 2012

Contact this candidate

Resume:

Design and Implementation of Scenarios for

Evaluating and Testing Distance Relays

D. Ristanovic, Student Member, IEEE, S. Vasilic, Student Member, IEEE, M. Kezunovic, Fellow, IEEE

This paper has two main objectives. First, to describe and

summarize the scenarios (special faults and normal operating

Abstract--This paper introduces a novel approach for

evaluating protective algorithms and testing protective relays. states) relevant for operation of distance relays. Second, to

The behavior has to be evaluated and tested for many scenarios in discuss their implementation using available network models

power system. The scenarios that encompass specific faults as well

and simulation tools. General fault events such as various

as a variety of the events in normal operating state are discussed.

types of shunt and serial faults including or excluding the fault

A set of power network models developed for studying distance

impedances are very well known and defined, and are not

relaying is introduced. Capabilities of these models, and some of

considered.

dedicated software tools such as Alternate Transient Program

(ATP), EUROSTAG and MATLAB'S Power System Blockset Various network models and simulation tools are utilized

(MATLAB/PSB), to implement described events are discussed. for comprehensive implementation of defined scenarios. The

An example of comparative simulation responses, by using

following models of actual power networks are used in

different programs for implementing the same model and event is

simulations: Stp-Sky and Nbelt-King section from Reliant

provided at the end.

Energy (RE) HL&P company developed earlier for testing

distance relays [6], and Glen Canyon-Flagstaff section from

Keywords--distance relays, power system modeling, power

system faults, electromagnetic transients, software tools. Western Area Power Administration (WAPA) company,

developed for evaluating fault location algorithms [7]. The

I. INTRODUCTION following software packages are used in the modeling: ATP

T [8], EUROSTAG [9], and MATLAB/PSB [10,11].

HE problem of selecting the right transmission line relay

The paper is organized as follows. Selected models of actual

settings or the best relaying algorithm to ensure proper

power networks are given in section II. Brief descriptions of

operation for all possible scenarios is the problem of making

specific fault events as well as events in the normal operating

sure that both dependability and security of the relay operation

state are given in section III. Section IV analyses the

are maintained in all cases [1]. To solve this problem,

capabilities of selected power network models and available

extensive evaluation of relaying algorithms and testing of

simulation tools for implementing defined events. An example

protective relays is performed. Evaluating and testing distance

of comparative simulation results obtained by ATP and

relays may require modeling of the power network and

MATLAB/PSB for the same model and fault event is shown in

performing complex simulation for a variety of scenarios.

section V. The conclusion is given at the end.

These scenarios encompass many events, including faults and

normal operating states. Diversity of power network models

II. POWER NETWORK MODELS

and software packages is important in proper implementation

of the scenarios. Three typical 345kV power system sections, two from RE

Some definitions of the scenarios that may cause unwanted HL&P and one from WAPA, were modeled for relay testing

operation of transmission line relays were analyzed in [2]. and simulation studies. The modeling involved two major

Practical method to evaluate values measured by distance steps: first, obtaining reduced Thevenin equivalent circuits for

relays in case of mutual coupling of parallel lines was all the boundary buses; second, detailed modeling of all the

described in [3]. Faults with time variant impedances, as well elements of the studied section. The reduced network

as the effect of load variations under normal conditions were equivalents were obtained by using the load flow and short

analyzed in [4]. Problems related to power swings were circuit data offered by RE HL&P and WAPA. The models

discussed in [5]. were verified using both the steady state and transient state

results. Figs. 1-3 show one-line diagram of the reduced

This study was supported by an Army/EPRI contract # WO 8333-05 equivalent for all three used sections.

between EPRI and Carnegie Mellon University, and has been carried out by

Stp-Sky section model (Fig. 1) has 9 buses and 11 lines,

Texas A&M University under the subcontract # 542***-***** titled "Self-

defined as lumped and distributed parameter models. Two

Evolving Agents for Monitoring, Control and Protection of Large, Complex

Dynamic Systems". important lines were modeled as frequency dependent, the rest

D. Ristanovic, S. Vasilic and M. Kezunovic are with the Department of

as frequency independent (constant parameter). Even though

Electrical Engineering, Texas A&M University, College Station, TX 77843-

the original system includes mutual coupling between some of

3128 USA (e-mails respectively: ******@**.****.***, ********@**.****.***,

*******@**.****.***).

the lines, this simplified model does not represent it. These reduced systems are convenient for producing fault

waveforms to be used for transient testing of distance relays.

HILL E9

SKY

III. AN OVERVIEW OF SPECIFIC SCENARIOS RELATED TO THE

FAULTS AND NORMAL OPERATION

MARION E8

This section describes some of the scenarios relevant for

studying operation of distance relays widely used today.

SPRUCE

Simulation results from the scenarios could be used in

HOLMAN E7

evaluation and tuning the existing and any new transmission

line relaying algorithms.

LHILL STP WAP

A. Specific Fault Events

Faults in reverse direction. Fault current direction can

change in one circuit when circuit breakers open sequentially

Z3 Z4 Z5

to clear the fault on the parallel circuit. A system configuration

that could result in current reversals is shown in Fig. 4. For a

DOW

E3 E4

fault on line L1, we may suppose that the circuit breakers do

not operate simultaneously. We assume that the circuit breaker

Fig. 1. Model of RE HL&P Stp-Sky Power Network Section

CB2 operates first, causing the direction of the current flow in

line L2 to reverse, before the circuit breaker CB1 opens. The

Nbelt-King section model (Fig. 2) has 6 buses and 6 lines,

change in current direction may cause improper operation of

described with frequency independent lumped parameter

permissive overreaching distance protection schemes and

models without mutual coupling.

directional ground-fault blocking schemes. Protection can see

the fault in the opposite direction to what was initially detected

(distance protection settings of these elements must exceed

150% of the line impedance at each terminal). The race

between the operating and resetting actions of the

WHART KING TMBAL E5

NBELT

overreaching distance elements at each line terminal can cause

the permissive overreach element to trip the healthy line.

Similar situation can occur in the directional ground fault

blocking scheme application.

Fig. 2. Model of RE HL&P Nbelt-King Power Network Section

Glen Canyon-Flagstaff section (Fig. 3) includes 3 buses and CB1 CB2

4 lines modeled as frequency independent distributed

Strong Weak

parameter lines, with mutual coupling between some of the Fault

lines. Only simplified representation of the model has been

shown in Fig. 3. The model also includes series capacitors CB3 CB4

with capacitor bank, discharge circuit, MOV protection, as S

well as surge arresters.

Fig. 4 Fault in reverse direction

PINNACLE

GLEN

FLAGSTAFF

PEAK

CANYON

Cross-country faults. They can occur between mutually

E1 E2

coupled lines (generally speaking between lines on the same

tower). A fault can occur, for instance, between phases A and

B but the phases belong to different lines on the same tower.

Fig. 3. Model of WAPA Glen Canyon-Flagstaff Power Network Section An example of the system configuration is shown in Fig. 5.

The situation becomes critical if the fault is near one of the

substations, for instance substation S. Protective relays on both characteristics. Weak infeed characteristics could also be

lines at substation R will detect A-B-G fault in the forward found when small generators are installed and connected to the

direction. At substation S, relay on L1 will detect A-G fault in system, or when some of the generators are occasionally off

the forward direction and relay on L2 will detect B-G fault in line. Several protection complications may occur due to the

the forward direction. If the fault location is moved away from weak infeed: there may be insufficient current contribution to a

the bus at substation S, the relays in substation R will also fault on the protected line for a relay to reliably detect a fault.

detect correct single-phase-to-ground faults. Condition shown In case of multiterminal lines with a weak source at one

in Fig. 5 may result in undesired tripping of all three phases of terminal as compared to the other terminals, protection at the

both lines at substation R (instead of a single phase tripping of weak source will not detect faults beyond the tap as

each line), and proper single phase tripping at substation S. successfully as relays at a strong source.

The undesired operation of the relays at substation R can occur

because they must rely on the local phase selection to L1

determine the fault type and which phase or phases to trip.

Fault

Zm0

A-G forw A-G forw

B-G forw B-G rev

Fault A-G

A-G forw A-G rev

B-G forw B-G forw

L2 Fig. 6 Parallel line out of service or faulted

R Fault B-G

Switching on-to-fault (recognizing fault after energizing the

line). This should occur following manual circuit breaker

Fig. 5 Cross-country fault

switching on-to a persistent fault. In such case, three pole

instantaneous tripping (and auto-reclose blocking) should

Faults caused by unsymmetrical supply. They are of interest

occur for any fault detected on the protected element. One

in certain operating conditions of transmission network.

complication is possible in case of switching on-to a fault close

Unsymmetrical supply may cause unbalanced conditions in

to the remote line end, when an underreach distance protection

transmission network and initiate, either proper or improper,

scheme is used. If the fault is not recognized as an immediate

operation of protective relays. Although conventional distance

fault after the circuit breaker closing, fault clearance will be

relays usually do not misoperate for this event, the event has to

unnecessarily delayed.

be taken into account for design and testing of any new

Evolving faults. They start as a single-phase-to-ground fault

relaying algorithm.

and then involve additional phases during the time that the

Faults with time variant fault impedance. The fault

initial fault is being cleared or during the circuit breaker dead

impedance variation is due to the variable impedance of the

time of the original faulted phase. Evolving fault may lead to

arc itself. This is due to the varying fault and environmental

difficulties in coordinating the ground-fault relays and

conditions that affect the arc formation and intensity. The

overcurrent relays.

effect of the remote infeed may further contribute to the

misinterpretation of the fault impedance measured by distance B. Events in the Normal State

relays. In that case a distance relay may operate for a forward Source voltage variation. It can occur due to various

external fault, or may not operate at all for an internal fault. changes in the power system network configuration. Response

Parallel line is out of service. This can cause unwanted of the automatic voltage control at the generators also

operation of distance relays. When overhead lines are contributes to the voltage variation. The result is the varying

connected in parallel or run in close proximity for either whole measured voltage at the relay location. Hence voltage

or part of their length, mutual coupling exists between the two measurements of protection relays are influenced by the

circuits. Typical application where the effects of mutual system voltage even though protected line may not be faulted.

coupling should be addressed is the case with parallel line out Load variation. It may generally have the major influences

of service and grounded at both ends (Fig. 6). For the case on protective relay setting and operation. Usually, load is

shown in Fig. 6, a ground fault at the remote bus may result in varying slowly and gradually due to the random changes in

incorrect operation of the distance ground fault elements for power consumption. In some situations, load change may be

zones 1 and 2. It may be desirable to reduce the distance huge such as in the case of on/off switching of large customers,

ground fault reach for zones 1 and 2 for this application. To or due to a fault or power outages. Load current level may

ensure adequate coverage an alternative reach setting may be influence fault detection sensitivity. In some cases, load

required. increase can lead to encroachment into a zone of distance

Weak infeed system. It may be considered whenever there protection that may further lead to unnecessary tripping.

are sources with high impedances in the network. Long line Line switching. It is an event in the normal state that

transmission systems with remote generation may have these

significantly contributes to the values measured by protective

IV. CAPABILITIES OF THE POWER NETWORK MODELS AND

relays. Switching of parallel or adjacent lines can cause events

SIMULATION TOOLS FOR IMPLEMENTING THE SCENARIOS

similar to source voltage or load variation. Sudden changes of

voltages and currents in some instances may lead to undesired The summary of simulation capabilities of available power

operation of protective relays. Similar phenomena may also network models: Stp-Sky, Nbelt-King and Glen Canyon-

occur if there is a transformer at the remote end of a Flagstaff, as well as available software tools: ATP,

transmission line (harmonic occurrence during on/off EUROSTAG and MATLAB/PSB for implementing defined

switching of the power transformer). scenarios is provided in Table I. Plus and minus signs in the

Line parameters variation. This can occur due to some table indicate whether or not the implementation is feasible.

external influence (temperature changes, humidity variations,

A. Power Network Models

soil resistivity variations, etc.). Line parameters are also

The following faults and events are not specific to any

dependable on the aging factor. If these variations are

particular network configuration and therefore there are no

significant, this can influence the values measured by distance

special requirements for implementing the models: faults with

relays.

time variant fault impedance, switching on-to fault, evolving

System frequency variation. It may occur due to the

faults, source voltage variation, load variation, line switching,

transient events in the network and due to the unbalance

and system frequency variation. Since there is no any specific

between the generated and consumed active power. Even

network configuration necessary to create conditions to

though the operation of distance relays should not be

simulate these events, all the network models are suitable.

influenced by this event, it could be interesting to use it to

Faults in reverse direction could be modeled in any of the

evaluate any new relaying solution.

available networks. Although there are no parallel, double

Power swings. Those are oscillations in the power flow,

circuit lines in Stp-Sky and NBelt-King models, it is possible

produced by various power system disturbances. They can be

to simulate current reversal through the network rings.

caused by faults, loss of synchronism across a power system,

Cross-country faults are related to the case of two different

or changes in the direction of the power flow due to the line

lines being on the same tower. This event could be modeled in

switching. Such disturbances can cause generators on the

any of the available networks, except NBelt-King Section,

system to accelerate or decelerate to adapt to the new power

since it does not have parallel lines.

flow conditions, which in turn leads to a power swing. The

Even though network models do not comprise detailed

result of a power swing may cause the impedance measured by

models of the generators (they include only simple source

a distance relay to move away from the normal load area and

representation), faults caused by unsymmetrical supply could

into one or more of its tripping characteristics. Stable relay

be created by disconnecting a particular phase in the network.

operation during the power swings is very important to avoid

undesirable relay tripping.

TABLE I SUMMARY OF THE EVENTS IMPLEMENTATION IN THE AVAILABLE POWER NETWORK MODELS AND SIMULATION TOOLS

Network Model Simulation Tool

Event Glen Power

Nbelt -

Stp - Sky Canyon - ATP Eurostag System

King

Flagstaf Blockset

Faults in reverse direction + + + + + +

Cross-country faults + - + + + +

Faults caused by unsymmetrical supply + + + + + +

Faults with time variant fault impedance + + + - - -

Fault

events The parallel line out of service - - - + - -

Weak infeed - - + + + +

Switching on-to fault + + + + + +

Evolving faults + + + + + +

Source voltage variation + + + + + +

Load variation + + + - + -

Normal Line switching + + + + + +

state

events Line parameters variation - - - - - -

System frequency variation + + + + + -

Power swings - - - + + +

Stp-Sky and Glen Canyon-Flagstaff-Pinnacle Peak are week Whenever one of the parallel lines is grounded at both ends,

infeed models. Nbelt-King model is not suitable to simulate three phases of the grounded line act similarly as the ground

the event. wires but with unsymmetrical position with respect to the

For parallel line (line on the same tower) out of service, the phases of the line in service. Simulating this event by

model should reflect the mutual coupling between parallel EUROSTAG or MATLAB/PSB, externally calculated value

lines. Nbelt-King model does not comprise (involve) parallel for zero sequence impedance has to be manually plugged into

line, and Stp-Sky model does not reflect explicitly this the system description.

relationship. Only Glen Canyon-Flagstaff model deals with Line parameter variation can not be implemented in any of

mutual coupling and consequently is suitable for modeling this mentioned software tools, because they do not provide

event. capabilities for continuous variation of line parameters due to

Even though line parameter variation is not related to any weather and/or aging conditions.

particular network configuration, available models do not

V. SIMULATION RESULTS

describe this variation as a function of external influences

(aging, temperature, pressure, etc). The network configurations An example of comparative simulation responses of ATP

provided in this paper are not relevant for modeling of such and MATLAB/PSB programs for the same faulted case is

event. shown to emphasize their characteristics. Stp-Sky power

Network model should comprise double-end infeeds in network model has been implemented in both programs, ATP

transmission lines and detailed model of the generators to and MATLAB/PSB. The model has not been implemented in

reflect power swings. Although all of the network models are EUROSTAG, since the voltage and current waveforms are not

of the ring type comprising double-end infeeds, none of them accessible, but only their effective values. Stp-Sky model is

provides detailed modeling of the synchronous machines, and used for simulating phase A-G fault on its Stp-Sky

therefore are not suitable for modeling this event. transmission line. Fault has been placed in the middle of the

line, with zero fault impedance and it starts at 20ms after

B. Software Tools

simulation begins. Voltages and current signals measured by

All three software tools, ATP, EUROSTAG and

protective relay at the Sky bus are shown in Figs. 7-8.

MATLAB/PSB, do include appropriate capabilities to

Similarity between responses in both programs is obvious.

simulate the following events: faults in reverse direction,

Some differences exist for the voltage signals during one/two

cross-country faults, faults caused by unsymmetrical supply,

periods after the fault starts, due to variable integration step

weak infeed, switching on-to fault, evolving faults, source

size in MATLAB/PSB contrary to fixed integration step size

voltage variation, line switching, and power swings.

in ATP.

For implementing cross-country faults, software packages

involve simulation of various types of faults occurring between

500k

different lines. Software packages allow phase switching and ATP

400k B

A C

they are suitable for simulating unsymmetrical supply. They 300k

comprises modeling of various types of fault at different 200k

subsequent time intervals at the same fault location, and they 100k

Vsky (V)

are suitable for implementing evolving faults. Instrument 0

-100k

transformer failure can be modeled by applying current or

-200k

voltage signals.

-300k

EUROSTAG enables implementation of various load -400k

models and is convenient for simulating load variation. -500k

0 .00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08

However, it can not simulate time varying fault impedance.

Tim e (s)

ATP and MATLAB/PSB provide only time invariant load and

fault impedances, and are not convenient for simulating these 500k

MATLAB

400k B C

events. A

300k

ATP and EUROSTAG can simulate continuous variation of 200k

the system frequency and may be used for implementing this 100k

event. MATLAB/PSB deals only with the time invariant

Vsky (V)

system frequency. -100k

Each software package provides detailed modeling of the -200k

synchronous machines, and so it may be used for -300k

-400k

implementing the power swings.

-500k

ATP, EUROSTAG and MATLAB/PSB provide detailed 0 .00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08

modeling of the overhead lines. Except ATP, other two Tim e (s)

packages can not compute a variation of the zero sequence line Fig. 7. ATP and MATLAB comparative voltage responses for the same fault

conditions

impedance due to the changes in the line configuration.

for microprocessor-based transmission line protective relays, Part I and

ATP II, IEEE Trans. Power Delivery, vol. 12, no. 1, pp. 134-156, Jan. 1997.

[2] IEEE Guide for Protective Relay Applications to Transmission Lines,

IEEE Standard C37.113-1999, Sep. 1999.

[3] L. M. Popovic, A practical method for evaluation of ground fault

1k current distribution on double circuit parallel lines, IEEE Trans. Power

Csky (A)

Delivery, vol. 15, no. 1, pp. 108-113, Jan. 2000.

[4] D. I. Jeerings and J. R. Linders, Ground resistance-revisited, IEEE

-1k

Trans. Power Delivery, vol. 4, no. 2, pp. 949-956, Apr. 1989.

-2k

[5] A. Mechraoui and D. W. P. Thomas, A new blocking principle with

-3k

phase and earth fault detection during fast power swings for distance

-4k

protection, IEEE Trans. Power Delivery, vol. 10, no. 3, pp. 1242-1248,

-5k July 1995.

0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08

[6] M. Kezunovic and Q. Chen, A novel approach for interactive

Tim e (s) protection system simulation, IEEE Trans. Power Delivery, vol. 12,

no. 2, pp. 668-674, Apr. 1997.

[7] A. Gopalakrishnan, M. Kezunovic, S. M. McKenna and D. M. Hamai,

M ATLAB

Fault Location Using the Distributed Parameter Transmission Line

Model, IEEE Trans. Power Delivery, vol. 15, no. 4, pp. 1169-1174,

2k Oct. 2000.

A B

[8] CanAm EMTP User Group, Alternative Transient Program (ATP) Rule

Csky (V)

Book, Portland, 1992.

[9] Tractabel-EDF, "EUROSTAG Software", Release Notes, ver. 4.1, Dec.

-1k

2000.

-2k

[10] The MathWorks, Inc., Using MATLAB, Natick, Jan. 1999.

-3k

[11] The MathWorks, Inc., "Power System Blockset User's Guide", Natick,

-4k

MA, Jan. 1999.

-5k

0 .00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08

VIII. BIOGRAPHIES

Tim e (s)

Fig. 8. ATP and MATLAB comparative current responses for the same fault

Dragan Ristanovic (S 01) received his B.S. in electrical engineering from

conditions

University of Belgrade in 1996, and is a M.S. candidate in electrical

engineering at Texas A&M University. His research interests are in protective

VI. CONCLUSION relaying and digital simulations in power systems.

This paper discusses distance relaying evaluation in many

Slavko Vasilic (S'00) received his B.S. and M.S. degrees in electrical

real situations in power networks. Specific fault events, as well

engineering from University of Belgrade, and is a Ph.D. candidate in

as events in normal operating states important for distance electrical engineering at Texas A&M University. His research interests are

relays operation have been described. The set of available neural networks, fuzzy logic, genetic algorithms, robust and adaptive control,

power network models which may be used for the events and their implementation in power systems control, protection and

monitoring.

modeling and simulation have been investigated. The

capabilities of these models and some of widely used

Mladen Kezunovic (S'77, M'80, SM'85, F'99) received his Dipl. Ing. degree

simulation tools for implementing a variety of fault events and from the University of Sarajevo, the M.S. and Ph.D. degrees from the

operating states have been studied. A brief summary has been University of Kansas, all in electrical engineering, in 1974, 1977 and 1980,

respectively. He has been with Texas A&M University since 1987 where he is

shown in a tabular form. One of the power network models has

the Eugene E. Webb Professor and Director of Electric Power and Power

been used to simulate the fault on one of the lines by using Electronics Institute. His main research interests are digital simulators and

ATP and MATLAB/PSB programs. Comparative results show simulation methods for equipment evaluation and testing as well as

application of intelligent methods to control, protection and power quality

similarity between signals obtained in both programs.

monitoring. Dr. Kezunovic is a registered professional engineer in Texas, and

a Fellow of IEEE.

VII. REFERENCES

[1] Power System Relaying Committee, Working Group D5 of the Line

Protection Subcommittee, Proposed statistical performance measures

Contact this candidate