Citation
Implementation and application of a digital simulator for protective relay testing

Material Information

Title:
Implementation and application of a digital simulator for protective relay testing
Creator:
Hamai, Daniel Masao
Publication Date:
Language:
English
Physical Description:
xi, 108 leaves : illustrations ; 29 cm

Thesis/Dissertation Information

Degree:
Master's ( Master of Science)
Degree Grantor:
University of Colorado Denver
Degree Divisions:
Department of Electrical Engineering, CU Denver
Degree Disciplines:
Electrical engineering

Subjects

Subjects / Keywords:
Protective relays -- Testing ( lcsh )
Digital computer simulation ( lcsh )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Bibliography:
Includes bibliographical references (leaves 107-108).
General Note:
Submitted in partial fulfillment of the requirements for the degree, Master of Science, Electrical Engineering.
Statement of Responsibility:
by Daniel Masao Hamai.

Record Information

Source Institution:
University of Colorado Denver
Holding Location:
Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
31159813 ( OCLC )
ocm31159813
Classification:
LD1190.E54 1994m .H36 ( lcc )

Downloads

This item has the following downloads:


Full Text
IMPLEMENTATION AND APPLICATION OF A DIGITAL
SIMULATOR FOR PROTECTIVE RELAY TESTING by
Daniel Masao Hamai B.S., University of Colorado, 1986
A thesis submitted to the Faculty of the Graduate School of the University of Colorado at Denver in partial fulfillment of the requirements for the degree of Master of Science Electrical Engineering 1994


This thesis for the Master of Science degree by Daniel Masao Hamai has been approved for the Department of Electrical Engineering
by
William R. Roemish


Hamai, Daniel Masao (M.S., Electrical Engineering) Implementation and Application of a Digital Simulator for Protective Relay Testing Thesis directed by Professor Pankaj K. Sen
ABSTRACT
Power utilities have used many testing techniques to evaluate protective relay performance prior to installing the relays in the system. The actual relay characteristics can only be determined by subjecting the relays to realistic system conditions. With the recent advances in computer hardware and software, utilities have an additional tool to simulate these conditions.
This thesis explores the implementation of a digital simulator for testing protective relays. The simulator injects current and voltage waveforms resembling actual fault conditions into the relays to evaluate relay performance. In contrast to other testing methods, the waveforms include both transient and steady-state components to provide more realistic test signals.
The simulator consists of a dedicated computer, digital-to-analog (D/A) converter modules, and high power amplifiers. The waveforms can be created by programs such as the Electromagnetic Transients Program (EMTP) or captured from actual power system faults by digital fault recorders (DFRs). In either case, the waveforms are represented digitally. The dedicated computer replays the waveforms through the D/A converters. The amplifiers
iii


transform the low voltage signals of the D/A converters into the currents and voltages to be applied to the relay. The amplifiers drive the relays at current and voltage levels usually found on the secondaries of the instrument transformers in the power system.
The related topics of protective relay advances, relay testing practices, and instrument transformer modeling in EMTP are also discussed in this thesis. As a specific application, the testing of the California-Oregon Transmission Project (COTP) protective relays with the digital simulator is examined.
The simulator can provide the relay engineer with added insight and understanding about the transient performance of the relays. The simulator can be extremely useful in validating relay settings for complex protection schemes.
This abstract accurately represents the content of the candidate's thesis. I recommend its publication,
Signed:!
Pankaj K. Sen
iv


Acknowledgmen ts
The support and assistance of everyone at Western Area Power Administration is deeply appreciated. Special thanks to Steve McKenna and Joel Bladow for their initial and on-going technical advice regarding the digital simulator project. Thanks to Matt Yakab for the hardware support in the Western Laboratory. In addition, thanks to Gary Zevenbergen for the use of several of his programs and models in this thesis.


CONTENTS
CHAPTER
I. INTRODUCTION............................................ 1
II. HISTORICAL REVIEW OF RELAY ADVANCEMENTS................. 7
III. PROTECTIVE RELAY TESTING................................12
Types of Relay Testing................................12
Historical Reviev of Transient Testing................18
IV. INSTRUMENT TRANSFORMER MODELING IN EMTP.................23
Current Transformer (CT) Models.......................24
Coupling Capacitor Voltage Transformer (CCVT) Models..28
V. DEVELOPMENT OF WESTERN'S SIMULATOR......................30
VI. PREPARATION OF SIMULATOR DATA...........................34
Data Preparation......................................34
IEEE Standard Common Format for Transient Data Exchange (COMTRADE)...................................39
Digital Processing Theoretical Considerations.........40
VII. USER INTERFACE DESCRIPTION..............................45
Explanation of the Main Menu..........................45
Default D/A Port Assignment Option..................46
Manual D/A Port Assignment Option...................47
Previous Case Option................................48
Set Program Parameter Option........................49
Explanation of the Run Menu...........................51
vi


Cycle Steady-State Option............................52
Cycle Steady-State and Execute Input File Option....52
View the Port Assignments Option.....................53
VIII. HARDWARE DESCRIPTION....................................54
Simulator Computer................................... 54
Current Amplifiers.....................................58
Voltage Amplifiers.....................................58
IX. DETAILS OF SOFTWARE IMPLEMENTATION......................60
Soft Start Feature.....................................60
Array Manipulation.....................................64
Input Data Array.....................................65
Digital-to-Analog Converter Arrays...................66
D/A Correction Routines................................71
X. DIGITAL SIMULATOR APPLICATIONS..........................72
California-Oregon Transmission Project.................73
Protective Relay Schemes.............................75
Relay Specifications.................................76
Transient Testing for Line Relay Evaluation..........82
Transient Testing for Relay Settings Verification...88
XI. CONCLUSION..............................................91
Future Work............................................93
Western's Experience...................................94
APPENDIX
A. EMTP MODEL USED FOR SIMULATOR FAULT INITIATION SIGNAL...95
vii


B. BPA PLOT PROGRAM ''PROFILE" FILE
96
C. TYPICAL INPUT SESSION USING THE DEFAULT PARAMETER
MENU OPTION............................................97
D. ABBREVIATED INPUT SESSION FOR DEFINING EACH D/A PORT...102
BIBLIOGRAPHY...................................................107
viii


FIGURES
Figure
1.1 Digital Simulator Layout and Corresponding Power
System Diagram............................................ 4
1.2 Major System Components of the Simulator................... 5
3.1 Steady-State Test Signals..................................13
3.2 Dynamic Test Signals.......................................15
3.3 Transient Test Signals................................... 16
3.4 Expansion of the Mho Characteristic........................17
4.1 EMTP CT Model............................................. 25
4.2 Test Circuit for CT Saturation Evaluation..................27
4.3 CT Transient Response for the EMTP Model...................27
4.4 EMTP CCVT Model............................................28
6.1 Digital Simulator Block Diagram............................35
6.2a 10 kHz Waveform Sampled at 200 kHz.........................42
6.2b 10 kHz Waveform Sampled at 9.5 kHz.........................42
6.2c 9.5 kHz Sampled Waveform Overlaid on the 200 kHz
Sampled Waveform...^......................................42
8.1 Front View of the Concurrent MC5450........................55
8.2 Connections between the Computer and Amplifiers............57
9.1 Time Frames of a Waveform..................................61
9.2 Voltage Waveform Showing Soft Start Feature...............63
9.3 Current Waveform Showing Soft Start Feature...............64
9.4 Flow of Data Through Circular Buffer.......................70
IX


10.1 Vicinity Map for COTP......................................74
10.2 COTP Line Relaying for the Captain Jack-01inda,
Olinda-Tracy, and Tracy-Los Banos Lines....................79
10.3 COTP Line Relaying for the Tracy-Tesla Line.................81
10.4 COTP System Diagram Showing Fault Locations.................85
10.5 Typical Currents and Voltages for COTP Relay Testing.......89
x


TABLES
Table
8.1 Convention for Mapping DA08F Channels to Program Ports....56
8.2 Current Amplifier Specifications..........................58
8.3 Voltage Amplifier Specifications..........................59
9.1 Array Description and Function............................65
10.1 Summary of Simulator Applications........................72
10.2 Captain Jack Olinda 500 kV Line External Fault
Summary Table............................................86
10.3
Captain Jack Olinda 500 kV Line Internal Fault Summary Table....................................
87


CHAPTER I
INTRODUCTION
Power utilities have always experienced difficulty in fully assessing protective relay performance before actually installing the relays in the power system. Relay operation usually occurs in response to random disturbances on the power grid caused by lightning, ice, wind, natural disasters, and human interaction with the system. Through the years, the relays themselves have changed. The physical components and the measuring principles used by protective relays have evolved with the advances in technology.
Despite the unpredictable environment, power utilities must do all they can to ensure reliable relay operation on the system. Customers demand it. Utilities have devised various testing methods for commissioning relays and validating their settings.
One of the earlier methods involved staged fault tests where actual faults were placed on the power grid to evaluate relay response. Later, the Transient Network Analyzer (TNA) was used which was a scaled power system model composed of discrete components.
Although both of these methods are still in use, several relay manufacturers and utilities are now turning to digital simulators to test relays in a laboratory environment. Computer workstations controlling high power amplifiers can simulate fault conditions for the relays. This thesis will explain the implementation of the
1


digital simulator developed by Western Area Power Administration (Western) and describe specific applications of the simulator.
Western Area Power Administration is a federal power marketing agency that was established in 1977 under the authority of the United States Department of Energy. Western markets power from fifty hydroelectric power plants operated by the Bureau of Reclamation, the United States Army Corps of Engineers, and the International Boundary and Water Commission, as well as a percentage of coal-fired power from a plant near Page, Arizona. Maximum operating capability in 1992 was more than 10,400 MW.
Western developed the digital simulator with the goal of testing relays back-to-back for commissioning and troubleshooting. Back-to-back testing refers to the practice of bringing both relay terminals of a transmission line into the laboratory environment, providing appropriate communication channels between the relays, and testing the relay scheme as an integral system. In this type of testing, transient current and voltage waveforms corresponding to fault conditions are injected into both line relays simultaneously to evaluate relay performance. The value of this testing has increased as the use of series capacitors, static var compensators, and single-pole tripping schemes have become more popular. In particular, the presence of series capacitors has proven to be a challenge to many protective relays; simulator testing of series compensated line relays can often aid in fine-tuning settings for secure and dependable operation.
2


The simulator uses digital data from sources such as the Electromagnetic Transients Program (EMTP) and digital fault recorders (DFRs). EMTP is extremely flexible in its capabilities since it can model transmission lines, transformers, series compensation, breakers, instrument transformers, fault switches...whatever is of interest in the power system. Whereas DFRs record faults on existing systems, EMTP can be used to model transmission systems which are in the process of being constructed. In simulator applications, EMTP can be used to model a fault or other system transient and provide the current and voltage waveforms of the two terminals of the transmission line. Internal faults, external faults, evolving faults, sequential tripping, and switch-onto-faults are typical examples of transient events that EMTP can be used for simulator applications. The simulator can replay digital fault recorder data but the user must ensure a proper sampling rate and accurate data representation (i.e. absence of noise and distortion in the fault waveforms). Western has found these factors to be major problems affecting DFR data. This thesis will concentrate on EMTP generated waveforms as the primary source of data.
A dedicated computer is used to send out the digitized waveforms through digital-to-analog (D/A) converters to the current and voltage amplifiers. With the proper gain and scaling, the waveforms are amplified to levels which would actually be present on the secondaries of the current transformers and voltage
3


transformers in the real power system. These currents and voltages are sent to the relays being tested. Figure 1.1 shows the simulator layout and the corresponding power system diagram.
TERMINAL #1
H
a?
o
TERMINAL #2

H
RELAY
VOLT. AMP. PH. A
VOLT. AMP, PH. B
VOLT. AMP. PH. C
CURR. AMP. PH. A
CURR. AMP. PH. B
CURR. AMP PH. C
EEID
SIMULATOR COMPUTER
COMM.
OELAY
VOLT. AMP.
PH. A
VOLT. AMP.
PH. B
VOLT. AMP.
PH. C
CURR. AMP.
PH. A
CURR. AMP.
PH. B
CURR. AMP.
PH. C
RELAY
Figure 1.1: Digital Simulator Layout and Corresponding Power System Diagram
A


The computer is a real-time data acquisition machine with built-in D/A converters. Two sets of three-phase voltage and current amplifiers are used for testing. A communication channel is provided with an adjustable time delay to simulate the travel time of the pilot signal.
Figure 1.2 identifies the major system components and flow of data through the simulator.
Figure 1.2: Major System Components of the Simulator
This diagram shows the digital input data from EMTP simulations or DFRs; the control of the digital data by the simulator program; the conversion of digital data to analog signals via the on-board D/A converters; the amplification of low voltage analog signals by the
5


current and voltage amplifiers; the transmission of the transient waveforms to the relays; and the recording of the relays' performance by a dedicated DFR.
The following chapters will discuss all of the components shown in Figure 1.1 in detail. Topics will include a brief history of protective relay development, relay testing techniques with an emphasis on transient testing, EMTP modeling of instrument transformers, details regarding actual implementation of the simulator, and applications of Western's simulator.
6


CHAPTER II
HISTORICAL REVIEW OF RELAY ADVANCEMENTS
The definition of a protective relay referred to in this thesis will be a device that detects abnormal power system conditions, such as short circuits, and initiates actions to isolate these conditions. In practice, the protective relay works in combination with circuit breakers that physically isolate the short circuit or fault from the rest of the system.
There are three major classifications of protective relays which refer to the level of technology at the time of their development: the electromechanical relay, the solid state relay, and the microprocessor (or digital) relay. At first, all relays were electromechanical relays. Two basic operating principles were used for this type of relay: electromagnetic attraction and electromagnetic induction. Most electromagnetic attraction relays operated by means of a plunger being drawn into a solenoid or an armature being attracted to the poles of an electromagnet [1]. Instantaneous overcurrent relays often used this method. Electromagnetic induction relays followed the principle of the induction motor and created operating torque by induction in a rotor [1J. Inverse time overcurrent relays often used electromagnetic induction. By the 1930s, electromechanical inverse time overcurrent relays with instantaneous elements were used to
7


coordinate relays with downstream power fuses [2]. In that time period, the two main relay manufacturers in the United States were the General Electric Company and Westinghouse Electric Corporation.
Two major performance issues affecting electromechanical relays were their tendency to overreach and overtravel. On an electromagnetic attraction relay, the direct current (dc) component of the fault waveform significantly affected its operation. With severe and long-lasting dc components, an overcurrent relay could pick up even though the steady-state value of the offset wave was less than the pickup setting [1]. This overreaching tendency had to be considered when setting the relay.
Since the electromechanical relays had moving parts, relay overtravel was an important issue. For example, in an induction-type overcurrent relay, an induction disk would operate in proportion to the current provided. If the current was suddenly removed, the induction disk would continue to move due to inertia. In closely coordinated applications, the engineer had to account for relay overtravel.
The introduction of the transistor by Bell Laboratories in 1948 eventually made a significant impact on the relaying industry [2]. By the late 1950s, the transistor had successfully been used to replace some of the moving parts in protective relays. However, due to the robustness of electromechanical relays and their long history of use, relays using transistor technology have not completely replaced electromechanical relays. In fact, most U.S.
8


utilities still have electromechanical relays on their systems in 1994.
The solid state relays are characterized by diode rectified bridges, level detectors, comparators, and timers. The power system current and voltages are often full-wave rectified to dc values which can be compared to threshold levels used for tripping decisions. Comparator circuits are also used to calculate the degree of coincidence between polarizing and operating quantities. Due to the use of transistors, extremely high speed operating times --- less than one cycle --- can be attained in some relay designs.
As the solid state technology progressed, relays began appearing with microprocessors. These relays are characterized by analog-to-digital (A/D) converters which sample and digitize the power system currents and voltages. The microprocessor functions as a signal processor and ultimately makes the operating decisions.
Features of the microprocessor based relays include lower cost, multiple functions, self-checking and monitoring diagnostics, and ability to change settings remotely via computer modems. These relays can often implement all of the commonly used pilot schemes used by relay engineers; the logic is inherent to the relay and the engineer simply chooses the scheme through dip switches or menus on a monitor. As opposed to electromechanical relays where relay integrity is tested during maintenance intervals or, undesirably, during an actual fault, the microprocessor based relay can automatically run self-checking tests during its idle time. In
9


most designs, a program routinely runs every few minutes to check the power supply, EPROMs, settings integrity, and other hardware. Another major advantage of the microprocessor relay is the ability to call the relay by modem and remotely change the relay settings based on power system conditions. Most of the microprocessor relays, however, cannot match the operating times of the fastest solid state, analog relays.
There may be a downside to the high technology relays. Typically, electromechanical relays were left in operation for thirty year lifetimes. Years of consistent operation by specific types of relays were attained. The same relays were applied in so many places that the relay engineer gained a high level of confidence in the way the relay performed. In some ways, the microprocessor relay has a very short lifetime. This is not to say that these relays have a higher failure rate, but the technology changes so quickly that the relays become obsolete in just a few years. Thus, the years of consistent use and operation do not exist as compared to the electromechanical relays. Today, the relay engineer is bombarded by manufacturers' claims of faster operating times and more advanced relay features. In many instances, the engineer is learning new setting procedures every few years.
The fact that relay technology is rapidly changing may reinforce the need for relay testing. To be sure that a relay package will meet the system requirements for speed and security,
10


testing the relay in a laboratory environment before it is placed on the power grid may become a common practice. To be effective, the relay engineer will require the ability to subject the relays to numerous faults in a short period of time to evaluate relay performance since identical relays are not purchased year after year.
11


CHAPTER III
PROTECTIVE RELAY TESTING
This chapter highlights the differences between the types of relay tests and provides details about the evolution of transient testing.
Types of Relay Testing
There are three main types of relay testing in common use: steady-state tests, dynamic (or pseudo-transient) tests, and transient tests. Each test has unique properties and specific applications.
In steady-state tests, the operating quantity (usually current or voltage) is changed very slowly until the relay just operates [3]. The test signal cannot be changed faster than the operating time of the relay in order to allow the relay a chance to operate. For relays which depend on more than one input to operate, only one quantity is varied at a time while the others are held constant.
For example, since distance relays depend on both voltage and current to make an operating decision, both quantities would be connected to the relay. The voltage would be decreased while the current is held constant until the relay operates. The test signals for this example would be similar to those shown in Figure 3.1. The use of voltage as the operating quantity in this
12


example is known as non-destructive testing since the short-time ratings of the relay are not exceeded.
Figure 3.1: Steady-State Test Signals
Due to the procedure of the steady-state test, the relay operating time cannot be evaluated. It should be noted that this test does not resemble the actual conditions on the power system. Faults normally are associated with a decrease in voltage and a corresponding increase in current. Even with this limitation, the steady-state test is useful. Many utilities use the steady-state test to check for relay integrity on relays installed on the power system. During periodic maintenance intervals, relay technicians
13


carry portable test sets into remote substations to perform steady-state tests. These tests can be compared to previous years' results to ensure the relays are still operating correctly and within specifications.
In dynamic testing, operating quantities are changed simultaneously. These tests are characterized by abrupt switching of pre-fault, fault, and post-fault states and do not simulate the normal transition between fault states [3]. As a result, the dc offset associated with fault initiation at a point other than a voltage peak is not present. Since the power system is a three-phase system, multi-phase faults will always produce a dc component in one of the phases in the actual power system. Ignoring this component is an obvious shortcoming of this type of testing. The fact that the operating quantities consist only of the fundamental frequency is another limitation of dynamic testing.
In the example of testing a distance relay, a dynamic test would apply 60 Hz nominal voltage and current to the relay during the pre-fault condition. The fault would be initiated by providing the relay diminished voltage and elevated current. Next, the relay would be given post-fault voltage and current. Since the fault states are merely switched in, however, the post-fault conditions may not always begin at current zeros as in actual breaker openings. The test signals for this example would be similar to those shown in Figure 3.2.
14


Figure 3.2: Dynamic Test Signals
Transient tests use voltage and current waveforms that closely resemble actual fault conditions on the power system. They often look like the waveforms on digital fault recorders (DFRs) that capture actual faults. The tests consist of pre-fault, fault, and post-fault conditions; they contain the fundamental frequency as well as harmonics, including the dc component associated with the fault condition. Today, analog simulators using a model power system and digital simulators using EMTP generated waveforms can perform transient testing. Figure 3.3 shows the voltage and current waveforms for typical transient testing.
15


TEST VOLTAGE
Figure 3.3: Transient Test Signals
Several papers have discussed the application of dynamic testing and transient testing. Henville et al. [3] discuss several cases where dynamic testing identified potential problems with protective relays. One example involved the well-known expansion of the mho characteristic of many distance relays. Depending on the power system's source impedance ratio and the method that the distance relay uses healthy phase voltages for polarization, the relay can expand the dynamic characteristic in comparison to the steady-state characteristic. This is demonstrated in Figure 3.4.
16


Figure 3.4: Expansion of the Mho Characteristic
The dynamic tests were used to determine the amount of mho expansion and resulted in redesign of the relay by the manufacturer to prevent unwanted operation beyond the steady-state reach.
Henville et al. [3] also list several of the limitations of dynamic testing. They concede that some very high speed relays may misoperate during dynamic testing due to the unrealistic transition between pre-fault and fault conditions that would otherwise be secure on the actual power system and during full transient testing. Also, line relays being tested for series compensated applications may show inaccurate responses since actual power system signals tend to display higher frequency components due to
17


the series capacitors.
Alexander et al. [4] discuss the differences between dynamic testing using Doble amplifiers and transient testing using an analog model power system. Their conclusion states that to fully evaluate the performance of a protective relay, realistic waveforms of voltages and currents must be used. These waveforms should consist of the transient components that affect relay performance including the dc offsets, high frequency components, and instrument transformer transients.
The decision to use dynamic testing or full transient testing should depend on the degree in which the engineer must know the relay's behavior on the actual power system. When the engineer is attempting to evaluate the performance of a number of different relay types for a specific application and intends to purchase relays based on this evaluation, full transient testing seems to be warranted.
Historical Review of Transient Testing Prior to the development of power system models in the laboratory environment, protective relay performance on a transient basis had to come from actual operation on the real power system. Relay manufacturers were able to develop the electromechanical relays and perform steady-state tests on them using scaled resistors, inductors, and capacitors; then, the relays were released for production and installed on the system. It was not
18


until the relays actually operated on the utility's power grid that the transient performance could truly be evaluated. In fact, the manufacturers relied on the fact that the longer the relay stayed on the market, the more data of actual system faults could be documented. This live testing allowed improvements to be made by the manufacturer based on field experience [5].
As time progressed, the manufacturers and utilities realized the advantages of being able to model the power system and test the relays in the laboratory. The development of the Transient Network Analyzer (TNA) which is capable of general transient studies was quickly adapted to the task of investigating the transient performance of protective relays. The TNA was able to model shunt and series capacitors, reactors, loads, surge arresters, circuit breakers, current transformers, and potential transformers as well as the distributed effects of transmission lines by using discrete resistors, inductors, and capacitors to represent the real power system. Components were sized according to the actual transmission line characteristics and were rated for the secondary current and voltage values expected during fault conditions. Once the transient was initiated, the currents and voltages propagated through the model system in real time just as they would on the actual system. However, the physical size and cost of the TNA limited its use [5,6].
The TNA study consists of three distinct phases: interconnection of the individual elements, running of the tests, and
19


evaluation of the results. The set-up phase can be very time and labor intensive which explains some of the high costs of a TNA study. Running the tests can be very efficient once the model is set up. Changing certain parameters such as the type of fault (i.e. single-line-to-ground or three-phase) or the value of fault resistance could often be done extremely easily by changing a potentiometer on the TNA panel. However, modifications to the system configuration such as adding another transmission line or changing parameters of a transformer could take considerable effort. As in most cases when interpreting results, the analysis can be extremely tedious but is often eased by the use of computers to organize the data [6].
With the advent of the Electromagnetic Transients Program (EMTP), digital modeling of the power system became possible. EMTP was developed by Dr. Hermann Dommel in the early 1960s at the Munich Institute of Technology. He continued advancing the program with Dr. V. Scott Meyer while at the Bonneville Power Administration (BPA) [6]. The EMTP Development Coordination Group (DCG) was formed in 1982 to further develop the program. The Electrical Power Research Institute's (EPRI) participation with the DCG was formalized in 1984 [7].
The EMTP is a computer program designed to solve the mathematical equations derived from the lumped and distributed parameter circuits which model an electric power system. EMTP can output the transient current and voltage waveforms of interest to
20


the relay engineer.
It has only been in the last five years that a growing interest in digital simulators by relay manufacturers, utilities, and universities has occurred [5,8,9]. BPA can be credited as having one of the first digital simulators which took, output from digital studies and applied the waveforms to relays through amplifiers. BPA's first generation simulator, however, took up considerable physical space mainly due to the implementation of the high power amplifiers.
Today's digital simulators are characterized by a compact physical design at substantially less cost than their TNA counterparts. Computer workstations generally store the digital data and are responsible for controlling the digital-to-analog converters. From that basic system, the relay manufacturers and utilities have customized their simulators according to their needs. Relay manufacturers have opted to develop automation systems to make type testing more efficient. Type testing involves creating literally hundreds of EMTP cases with varying fault initiation angles, fault resistance, and initial load conditions.
To validate new relay ideas and concepts as well as test software changes in a developing digital relay prototype, the relay is subjected to these cases. The relay manufacturers need the ability to automatically load the fault cases, play the waveforms through the relay, store the results, and generate report summaries. One manufacturer estimates that its simulator can run a fault case
21


every five seconds, allowing up to 17,000 cases a day [5].
Some utilities have decided to develop elaborate user interfaces and post-processors for their simulators. BPA's digital simulator records the output waveforms using analog-to-digital (A/D) converters and displays them for the user. BPA is also pursuing the development of an expert system to analyze the enormous amount of data created from relay testing.
As more relay engineers become familiar with the simulators, features affecting ease of use and improving analysis will undoubtedly develop. Recent computer hardware and software advances have promoted interest in digital simulator testing by utilities who previously thought transient testing to be economically prohibitive.
22


CHAPTER IV
INSTRUMENT TRANSFORMER MODELING IN EMTP
In actual power system applications, the instrument transformers perform the vital task of reproducing the primary current and voltage waveforms for the protective relays in secondary quantities. For most of the relays used by Western, the rated input current is five amperes and rated line-to-neutral input voltage is 66.4 volts. Instrument transformers consist of current transformers (CTs), potential transformers (PTs), and coupling capacitor voltage transformers (CCVTs).
Due to the choice of the instrument transformers' ratio and their basic design, the secondary waveforms are not always accurate
representations of the primary waveforms ---- even to the extent
that relay operation is affected. If the relay engineer sizes the CT ratio too small for the available fault current, saturation of the CT may occur during fault conditions. This will produce distorted secondary current waveforms which will affect relay performance. Due to the inherent design of a CT, the presence of dc in the current waveform will affect the CT's ability to reproduce this waveform to varying degrees. The capacitors in CCVTs have a natural response to sudden changes in primary voltage which tends to smooth the changes in the secondary waveforms.
When using EMTP as the source of data, the effects of
23


saturation and transient response of the instrument transformers should be modeled. The modeling of CTs and CCVTs is not a trivial task. Although EMTP makes available all of the components needed for modeling the instrument transformers, the actual values of transformers, resistors, capacitors, and inductors are not always readily available. Also, the effect of stray capacitance is not a component value listed on a parts list. To model these effects, knowledge of the physical design of the instrument transformer is required. Work has been done by various sources to model CTs and CCVTs in EMTP and verify the response of the models with actual field values in the frequencies important to protective relays.
The instrument transformers should be accurate in the frequency range of 0-3 kHz since most relays have input filters with cutoff frequency at 3 kHz [10].
Western has not investigated the modeling effects of PTs because the installations requiring simulator testing to date have used CCVTs. The simulator testing has been done on systems at voltage ratings of 230 kV and above. Due to economic factors at these voltages, CCVTs have been used to provide voltages to the protective relays.
Current Transformer (CT) Models The CT model used by Western in relay testing was provided by the Bonneville Power Administration. The topology is similar to
24


the models described in recent literature [10,11] and is shown in Figure 4.1.

CT Model
Rsurge
0.10
Xp
j.00033
Rmag
0.6
I
Rs
AA/V
4.5E-6
Saturation
Curve
T
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------j
Zburden
8.3E-6
Rsurge=surge resistance Xp=primary leakage reactance Rmag=core loss Rs=secondary winding resistance Zburden=cable lead
resistance and relay burden
all vaIues i n ohms
Figure 4.1: EMTP CT Model
As seen from Figure 4.1, the CT can be represented by the saturable transformer model in EMTP. The primary leakage reactance (Xp) will be negligible in CTs with fully distributed windings. The parameters are usually attainable from the CT secondary excitation curves provided by the transformer and breaker manufacturers who
25


install the bushing CTs. Impedances associated with the relay burden and cable leads need to be explicitly defined on the secondary of the transformer models. The burden is normally available in relay manufacturer specifications. Resistance values for the cable leads can be determined by the cable size and length. These impedances affect the error in the CTs when combined with substantial fault current. In Western's experience, modeling an unreasonably high relay burden has resulted in saturated current waveforms that were not indicative of the real system.
To validate the CT model, the system configuration used by Kezunovic et al. [11] was duplicated in EMTP using the Western model. A 1200:5 CT was used in their case study. Since Western's model uses a 5:5 ratio, the relay burden of 4.0 + j6.93 ohms used by Kezunovic et al. [11] had to be reflected to the high side of the CT. This results in the apparent impedance of 0.000069 + jO.00012 ohms. The test circuit modeled in EMTP and the resulting waveforms are shown in Figures 4.2 and 4.3, respectively. The waveforms compare well with those documented by Kezunovic et al. [11].
26


Rsc Xsc
Figure 4.2: Test Circuit for CT Saturation Evaluation
Figure 4.3: CT Transient Response for the EMTP Model
27


Coupling Capacitor Voltage Transformer (CCVT) Models Western uses the CCVT model developed and validated by Kezunovic et al. [12]. The circuit model for a 345 kV Trench Electric CCVT similar to those used on Western's system is shown in Figure 4.4.
resistor and inductor values in ohms
Figure 4.4: EMTP CCVT Model [12]
The work by Kezunovic et al. [12] has shown that the stray capacitance (Cc) from the compensating inductor has significant effect at higher frequencies. The stray capacitance (Cp) of the
28


step-down transformer primary winding also has an effect but to a somewhat lesser degree- These parameters should be derived as accurately as possible when using CCVTs of different manufacturers and design.
29


CHAPTER V
DEVELOPMENT OF WESTERN'S SIMULATOR
Many factors influenced Western's decision to develop the digital simulator for relay testing. Since environmental concerns have made constructing new transmission lines more difficult, existing transmission must handle increasing levels of power transfer. Also, existing generation is transported over large distances from areas with excess generation to the heavily populated load areas. This is especially noticeable in the Western United States where many projects have strengthened the ties into California.
Due to the operation of the power system closer to its stability limits, protective relays must operate faster and yet remain secure for faults outside their zones of protection.
Today's fastest relays claim operating times of less than one cycle. Utilities are also placing demands that protective line relays have phase selection logic to determine the faulted phase on single-line-to-ground faults; this logic is used on single-pole tripping and reclosing schemes to help maintain stability.
To evaluate advertising claims and to place relays on an equal basis, Western found that transient relay testing was necessary.
An additional force driving simulator development was Western's involvement in the California-Oregon Transmission Project (COTP).
30


This project involved construction of a 500 kV transmission line from the southern Oregon border to Redding, California. The line relay specification required all potential bidders to submit to transient relay testing before award of the relay contract.
Western was responsible for performing the testing.
At the time of COTP, digital simulators capable of full transient relay testing were not commercially available. There were computer workstations on the market that allowed digital signals to be converted to analog signals. However, an integrated system that would allow digital waveforms to be played back and amplified to levels found during fault conditions was not commercially produced.
Bonneville Power Administration (BPA) developed an in-house digital simulator in the raid-1980's. However, Western had several constraints that would have made using BPA's simulator very difficult. In determining the direction of Western's simulator, the following implementation goals were set:
1) Western had very limited lab space. The simulator was to reside in a room approximately 12 feet by 12 feet.
2) Western wanted to use commercially available components in creating the integrated system. This requirement referred to the computer hardware as well as the current and voltage amplifiers.
3) The current and voltage amplifiers had to have maximum output ratings to model faults on the majority of
31


Western's lines. This meant that the current amplifiers had to be able to output 180 amperes peak momentarily and the voltage amplifiers had to have a maximum range of 300 volts.
After substantial investigation, hardware meeting our constraints was found and procured. The computer was made by the Concurrent Computer Corporation. The voltage amplifiers were manufactured by the D.I.R. Corporation. The current amplifiers were made by the Techron Corporation. The description of the equipment specification can be found in Chapter VIII "Hardware Description."
A team of engineers and technicians was formed to integrate the components into a working simulator. The team resolved hardware-software interface issues, developed the simulator program code, constructed cabinets to house the amplifiers, and completed acceptance testing on the amplifiers.
There were a number of difficulties in this project. One of the first hurdles was to configure the computer to most efficiently output and synchronize fourteen D/A signals. The UNIX operating system had to also be learned. The voltage amplifiers had significant design problems that were fixed primarily by Western personnel.
Western's simulator is showing its age. Since the time of the initial hardware purchase nearly five years ago, significant advancements in computer capability have emerged. Today, cards are available with D/A outputs that can be used with personal computers
32


to do what the Concurrent computer does at less cost. The increased clock speeds and available random access memory (RAM) of today's computers would enhance the performance of the simulator. Also, the larger hard disks today would make storing and retrieval of the large EMTP files much easier; this would decrease the amount of file transfer time that exists with the present simulator.
33


CHAPTER VI
PREPARATION OF SIMULATOR DATA
An effective procedure for preparing EMTP generated data is usually developed and tailored by the engineer who is creating the simulator test cases. However, a basic outline of the process is generally helpful to prevent omitting critical items. In addition, an understanding of the input format used by the simulator and basic concepts in digital sampling theory can prove quite useful to the engineer obtaining questionable results from the simulator.
Data Preparation
The process of generating data, transferring the data to the simulator computer, and organizing the data on the simulator is relatively straightforward. Figure 6.1 shows the digital simulator block diagram from the sources of data to the ultimate recording of the output waveforms and relay contact operation. At the present time, EMTP resides on the VAX mainframe computer. Regardless of the source, the data input to the simulator must be formatted according to the IEEE Standard Common Format for Transient Data Exchange (COMTRADE).
34


Figure 6.1: Digital Simulator Block Diagram
As shown in Figure 6.1, one source of data is the actual recording of a disturbance on the power system using a DFR. As long as the DFR adheres to the COMTRADE standard, the digital data can be ported to the simulator computer and used directly by the simulator program. For a valid run, however, the data must be properly filtered for the desired sampling rate.
35


There are a number of steps to prepare the data for the simulator from the second source, EMTP. First, and of utmost importance, is to accurately model the power system of interest in EMTP. All pertinent system components should be included in this model. In general, the model will include frequency dependent line models, constant parameter line models, series capacitors including metal oxide varistor (MOV) protection with or without triggered gaps, shunt reactors, voltage sources with appropriate load flow, Thevenin equivalent impedances, current transformer models, and coupling capacitor voltage transformer models. Several other constraints are placed on the EMTP user when setting up the data for the simulator, as described below.
The simulator's optimal clock speed to output waveform points is 70 microseconds. Western has found the following process to work successfully:
The EMTP case should be executed with a time step of 10 microseconds.
The simulator program requires at least three cycles of steady-state from the EMTP model before the event (i.e. the fault) is initiated. If the filter program on the VAX is used, the EMTP data case should be set to provide at least four cycles of steady-state. The filter program deletes the first cycle as invalid data.
Every computed point should be saved to the EMTP plot (.PL4) file. This can be done by setting the "IPLOT"
36


parameter in the EMTP miscellaneous card to 1 [7].
The maximum values for the current and voltage waveforms should be saved to the EMTP output (.OUT) file. To accomplish this, the "MAXOUT" parameter in the EMTP miscellaneous card should be set to 1 [7]. A Fortran program has been written to read the EMTP output file and print the maximum and minimum values for the requested nodes and branches. The program can be executed by typing "VARMAX" at the VAX prompt. This information is useful in setting the variable gain on the current and voltage amplifiers.
The .PL4 file should be low pass filtered with a cutoff frequency of 5 kHz and decimated to every seventh point. This will result in a 70 microsecond time step for the simulator waveforms. Frequency content will be reduced but the low pass digital filtering prior to decimation will prevent aliasing. A routine has been written to run the EMTP case in batch mode by typing "EBAT at the VAX prompt. From the "EBAT" menu, the user may request the data to be low pass filtered, decimated, and saved to a file. The new file will have an identical case name except the new time step will be appended to the file name. As noted above, the program deletes the first cycle of filtered data due to initialization errors in the digital filter.
37


* A step signal is required at the time of fault initiation. The simulator expects a node voltage variable set to zero volts during the steady-state period; at the time of the fault, the node voltage should be switched to five volts.
An EMTP data file which can be used to accomplish this is provided in Appendix A.
The EMTP case file names should not be longer than ten characters (excluding the extension name; i.e. .PL4). This limitation ensures that file names are unique on the simulator computer.
After the EMTP .PL4 file has been created, filtered, and decimated, the binary file must be translated into the ASCII format of the COMTRADE standard. This can easily be done using a plotting program developed by BPA. An example of the profile file used by the Plot Program for this purpose is shown in Appendix B.
Once the ASCII formatted COMTRADE file is created, the data is ready to be ported to the simulator computer. The transfer can be easily and quickly done using the Ethernet connection between the VAX and simulator computers. Standard Ethernet commands are used to connect the computers and transfer data. After the file is transferred, the waveform data can remain in ASCII format or can be translated into binary format using a Fortran program residing on the simulator computer named "conv_bin." The advantages of using a binary format are a reduction in file size and faster reading time by the simulator program. The simulator program, however, will
38


accept either format. The user may invoke the simulator program by entering "relaytst" at the computer prompt.
IEEE Standard Common Format for Transient Data Exchange (COMTRADE)
A brief description of the COMTRADE standard may provide insight to possible format problems in the simulator data.
Many power system devices have the ability of storing and using digital data: digital fault recorders, real-time oscillography, computer based protective relays which digitally store fault events, and digital simulators. Initially, many of the manufacturers used proprietary methods of formatting the digital data and marketed software for data display and analysis. To manipulate the digital data beyond the standard programs, power utility programmers were forced to comply with all of the different formats for digital storage in order to write their own code.
As the name implies, COMTRADE defines a common format for storing and exchanging digital data of transient events. Although different devices may use the data differently, the standard makes it possible for these devices to physically exchange data.
COMTRADE defines three files which completely describe the digital data [13]. The header file which has a .HDR suffix is similar to a comment file where textual descriptions of the event and device recording the data can be written. The header file must be an ASCII file and can often be created using a word processor.
39


The configuration file with a .CFG suffix contains information for a computer program to correctly read the actual transient data. The configuration file must follow a strict format so that a program can read this file. The specific format is described in the standard.
The last file is the data file which has a .DAT suffix. This file contains the actual values of the sampled transient data. The values are scaled by factors listed in the configuration file. The data file also has a strict format which is described completely in the standard. The data values may be stored in either ASCII or binary. However, data transferred between devices should be in the ASCII format. Data transferred in the binary format may be unreadable by computers that represent numbers differently than the host device.
Digital Processing Theoretical Considerations Since the simulator uses digital data, digital signal processing techniques must be considered. The Nyquist criteria must be followed when the EMTP data is low pass filtered and decimated. Recall that the optimal time step for the simulator is 70 microseconds. EMTP modeling requires that the time step be less than the travelling wave time of the shortest frequency dependent line. A 70 microsecond time corresponds to the travel time of a thirteen mile line segment. Since the user often models transpositions in the transmission line of interest, the EMTP time
40


step chosen is usually less than 70 microseconds. The EMTP waveforms, then, must be manipulated to arrive at a final sampling rate of 70 microseconds for the simulator. The user must be careful not to alias high frequency signals into the final data when decimating. Aliasing refers to the undesirable process of allowing high frequency components in the time function to impersonate low frequencies because the sampling rate is too low
[141.
An example of aliasing is shown in Figures 6.2a through 6.2c. Figure 6.2a shows a 10 kHz voltage waveform created from an EMTP case. The time step used was 5 microseconds which corresponds to a sampling frequency of 200 kHz. Every calculated point was saved and plotted. The identical EMTP case was run again except the program was instructed to save every twenty-first data point to the plot file (EMTP variable IPLOT = 21), giving an effective sampling frequency of 9.5 kHz. Figure 6.2b shows the resulting voltage waveform. By overlaying the two waveforms, Figure 6.2c shows how the sampling rate "created" the aliased waveform.
41


600.0
-600.0
Figure 6.2a: 10 kHz Waveform Sampled at 200 kHz
Figure 6.2b: 10 kHz Waveform Sampled at 9.5 kHz
Figure 6.2c: 9.5 kHz Sampled Waveform Overlaid on the
200 kHz Sampled Waveform
42


Sampling at a rate at least twice as high as the highest frequency present in the waveform will prevent aliasing. If the device storing or replaying the data cannot meet the required sampling rate, the data must be low pass filtered to attenuate the high frequencies. In the case of the simulator which has a minimum sampling rate of 70 microseconds, frequencies above 7 kHz must be removed prior to any decimating. Although some of the original data may be lost, this process will prevent the high frequency components from being folded into the lower frequency data.
As discussed in the first section of this chapter, a low pass filter routine can be called automatically after running the EMTP case. A second order Butterworth digital filter is used. The filter equation is shown below:
w *
G(s) = ------------------
s2 + /2wcs + wc2
where wc = cutoff frequency
Using the bi-linear transform:
w ^ wc
f2 1-z"1l |2 * || + Wc2
[t 1+z"1] [Uz-1]
wc2T2(l + 2z_1 + z-2)
(4+2/2wcT+wc2T2) + (2wc2T2-8)z-1 + (4-2/2wcT+wc2T2 )z-2 where T = sampling interval z_1 = unit delay operator
43


The digital filter routine was written in a stand-alone program in Fortran.
44


CHAPTER VII
USER INTERFACE DESCRIPTION
This chapter describes the man-machine interface of the simulator program that allows the user to set all parameters and initiate replaying of the transient waveforms. There are two primary menus that are used to (1) select the waveforms of interest and (2) begin the simulation. The Main Menu and Run Menu are described in the following sections.
Explanation of the Main Menu
The Main Menu directs the user to enter the COMTRADE input file of interest; to use the default D/A port assignment or to assign a waveform to each D/A port; to use the port assignment of the previous run; to modify or create the scaling factors; or to exit the program. The main menu text is shown below:
1) Select Simulator File and Use the Default D/A Port
Assignment
2) Select Simulator File and Define D/A Port Assignment
3) Use the Previous D/A Port Assignment with a New Simulator
File
10) Set Program Parameters
99) Exit the Program
45


Default D/A Port Assignment Option
The Default D/A Port Assignment option frees the user from having to assign each port individually every time the program is run. The program assumes the following port assignments:
D/A Output Port 1:
D/A Output Port 2,3,4:
D/A Output Port 5,6,7:
D/A Output Port 8:
D/A Output Port 9,10,11:
Trigger Output
Ph. A, Ph. B, Ph. C Terminal #1 Currents
Ph. A, Ph. B, Ph. C Terminal #1 Voltages
Trigger Output
Ph. A, Ph. B, Ph. C Terminal #2 Currents
D/A Output Port 12,13,14:
Ph. A, Ph. B, Ph. C Terminal #2 Voltages
Each output port refers to a D/A channel as explained in Chapter VIII "Hardware Description."
The program prompts the user for the simulator case name and then lists the available branch currents from the COMTRADE input file. The user is asked to enter the branch name for each phase current at terminal #1. The user then sees the available node voltages from the COMTRADE input file and enters the bus name for each phase voltage. The user must enter appropriate current transformer (CT) and potential transformer (PT) ratios. It is recommended that the user simply input unity values for the CT and PT ratios and set the scaling factors according to the section "Set
46


Program Parameter Option."
The user is then prompted for the same information for terminal #2.
Next, the user is prompted for event trigger information. The D/A port with the trigger output waveform is intended to be connected to the trigger input of a digital fault recorder. The trigger output waveform is expected to be a step change waveform with the rising edge at the time of the event. This allows the user to time stamp the event initiation.
If the simulator file does not contain a trigger waveform, the user will be prompted for the number of time steps at which the rising edge should be placed. For example, if the event occurred three cycles into the event file and the time step was 70 microseconds, the user should input a value of 714 ([3 cyc/60 eye] *70E-6 = 714) time steps for the rising edge of the trigger. Then, the D/A port with the trigger waveform will send out the rising edge three cycles after the user requests the entire event waveform to be replayed. A typical input session using the default port assignment option is shown in Appendix C. Diagrams showing the connections between the computer and the amplifiers are shown in the Chapter VIII "Hardware Description."
Manual D/A Port Assignment Option
The user has the flexibility to assign the waveform to each D/A port. After the simulator case name is entered, the user is
47


provided the following menu to define the type of waveform:
Enter the Type of Waveform for D/A Output Port "N"
BC: Branch Currents
NV: Node Voltages
TR: Trigger Signal
NU: D/A Output Port Not Used
RU: Remaining D/A Output Ports Not Used
EX: Abort Input Session; Return to Main Menu
The branch currents and node voltages are taken from the COMTRADE input file. The trigger signal is typically a node voltage which has a rising edge at the event initiation. As in the Default Port Assignment Option, if a trigger signal is not available, the user may input the number of time steps where the rising edge of the fabricated signal should be placed. The user can assign a port to not be used which will simply output zeros during the simulator run. The user can also force the present D/A port assignment and all remaining ports to have "Not Used" status. An abbreviated input session using this option is shown in Appendix D.
Previous Case Option
This option allows the user to select another simulator case name and use the same port assignments of the previous run. The user is allowed to use this option only after a successful run has
48


taken place. The program will check that the present CQMTRADE input file and previous COMTRADE input file order the branch currents and node voltages in exactly the same way before allowing this option to proceed.
Set Program Parameter Option
In this option, the user is prompted to change the scaling factors for the current waveforms, voltage waveforms, and trigger signals. This information is stored in an auxiliary file called AUXDRFL.DAT in the user's working directory. The user should avoid creating or editing this file.
This option prompts the user for important parameters needed for proper scaling of the digital data. First, the routine asks the user to enter the actual current transformer (CT) ratio used on the power system and the current amplifier gain. If the current data is represented in secondary amperes, however, the user should enter a value of unity for the CT ratio. The program calculates a maximum current value and checks that all current waveform data is equal to or less than this maximum value. This value is calculated as follows:
CURRMAX = CTR CURE AMP GAIN 5 (21X-1)
(?rr)
where CTR = Current Transformer Ratio
CURR AMP GAIN = Current Amplifier Gain
49


Similarly, the user is asked to enter the actual potential transformer (PT) ratio used on the power system and the voltage amplifier gain. Again, if the voltage data is represented in secondary volts, the user should enter a value of unity for the PT ratio. The program checks that all the voltage waveform data is less than or equal to the calculated maximum voltage value. This value is calculated as shown below:
VOLTMAX = PTR VOLT AMP GAIN 5 (211-!)
PIT)---------------
where PTR = Potential Transformer Ratio
VOLT_AMP_GAIN = Voltage Amplifier Gain
If the data values exceed the calculated maximum values, the program will not continue. If the above procedure is followed, the instrument transformer ratios asked for in the input sessions should be entered as unity values.
If the trigger signal is a node voltage with a maximum value of five volts as suggested in Chapter VI "Preparation of Simulator Data," the user should enter a value of six for the maximum scaling factor. A slightly larger value is required due to the overshoot caused by the digital filter.
50


Explanation of the Run Menu
After Option 1, 2, or 3 of the Main Menu is selected, a listing of the port assignments that the user has selected is provided. The user is asked if any modifications are necessary.
At this point, the user has the ability to change one or all of the port assignments. Once all of the port assignments are correct, the following Run Menu is displayed:
SIMULATOR RUN MENU:
1) CYCLE ZERO OUTPUT...
ENTER "CTRL C" TO CYCLE STEADY-STATE...
ENTER "CTRL C" TO CYCLE ZERO OUTPUT.
2) CYCLE ZERO OUTPUT...
ENTER "CTRL C" TO CYCLE STEADY-STATE...
ENTER "CTRL C" TO EXECUTE THE ENTIRE INPUT FILE... AUTOMATICALLY CYCLE ZERO OUTPUT AT THE END OF RUN. 10) VIEW THE D/A OUTPUT PORT ASSIGNMENTS 90) RETURN TO THE MAIN MENU
During actual replaying of the waveforms, the user controls the state of the program using a "Ctrl C" interrupt. Whenever this interrupt occurs, the program execution is altered and the waveform data is modified as described in Chapter IX "Details of Software Implementation."
51


Cycle Steady-State Option
This option is primarily used during the initial setup of the amplifiers and protective relays. The 60 Hz steady-state waveforms are continuously cycled to allow the user to verify correct gains on the amplifiers and calibrate other recording equipment. This option is not used for transient testing since the fault event is not sent out through the D/A converters.
When this option is first entered, the D/A converters are set to output zero values. Upon the first "Ctrl C" interrupt, the D/A converters are instructed to constantly repeat the first cycle of the transient waveform which should be the prefault, steady-state condition. At this time, the user can check the steady-state currents and voltages using ammeters and voltmeters. A second "Ctrl C" interrupt instructs the D/A converters to output zero values again. A third "Ctrl C" interrupt is required to exit the simulation and return the user to the Run Menu.
Cycle Steady-State and Execute Input File Option
This option is used to cycle the steady-state waveforms for a user-defined time and then replay the transient waveforms in their entirety. This is the option to choose for transient testing.
As in the Cycle Steady-State Option, the D/A converters are initially set to output zero values. This allows the user to make sure that the equipment is ready for a test run. Upon the first "Ctrl C" interrupt, the D/A converters are instructed to constantly
52


repeat the first cycle of the transient waveform which should be the steady-state condition. This provides the protective relays with prefault currents and voltages, resembling the actual power system environment. A second "Ctrl C" interrupt instructs the D/A converters to output the entire input file which usually consists of a fault simulation. After completion of the input file, the D/A converters are automatically set to cycle zeros again. A third "Ctrl C" interrupt is required to exit the simulation and return the user to the Run Menu.
View the Port Assignments Option
This option lists the D/A port assignments for information purposes only. The user is not allowed to modify these assignments from this option.
53


CHAPTER VIII
HARDWARE DESCRIPTION
This chapter describes the hardware performance of the simulator computer, current amplifiers, and voltage amplifiers. In addition, explanations and diagrams are provided to show the proper connections between the computer and amplifiers.
Simulator Computer
The simulator computer is responsible for organizing and controlling the digital data to the power amplifiers. To accomplish these tasks, the MC5450 computer, manufactured by the Concurrent Computer Corporation, uses a 32-bit Motorola MC68020 CPU with a math co-processor running at 20 MHz. Programs can access four megabytes of RAM; an additional two megabytes reside on a separate board for graphics support. The 142 megabyte hard drive is used to store the COMTRADE data files. Figure 8.1 shows a basic drawing of the Concurrent computer. The major devices are shown.
54



ZoN/C
ID
r
ON/OFF
SWITCH
-5 1/4" FLOPPY OISK DRIVE
-1/4" TAPE DRIVE

EF12M MODULE
DA08F MODULE 1
DA08F MODULE 2
i-142 M8 HARD DRIVE
Figure 8.1: Front View of the Concurrent MC5450
The synchronizing clocks and digital-to-analog (D/A) converters reside on separate boards installed in the MC5450. The EF12M Extended Function Data Acquisition Module contains sixteen 12-bit analog inputs, two 12-bit D/A outputs, a bi-directional 16-bit I/O port, and a clock module with five programmable counters. At the present time, only the clock functions are used on this module.
The simulator program requires fourteen D/A converters. This requirement was met by using two DA08F modules which contain eight 12-bit D/A converters on each board. The program uses the 5 volt
55


output range on the D/A modules. The resolution of the least significant bit is 10 V / 212, or 2.44 millivolts.
Table 8.1 shows the conventions used to map the channels on the DA08F modules to the D/A assignments established by the program.
Table 8.1. Convention for Mapping DA08F Channels to Program Ports
Program D/A Assignment Concurrent Module
D/A Output Port 1 Channel o, Top DA08F Module 1
D/A Output Port 2 Channel 1, Top DA08F Module 1
D/A Output Port 3 Channel 2, Top DA08F Module 1
D/A Output Port 4 Channel 3, Top DA08F Module 1
D/A Output Port 5 Channel 4, Top DA08F Module 1
D/A Output Port 6 Channel 5, Top DA08F Module 1
D/A Output Port 7 Channel 6, Top DA08F Module 1
D/A Output Port 8 Channel o, Bottom DA08F Module 2
D/A Output Port 9 Channel 1, Bottom DA08F Module 2
D/A Output Port 10 Channel 2, Bottom DA08F Module 2
D/A Output Port 11 Channel 3, Bottom DA08F Module 2
D/A Output Port 12 Channel A, Bottom DA08F Module 2
D/A Output Port 13 Channel 5, Bottom DA08F Module 2
D/A Output Port 14 Channel 6, Bottom DA08F Module 2
Figure 8.2 shows the proper connections for the simulator's D/A ports to the DFR and the amplifiers; it also shows the connections between the simulator modules.
56


Figure 8.2: Connections between the Computer and Amplifiers
57


Current Amplifiers
For the current amplifiers, six Techron 7700 Series power amplifiers were used. Table 8.2 lists the important specifications of the current amplifiers.
Table 8.2. Current Amplifier Specifications
Parameters Performance
Supply Voltage 240 volt, 60 Hz AC, 3-Phase
Maximum Output Current 180 amperes peak
RMS Output Current 45 amperes, averaged over one hour
Gain 20 amperes/volt, nominal
Input Impedance 20K ohms, differential
Output Impedance Greater than 500 ohms
Frequency Response DC to 30 kHz
Voltage Amplifiers
For the voltage amplifiers, six D.I.R. Model 9272-1 power amplifiers were used. Table 8.3 lists the important specifications of the voltage amplifiers.
58


Table 8.3. Voltage Amplifier Specifications
Parameters Performance
Supply Voltage 120 volt, 60 Hz AC
Maximum Output Voltage 300 volts
Gain 45 volts/volt, nominal
Frequency Response DC to 10 kHz
New voltage amplifiers have been ordered to replace the existing amplifiers. The new amplifiers will be produced by the Techron Corporation which is the same manufacturer as the current amplifiers. Western expects that the new amplifiers will exceed all of the present amplifiers' rating specifications.
59


CHAPTER IX
DETAILS OF SOFTWARE IMPLEMENTATION
This chapter describes the software implementation of the digital simulator. The simulator program is written in Fortran 77. Explanations of the simulator features and the associated code are provided in detail; they are intended to assist someone wanting to enhance the simulator program. Many of the variables and arrays are defined for the user in this chapter. The simulator computer was provided with data acquisition software already installed by the manufacturer. The program uses the resident library routines which require the data to be organized in structured data arrays.
Soft Start Feature
As discussed in Chapter III, full transient testing requires steady-state, 60 Hz current and voltage waveforms to be applied to the relay prior to a fault test. Then, the transient waveforms associated with a fault can be sent to the relay and the relay performance can be evaluated.
The simulator program continually repeats the first cycle of the waveform to simulate prefault conditions. This is the reason for requiring the EMTP simulations to provide three cycles of steady-state before an event or fault. A sketch labeling the important time frames of a waveform is shown in Figure 9.1.
60


h Steady-State Voltage
Voltage
During
Fault
Post-
Fault
Voltage
I-
0* 74.00
Time in Milliseconds
14S.0
Figure 9.1: Time Frames of a Waveform
EMTP uses a phasor steady-state solution to determine the initial value of each current and voltage waveform when starting the transient simulation [7]. Since the power system is a three-phase system, not all of the current and voltage waveforms will turn on at a zero-crossing. Given three voltage waveforms in steady-state, for example, there is a good chance that one waveform will start with a near-zero magnitude and another will start with a near-peak magnitude. In the waveform above, the voltage magnitude is near its peak value at time equal zero. That is, at time t=0, the voltage is zero; at time t=0+, the voltage is at its peak value. There was concern that the extremely high rate of rise due
61


to this step change may cause the amplifiers to overshoot, causing high magnitude transients to be applied to the relays being tested. In addition, these transients would require adequate settling time before the amplifiers would accurately track the input. To avoid these potential problems, a soft start feature was implemented.
The D/A converters are configured to access the data in specific arrays. By manipulating this data, the soft start technique ensures a smooth transition for each amplifier as it turns on. In the steady-state portion of the waveform, each line current is 120 degrees out-of-phase with respect to one another. Each phase of the voltage is similarly displaced. A scanning technique was implemented to store the time value that each waveform first crossed zero. The slow, gradual slope of the steady-state 60 Hz sine wave provided an excellent turn-on waveform for the amplifiers.
For a voltage waveform, a zero value was stored in the data array until the first zero crossing. From that point, the actual waveform data points were stored in the arrays. This was done for each voltage waveform. A sketch showing the original voltage waveform and the corresponding soft-start feature is shown in Figure 9.2.
62


Figure 9.2: Voltage Waveform Showing Soft Start Feature
For a current waveform, a zero value was stored in the data array for the entire first cycle and continued until the first zero crossing. Then, the actual waveform data points were stored in the arrays. A typical sketch of the original current waveform and a waveform having the soft start feature is shown in Figure 9.3.
63


Figure 9.3: Current Waveform Showing Soft Start Feature
The difference in the "zeroing" of the voltages and currents is due to the fact that most protective relays use voltage as a restraint quantity and current as an operate quantity. By allowing the voltage waveforms to turn on before the current waveforms, the soft start prevented nuisance operation of the relays during testing.
Array Manipulation
As mentioned in the introduction of this chapter, several arrays are used to organize the data from the COMTRADE files and to prepare the waveforms for output to the D/A converters. The waveform data is represented digitally. So, for each time step of, say, 70 microseconds, there is an instantaneous magnitude for each
64


signal that the user is interested in ---- namely, the relay
currents and voltages.
A list of all of the important arrays, their dimensions, and a brief description is shown in Table 9.1.
Table 9.1. Array Description and Function
ARRAY(DIMENSION) ARRAY DESCRIPTION AND FUNCTION
DAPORT(14,20000) Contains data points for the fourteen channels.
EXEPT1T06(1000000) Main buffer for the top DA08F.
EXEPT7T012(1000000) Main buffer for the bottom DA08F.
SSTATE1T06(300000) Contains steady-state waveforms for top DA08F.
SSTATE7T012(300000) Contains steady-state waveforms for bottom DA08F.
SWAP1ARR(300000) Contains two cycles of steady-state waveforms for top DA08F.
SWAP7ARR(300000) Contains two cycles of steady-state waveforms for bottom DA08F.
TEMP1EXE(300000) Contains the entire event waveforms for top DA08F.
TEMP7EXE(300000) Contains the entire event waveforms for bottom DA08F.
Input Data Array
One of the arrays used to organize the simulator input data is DAPORT which is a two dimensional array; the first index is dimensioned to 14 and the second index is dimensioned to 20,000. After the user has assigned the waveforms to each D/A port using the Main Menu options, the data points are read from the COMTRADE data file and stored in this array. The first index of DAPORT
65


corresponds to each of the fourteen ports. The second index is the time step associated with the first index's waveform. Note that the program can store a maximum of 20,000 time steps. For example, if a time step of 70 microseconds is used, the program cannot accept waveforms longer than 1.4 seconds.
The data is normalized using the maximum current and voltage values described in Chapter VII as it is stored in the DAP0RT array. This step is necessary to prepare the data for the 12-bit D/A converters. The output voltage of each channel of the D/A converters is given by:
Vout = 10 (code/212) volts
where code is the data in the DAP0RT array.
The structure of the DAP0RT array allows easy, straight-forward access when creating the arrays for the D/A converters.
Digital-to-Analog Converter Arrays
The remaining arrays organize the waveform data for easy retrieval by the D/A converters. Since there are two D/A converter cards, two instances of each type of array are required.
The program assigns a circular buffer to each converter card. The buffer is defined by an array name. It is considered circular because the D/A converters continuously cycle through the buffer to access data. Each circular buffer is partitioned into five sub-
66


buffers. At any given time, two sub-buffers are allocated to the D/A converters. The third sub-buffer is idle, ready to be sent to the D/A converters on demand; another is dedicated for the program to modify; and the fifth sub-buffer is a spare to provide a margin of safety [15]. The sub-buffer dedicated to the program will be referred to as the "free" sub-buffer.
By using the circular buffer technique, the D/A converters can output the values in one sub-buffer while the program changes the data in the free sub-buffer. When a change is required, the program keeps track of the number of sub-buffers it modifies until each one has the correct data. This process is used to swap in the proper waveforms when the user enters a "Ctrl C" interrupt.
Some experimentation was needed to derive the most efficient size of each sub-buffer. It was found that swapping waveform data is optimal if each sub-buffer contains approximately 40 cycles of data. Since the time step (defined as Fortran variable DELTAT) can be different for each simulation, the following variables were defined:
PTSINCYC = Number of points in a cycle
1.0
= INT --------------
60.0 DELTAT
CYCPERCHAN = Number of cycles per channel
50000.0
= INT ------------------------------- + 2
Number of Channels PTSINCYC
67


TOTAL = Total array elements in one sub-buffer
= Number of Channels CYCPERCHAN PTSINCYC
So, the Fortran variable TOTAL defines the size of each sub-buffer.
The circular buffers are designated EXEPT1T06 and EXEPT7T012; each is dimensioned for 1,000,000 indices. EXEPT1T06 contains the digital data that will be sent to the top D/A card; EXEPT7T012 contains the digital data that will be sent to the bottom D/A card. These buffers will be directly accessed by the computer's data acquisition applications routines.
The one-dimensional arrays SSTATE1T06 and SSTATE7T012 contain the steady-state 60 Hz waveforms. By accessing the DAP0RT array, the first cycle of the transient waveform is repeatedly stored in each of these arrays to the size of one sub-buffer. As becomes necessary, the steady-state waveforms will be stored into the circular buffers. The first two cycles of the SSTATE1T06 and SSTATE7T012 arrays contain the soft start feature.
The one-dimensional arrays SWAP1ARR and SWAP7ARR contain two cycles of steady-state waveforms to replace the soft start feature during execution of the simulation. The data is taken from the first cycle of the EHTP simulated data.
The one-dimensional arrays TEMP1EXE and TEMP7EXE contain the entire fault waveforms consisting of the three cycles of steady-state followed by the transient waveforms. As a software implementation requirement, the event waveform data must not exceed
68


the length of two sub-buffers. In the simulations to date, this requirement has not been a limitation. Typically, the waveforms will not completely fill the size of two sub-buffers. Any excess array space in TEMP1EXE and TEMP7EXE will be filled with zero values.
Figure 9.4 shows the flow of data through the circular buffer for one of the D/A converter modules. The technique is identical for the other D/A module except the array names are different. As an example, imagine that the user chooses the Run Menu option that replays the steady-state waveforms followed by the event waveforms. For simplicity, consider the data swapping of only one D/A module. When the user enters the first "Ctrl C," the program copies the values from the steady-state waveform array, SSTATE1T06, into the free sub-buffer. Recall that this array contains the soft start feature. As each of the remaining sub-buffers becomes free, the program stores the data from the steady-state array into the subbuffer but replaces the soft start waveforms with the contents of the SWAP1ARR array. Eventually, the D/A converters access the first modified sub-buffer and begin to cycle the steady-state waveforms. The program replaces the soft start waveforms in the first sub-buffer with the steady-state waveforms in the SWAP1ARR array after replaying them once. A similar swap of data occurs when the user requests the fault waveforms. After the transient waveforms are sent out, the program automatically fills the subbuffers with zero values.
69


o
SSTATETT06
ARRAY
SWAPTARR
ARRAY
\ Store SSTATE1T06 \ Upon FT rst Interrupt
\
\
\
\
\
\
\
TEMPIEXE ARRAY
Store TEMPTEXE Upon Second Interrupt
s' Store Zero9 After TEMPTEXE
EXEPTTT06
EXEPTTT06 SUB-BUFFER 3
EXEPTTT06 SUB-BUFFER 5
DA08F EXEPT1T06
D/A Converter SUB-BUFFER 2
Module .
EXEPT1T06 SUB-BUFFER 1
Analog
Output
Figure 9.4s Flow of Data Through Circular Buffer


D/A Correction Routines
A small amount of drifting by the D/A converters was noticed when the digital input values were set to zero. That is, during the simulation run when the converters were sending out the zero values stored in the sub-buffers, actual measured voltages ranged from 2 millivolts to 15 millivolts.
Software correction routines were written to add a fixed integer to each array value to correct the dc bias from the D/A hardware. This fixed integer value can be easily changed as needed in the future.
71


CHAPTER X
DIGITAL SIMULATOR APPLICATIONS
Western's digital simulator has been used in several applications since its initial development in 1990. The projects and a brief description of the simulator's use are shown in Table 10.1.
Table 10.1. Summary of Simulator Applications
Year System Simulator Function
1991 California-Oregon Transmission Project Pre-qualification of 500 kV Line Relays
1991 Glen Canyon-Flagstaff Line Analysis of 345 kV Line Relay Misoperation on Series Compensated Line
1992 California-Oregon Transmission Project Verification of Line Relay Settings
1993 Kayenta Advanced Series Compensation Verification of Line Relay Settings
1993 Thane-Sne 11 i sham Line (Alaska Power Administration) Verification of Fault Locating Capability on Transmission Line/Cable Application
All of the projects listed above enabled Western engineers to gain further insight into the transient response of the line relays tested. Although all of the projects are good candidates for a
72


case study, the testing associated with the California-Oregon Transmission Project (COTP) line relays will be discussed in detail. The following COTP topics will be addressed in this chapter: the protective relay schemes, the relay specifications,
and the transient testing used for evaluation and relay settings verification.
California-Oregon Transmission Project The California-Oregon Transmission Project involved the design and construction of the Third Pacific Alternating Current (AC) Intertie. Operated at 500 kV, the transmission line spans over 340 miles starting in southern Oregon and terminating in the vicinity of Redding, California. Figure 10.1 shows a map of the area. The project participants consisted of over thirty municipal utilities, investor-owned utilities, and state and federal agencies. The Transmission Agency of Northern California (TANC) was the project manager.
73


Figure 10.1: Vicinity Map for COTP [16]
The intertie is sectionalized into four line segments:
Captain Jack-Olinda line, Olinda-Tracy line, Tracy-Tesla line, and Tracy-Los Banos line. The Captain Jack-Olinda line is 66% series compensated with capacitor banks at Captain Jack and Olinda. The Olinda-Tracy line is 70% series compensated by means of three series capacitor banks, including one midpoint bank.
Prior to COTP, the intertie between the Pacific Northwest and California consisted of two 500 kV ac transmission lines and one
74


500 kV dc line. The three lines had a total capacity of 5,200 MU. The addition of COTP increases the transfer capability by 1,600 MU [16].
Construction of the Third Pacific AC Intertie has also provided system stability benefits. Prior to COTP, tripping of one of the 500 kV ac lines during a fault initiated a remedial action scheme that dropped generation and applied dynamic braking resistors to arrest system acceleration. Under certain conditions, 1,400 MW of generation could be dropped. Although braking resistors are still used, the addition of the third ac line allows one of the lines to be tripped without dropping any generation.
Protective Relay Schemes
Due to the importance of COTP, fully redundant protection schemes were implemented. Initially, each transmission line was designed to have two primary relay sets. The relay sets were provided with separate microwave channels over geographically different paths. This design allowed failure of one relay set or communication path without affecting the operation of the other relay set in any way. Ultimately, a third primary relay set was installed which utilized the microwave channels on both paths.
To improve system stability on the Pacific AC Intertie, the protection schemes on the Captain Jack-Olinda and Olinda-Tracy lines were required to implement single-pole switching. In this type of switching, only the faulted phase is tripped for single-
75


line-to-ground faults. The other two phases remain intact which allow electrical power to flow through the unfaulted phases. This effectively lowers the accelerating power during the fault and improves the chances of a stable system. Automatic reclosing occurs approximately thirty cycles after tripping the breakers. If the fault is still present, the relays must initiate three-pole tripping.
Relay Specifications
Western was given the task of issuing the line relay specifications for COTP and evaluating the proposals. Final selection was approved by TANC. Relay terminals for all four line sections were included in the specifications. The Captain Jack-Olinda line, Olinda-Tracy line, and Tracy-Los Banos line had basically identical relaying requirements. The eight-mile Tracy-Tesla line had requirements associated with a short line.
Several relay types were likely candidates for this project: traveling wave relays, distance relays, and current differential relays. Traveling wave relays use the changes in the current and voltage waveforms at the fault inception to make tripping decisions. Using this technique, the relays can easily operate under one cycle. Distance relays implementing pilot schemes base their operation on overreaching zones of protection and fault information from the remote end of the line. Current differential relays compare the direction of current at both terminals of the
76


line to make operating decisions. Current differential relays are typically applied on short transmission lines.
The relaying requirements were very stringent based on the importance of the intertie and the strict stability margins. The major requirements outlined in the relay specification for the Captain Jack-Olinda, Olinda-Tracy, and Tracy-Los Banos lines are given below [17]:
* Two primary relay sets (sets A and B) and at least one backup relay type (set C) were to be installed at each terminal. It was also stated that an additional step distance relay was to be provided as part of the primary relay set if (1) the primary set could not provide high speed protection for zone 1 faults upon loss of the communication channel or (2) the entire primary set had to be taken out of service during relay maintenance. A directional ground overcurrent relay was required at each terminal for additional backup protection.
* The two primary relay sets either had to be made by different manufacturers or, if made by the same manufacturer, use different measuring principles and conceptional designs. Western's intent for this requirement was to allow the differences in line relay designs to complement one another. For example, one relay design might favor detecting high resistance faults while the other relay design might provide faster operating times
77


for severe faults.
During single-line-to-ground faults, the pilot relay sets were to initiate single-pole tripping and reclosing or, optionally, three-pole tripping and reclosing. The latter scheme was needed to allow reclosing during outages of the line reactors which are required for arc suppression in single-pole applications. It was indicated that the Tracy-Los Banos line was to be operated with three-pole tripping and reclosing at the present time but the option was available to operate the line single-pole at a later date.
- Faults other than single-line-to-ground were to result in three-pole tripping and blocking of reclosing.
* Relay operation was to be very high-speed (typical 1 cycle or less) with the pilot channel in service; this did not include channel delay.
Figure 10.2 is a block diagram of the major relay requirements for these lines.
78


SET 'A'
(PILOT)
PILOT TRIP SIGNAL VERY HIGH SPEED (TYPICALLY ONE CYCLE OR LESS)
LOCAL VERY HIGH SPEED TRIP (TYPICALLY ONE CYCLE OR LESS FOR ZONE 1)
PILOT TRIP SIGNAL VERY HIGH SPEED (TYPICALLY ONE CYCLE OR LESS)
SET 'C (NON-PILOT)
LOCAL NON-PILOT TRIP HIGH SPEED (MAX. 2 CYCLES FOR ZONE 1)
TIME DELAYED TRIP FOR HIGH RESISTIVE FAULT
NOTE: RECLOSER. BREAKER FAILURE AND POLE DISAGREEMENT RELAYS NOT SHOWN
Figure 10.2: COTP Line Relaying for the Captain Jack-Olinda, Olinda-Tracy, and Tracy-Los Banos Lines [17]
79


The major requirements outlined in the relay specification for the Tracy-Tesla line are given below (17]:
* Two primary current differential relay sets (sets A and B) with step distance subsystems were required. As with the other lines, an additional step distance relay set (set C) and directional ground overcurrent relay were required.
* The relay sets were to use three-pole tripping and initiate automatic reclosing only for single-line-to-ground faults.
* Relay operation was to be very high-speed (typical 1 cycle or less) with the pilot channel in service; this did not include channel delay.
Figure 10.3 is a block diagram of the major relay requirements for this line.
80


SET A
(PILOT)
PILOT TRIP SIGNAL VERY HIGH SPEED (TYPICALLY ONE CYCLE OR LESS)
LOCAL VERY HIGH SPEED TRIP (TYPICALLY ONE CYCLE OR LESS FOR ZONE 1)
PILOT TRIP SIGNAL VERY HIGH SPEED (TYPICALLY ONE CYCLE OR LESS)
SET 'C'
(NON-PILOT)
LOCAL NON-PILOT TRIP HIGH SPEED (MAX. 2 CYCLES FOR ZONE 1)
STEP DISTANCE BACKUP
NOTE: RECLOSER. BREAKER FAILURE AND POLE DISAGREEMENT RELAYS NOT SHOWN
Figure 10.3: COTP Line Relaying for the Tracy-Tesla Line [17]
81


For all terminals except the Captain Jack terminal, reclosing relays, breaker failure relays, and pole disagreement relays were to be provided for each relay set.
Transient Testing for Line Relay Evaluation
The Captain Jack-Olinda, Olinda-Tracy, and Tracy-Tesla line relay sets were required to submit to an evaluation test using Western's digital simulator as a basis for award. The relay manufacturer was required to provide and deliver two sets of their proposed relay terminals to Western's Digital Test Facility in Golden, Colorado, where back-to-back line model testing would be performed. In addition, the manufacturer was to submit optimal relay settings for each line and provide a representative to participate and supervise the transient relay testing.
Western selected the fault events and performed the EMTP studies to create the current and voltage waveforms. In preparing the fault cases, Western engineers looked for likely places for relay misoperations. One particular location involved faults near the series capacitors due to the effect of the compensation. The engineers also wanted to simulate fault conditions that fully exercised the single-pole tripping logic. Western provided the personnel to set up and run the simulator testing.
The following types of faults were created for each line section:
Internal faults. These are faults within the relay's high
82


speed zone of protection; the relay is expected to operate on internal faults.
External faults. These are faults outside of the relay's high speed zone of protection; during the evaluation testing, the relay is not expected to operate on external faults.
Evolving faults. These are faults which initially involve a single-phase-to-ground and then shortly after develop into a multi-phase fault at the same location.
Cross-country faults. These are faults involving two separate fault locations that occur shortly after one another. In Western's cases, the first fault was an external fault that was followed by an internal fault. It was expected that the relay would operate slower on this type of fault compared with an internal fault alone since the relay typically sets up restraint tendencies during external faults.
Switch-onto-faults. These are faults that simulate a power circuit breaker closing into a de-energized line with a pre-existing three-phase-to-ground fault. This condition occurs, for example, when the personal grounds are inadvertently left on a line after line maintenance. These faults are generally a problem for distance-type relays with line connected CCVTs since the relays are not given pre-fault voltages.
83


Figure 10.4 shows a system diagram showing all of the fault locations.
For each type of fault, the following parameters were varied:
* Faulted phase. Faults involving single phases to ground (AG, BG, CG), two phases (AB, BC, CA), two phases to ground, three phases (ABC), and three phases to ground were modeled. The faulted phases were systematically varied along the line so that only a subset of these faults were placed at any one location.
* Fault resistance. For single-phase-to-ground faults, the resistance of the fault was either zero ohms or thirty ohms. The thirty ohm faults gave an indication of the relay's ability to detect high resistance faults.
* Fault incidence angle. By changing this angle, the fault was initiated at a different point on wave.
* Prefault loading. The megawatt loading was changed to provide the relay with different initial conditions.
* Temporary or permanent faults. On single-line-to-ground faults with automatic reclosing, some faults were removed for the reclosing cycle to see if the relay correctly refrained from tripping the second time.
Tables 10.2 and 10.3 summarize the types of external faults and internal faults that were required for the Captain Jack-Olinda line.
84


03
Ln
MERIDIAN GRIZZLY GRIZZLY
Figure 10.4: COTP System Diagram Showing Fault Locations


Table 10.2 [17].
Captain Jack Olinda 500 kV Line External Fault Summary Table
Test No. Fault Type Fault Resis. Incident Angle Load Level Reclosing Type
FX1.1A+ CG 0. 90 1900 MW None
FX1.1B AG 0. 90 1900 MW Single-Pole
FX1.1C ABCG 0. 90 1900 MW None
FX1.ID AC 0. 150 1900 MW None
FX1.2A BG 0. 150 1900 MW None
FX1.2B ABCG 0. 0 3500 MW None
FX1.2C BC 0. 150 1900 MW None
FX1.2D* AG+BG 0. 150+10ms 1900 MW None
FX1.3A AG 0. 90 3500 MW None
FX1.3B BCG 0. 0 1900 MW None
FX1.4A AG 0. 90 1900 MW Single-Pole
FX1.4B ABCG 0. 0 3500 MW None
FX1.4C ACG 0. 150 260 MW None
+ Test FX1.1A simulates double line outage of the existing two
Malin-Round Mountain lines under maximum loading with the
resulting large power swing through the Captain Jack-Olinda
line.
* Test FX1.2D consists of an external AG fault at location FX1
followed 10 msec later by an internal BG fault at location
FI1.2.
86


Table 10.3 [17].
Captain Jack. Olinda 500 kV Line
Internal Fault Summary Table
Test No. Fault Type Fault Resis. Incident Load Angle Level
FI1.1A CG 0. 0 1900 MU
FI1.1B* AG+AB 0. 0 1900 MW
FI1.1C ABCG 0. 0 1900 MU
FI1.2A AG 0. 0 1900 MW
FI1.2B CG 30. 0 1900 MW
FI1.2C AG 30. 150 3500 MW
FI1.2D** ' AG 30. 150 1900 MW
FI1.2E ABCG 0. 150 1900 MW
FI1.3A AG 0. 0 1900 MW
FI1.3B* * CG 30. 0 1900 MW
FI1.3C AG 30. 150 3500 MW
FI1.3D ABCG 0. 150 1900 MW
FI1.3E AC 0. 0 1900 MW
FI1.3F ACG 0. 0 1900 MW
FI1.3G* AG+AC 0. 0 1900 MU
FI1.4A BG 0. 150 260 MW
FI1.4B* * CG 30. 0 260 MW
FI1.4C AG 30. 150 3500 MW
FI1.4D ABCG 0. 150 260 MW
FI1.4E AC 0. 0 1900 MW
FI1.4F ACG 30. 150 1900 MW
FI1.4G* BG+AB 30. 150 1900 MW
FI1.5A BC 0. 0 260 MW
FI1.5B CG 30. 0 260 MU
FI1.5C AG 30. 150 3500 MW
FI1.5D* * AG 30. 150 1900 MW
FI1.5E ABCG 0. 150 260 MW
FI1.6A BG 0. 0 1900 MU
FI1.6B ABCG 0. 0 1900 MU
FI1.6C AB 0. 150 1900 MU
* Single line-to-ground fault that evolves into line-to-line-to-ground fault.
Full single-pole trip and reclose simulated, shunt reactors with neutral reactor switched in after the faulted phase opens.
87


In all, 122 faults were listed in the specification. Each relay set was scored based on its secure operation on external faults and its dependable and accurate operation on internal faults. Figure 10.5 shows a typical case of the currents and voltages sent to the protective relays during testing.
This testing procedure was extremely valuable for the evaluation process. Each proposed relay set was tested with the same waveforms and on the same simulator. Although the proposed relay sets were not subjected to every conceivable fault on the transmission line, the relay engineers were given equal performance information for each relay for a given fault. By using the same simulator during the evaluation tests, the performance of the D/A converters and amplifiers was identical for each relay manufacturer. Western believed in controlling as many variables as possible for a fair evaluation.
The actual relay sets procured exceeded the requirements in the specification. Based partly on the results of the evaluation tests, each step distance backup relay set (set C) on every line was changed to a primary pilot relay set. Due to the presence of the series capacitors on this system, reliable operation of the third relay set required a communication channel.
Transient Testing for Relay Settings Verification
After delivery of the primary relay sets and after completion of the control design, Western performed another series of
88


Case: 0TI-3B
Olinda Neatral Cuneot
2X00
o jo------------------------------------------------------------ ---------------------*----------1----------
-1000 ____________t__________________________________________________________________________________________________
OjO 43J7 84.75 150.1 173.5 214.9
Tiroe in MUItsecoo* 03TP5:ieKfrP.ODAjrCY]an-3B70JPU:l
Figure 10.5: Typical Currents and Voltages for C0TP Relay Testing
89


Full Text

PAGE 1

IMPLEMENTATION AND APPLICATION OF A DIGITAL SIMULATOR FOR PROTECTIVE RELAY TESTING by Daniel Masao Hamai B.S., University of Colorado, 1986 A thesis submitted to the Faculty of the Graduate School of the University of Colorado at Denver in partial fulfillment of the requirements for the degree of Master of Science Electrical Engineering 1994

PAGE 2

This thesis for the Master of Science degree by Daniel Masao Hamai has been approved for the Department of Electrical Engineering by Pankaj K. Sen Date

PAGE 3

Hamai, Daniel Masao (M.S., Electrical Engineering) Implementation and Application of a Digital Simulator for Protective Relay Testing Thesis directed by Professor Pankaj K. Sen ABSTRACT Power utilities have used many testing techniques to evaluate protective relay performance prior to installing the relays in the system. The actual relay characteristics can only be determined by subjecting the relays to realistic system conditions. With the recent advances in computer hardware and software, utilities have an additional tool to simulate these conditions. This thesis explores the implementation of a digital simulator for testing protective relays. The simulator injects current and voltage waveforms resembling actual fault conditions into the relays to evaluate relay performance. In contrast to other testing methods, the waveforms include both transient and steady-state components to provide more realistic test signals. The simulator consists of a dedicated computer, digital-toanalog (D/A) converter modules, and high power amplifiers. The waveforms can be created by programs such as the Electromagnetic Transients Program (EMTP) or captured from actual power system faults by digital fault recorders (DFRs). In either case, the waveforms are represented digitally. The dedicated computer replays the waveforms through the D/A converters. The amplifiers iii

PAGE 4

transform the low voltage signals of the D/A converters into the currents and voltages to be applied to the relay. The amplifiers drive the relays at current and voltage levels usually found on the secondaries of the instrument transformers in the power system. The related topics of protective relay advances, relay testing practices, and instrument transformer modeling in EMTP are also discussed in this thesis. As a specific application, the testing of the California-Oregon Transmission Project (COTP) protective relays with the digital simulator is examined. The simulator can provide the relay engineer with added insight and understanding about the transient performance of the relays. The simulator can be extremely useful in validating relay settings for complex protection schemes. This abstract accurately represents the content of the candidate's thesis. I recommend its Pankaj K. Sen iv

PAGE 5

Acknowledgments The support and assistance of everyone at Vestern Area Power Administration is deeply appreciated. Special thanks to Steve McKenna and Joel Bladow for their initial and on-going technical advice regarding the digital simulator project. Thanks to Matt Yakab for the hardware support in the Vestern Laboratory. In addition, thanks to Gary Zevenbergen for the use of several of his programs and models in this thesis. v

PAGE 6

CONTENTS CHAPTER I. INTRODUCTION. . . . . . . . . . . . . . . . 1 II. HISTORICAL REVIEY OF RELAY ADVANCEMENTS ....... 7 III. PROTECTIVE RELAY TESTING .................. 12 Types of Relay Testing ................................ 12 Historical Review of Transient Testing ............ 18 IV. INSTRUMENT TRANSFORMER MODELING IN EMTP .... 23 Current Transformer (CT) Models ....... 24 Coupling Capacitor Voltage Transformer (CCVT) Models 28 V. DEVELOPMENT OF WESTERN'S SIMULATOR .............. 30 VI. PREPARATION OF SIMULATOR DATA ................ 34 Data Preparation ........................ 34 IEEE Standard Common Format for Transient Data Exchange (COMTRADE) .............. 39 Digital Processing Theoretical Considerations ...... 40 VII. USER INTERFACE DESCRIPTION ..................... 45 Explanation of the Main Menu .................... 45 Default D/A Port Assignment Option .... 46 Manual D/A Port Assignment Option ................. 47 Previous Case Option ............................. 48 Set Program Parameter Option ....................... 49 Explanation of the Run Menu ........................ 51 vi

PAGE 7

Cycle Steady-State Option .................... 52 Cycle Steady-State and Execute Input File Option .. 52 View the Port Assignments Option ...... 53 VIII. HARDWARE DESCRIPTION ............... 54 Simulator Computer .................................... 54 Current Amplifiers .................................... 58 Voltage Amplifiers ............................ 58 IX. DETAILS OF SOFTVARE IMPLEMENTATION .............. 60 So t Start Feature .................................... 60 Array Manipulation .................................... 64 Input Data Array ................................... 65 Digital-to-Analog Converter Arrays .......... 66 D/A Correction Routines ........................ 71 X. DIGITAL SIMULATOR APPLICATIONS ..................... 72 California-Oregon Transmission Project ........ 73 Protective Relay Schemes ........... 75 Relay Specifications ................................ 76 Transient Testing for Line Relay Evaluation ......... 82 Transient Testing for Relay Settings Verification .. 88 XI CONCLUSION .................................... 91 Future Work ........................................... 93 Vest ern's Experience .................................. 94 APPENDIX A. EMTP MODEL USED FOR SIMULATOR FAULT INITIATION SIGNAL ... 95 vii

PAGE 8

B. BPA PLOT PROGRAM "PROFILE" FILE ............. 96 C. TYPICAL INPUT SESSION USING THE DEFAULT PARAMETER MENU OPTION .................. 97 D. ABBREVIATED INPUT SESSION FOR DEFINING EACH D/A PORT ... 102 BIBLIOGRAPHY .......................................... 107 viii

PAGE 9

FIGURES Figure 1.1 Digital Simulator Layout and Corresponding Power System Diagram ........................................... 4 1.2 Major System Components of the Simulator ..... 5 3.1 Steady-State Test Signals ........................... 13 3. 2 Dynamic Test Signals ...................................... 15 3.3 Transient Test Signals ................................... 16 3.4 Expansion of the Mho Characteristic ........... 17 4. 1 EMTP CT Model ............................................. 25 4.2 Test Circuit for CT Saturation Evaluation ............ 27 4.3 CT Transient Response for the EMTP Model ...... 27 4. 4 EMTP CCVT Model ........................ 28 6.1 Digital Simulator Block Diagram ............... 35 6.2a 10kHz Vaveform Sampled at 200 kHz .................. 42 6.2b 10kHz Vaveform Sampled at 9.5 kHz ......... 42 6.2c 9.5 kHz Sampled Vaveform Overlaid on the 200 kHz Sampled Waveform .......................................... 42 8.1 Front View of the Concurrent MC5450 ............. 55 8.2 Connections between the Computer and Amplifiers ........ 57 9.1 Time Frames of a Vaveform ....... ........................ 61 9.2 Voltage Vaveform Showing Soft Start Feature .......... 63 9.3 Current Vaveform Showing Soft Start Feature .......... 64 9.4 Flow of Data Through Circular Buffer ..................... 70 ix

PAGE 10

10.1 Vicinity Map for COTP .................................. 74 10.2 COTP Line Relaying for the Captain Jack-Olinda, Olinda-Tracy, and Tracy-Los Banos Lines ................. 79 10.3 COTP Line Relaying for the Tracy-Tesla Line ............. 81 10.4 COTP System Diagram Showing Fault Locations .............. 85 10.5 Typical Currents and Voltages for COTP Relay Testing ..... 89 X

PAGE 11

TABLES Table 8.1 Convention for Mapping DA08F Channels to Program Ports 56 8.2 Current Amplifier Specifications ........... 58 8.3 Voltage Amplifier Specifications ......................... 59 9.1 Array Description and Function ........................ 65 10.1 Summary of Simulator Applications ................ 72 10.2 Captain Jack Olinda 500 kV Line External Fault Summary Table ............................................. 86 10.3 Captain Jack Olinda 500 kV Line Internal Fault Summary Table ........................................... 8 7 xi

PAGE 12

CHAPTER I INTRODUCTION Power utilities have always experienced difficulty in fully assessing protective relay performance before actually installing the relays in the power system. Relay operation usually occurs in response to random disturbances on the power grid caused by lightning, ice, wind, natural disasters, and human interaction with the system. Through the years, the relays themselves have changed. The physical components and the measuring principles used by protective relays have evolved with the advances in technology. Despite the unpredictable environment, power utilities must do all they can to ensure reliable relay operation on the system. Customers demand it. Utilities have devised various testing methods for commissioning relays and validating their settings. One of the earlier methods involved staged fault tests where actual faults were placed on the power grid to evaluate relay response. Later, the Transient Network Analyzer (TNA) was used which was a scaled power system model composed of discrete components. Although both of these methods are still in use, several relay manufacturers and utilities are now turning to digital simulators to test relays in a laboratory environment. Computer workstations controlling high power amplifiers can simulate fault conditions for the relays. This thesis will explain the implementation of the 1

PAGE 13

digital simulator developed by Yestern Area Power Administration (Yestern) and describe specific applications of the simulator. Yestern Area Power Administration is a federal power marketing agency that was established in 1977 under the authority of the United States Department of Energy. Yestern markets power from fifty hydroelectric power plants operated by the Bureau of Reclamation, the United States Army Corps of Engineers, and the International Boundary and Yater Commission, as well as a percentage of coal-fired power from a plant near Page, Arizona. Maximum operating capability in 1992 was more than 10,400 MY. Yestern developed the digital simulator with the goal of testing relays back-to-back for commissioning and troubleshooting. Back-to-back testing refers to the practice of bringing both relay terminals of a transmission line into the laboratory environment, providing appropriate communication channels between the relays, and testing the relay scheme as an integral system. In this type of testing, transient current and voltage waveforms corresponding to fault conditions are injected into both line relays simultaneously to evaluate relay performance. The value of this testing has increased as the use of series capacitors, static var compensators, and single-pole tripping schemes have become more popular. In particular, the presence of series capacitors has proven to be a challenge to many protective relays; simulator testing of series compensated line relays can often aid in finetuning settings for secure and dependable operation. 2

PAGE 14

The simulator uses digital data from sources such as the Electromagnetic Transients Program (EMTP) and digital fault recorders (DFRs). EMTP is extremely flexible in its capabilities since it can model transmission lines, transformers, series compensation, breakers, instrument transformers, fault switches whatever is of interest in the power system. Yhereas DFRs record faults on existing systems, EMTP can be used to model transmission systems which are in the process of being constructed. In simulator applications, EMTP can be used to model a fault or other system transient and provide the current and voltage waveforms of the two terminals of the transmission line. Internal faults, external faults, evolving faults, sequential tripping, and switch-onto-faults are typical examples of transient events that EMTP can be used for simulator applications. The simulator can replay digital fault recorder data but the user must ensure a proper sampling rate and accurate data representation (i.e. absence of noise and distortion in the fault waveforms). Yestern has found these factors to be major problems affecting DFR data. This thesis will concentrate on EMTP generated waveforms as t'he primary source of data. A dedicated computer is used to send out the digitized waveforms through digital-to-analog (D/A) converters to the current and voltage amplifiers. Yith the proper gain and scaling, the waveforms are amplified to levels which would actually be present on the secondaries of the current transformers and voltage 3

PAGE 15

transformers in the real power system. These currents and voltages are sent to the relays being tested. Figure 1.1 shows the simulator layout and the corresponding power system diagram. I TERMINAL :11:1 TERMINAL :11:2 L j COMM. DELAY VOLT. AMP. VOLT. AMP. PH. A PH. A 1--VOLT. AMP. VOLT. AMP. PH. B PH. B .__ VOLT. AMP. VOLT. AMP. PH. C PH. C r-CURR. AMP. CURR. AMP. rPH. A PH. A CURR. AMP. IDIIIDII CURR. AMP. PH. B PH. B IDI t-CURR. AMP. CURR. AMP. PH. C RELAY PH. C RELAY SIMULATOR COMPUTER Figure 1.1: Digital Simulator Layout and Corresponding Power System Diagram 4 J

PAGE 16

The computer is a real-time data acquisition machine with built-in 0/A converters. Two sets of three-phase voltage and current amplifiers are used for testing. A communication channel is provided with an adjustable time delay to simulate the travel time of the pilot signal. Figure 1.2 identifies the major system components and flow of data through the simulator. VI COUPIITER DIGITAL FAULT RECORDER I I CONCURRENT COMPUTER MEWRY ..---SIMULATOR D/A 1-H AMPLIFIERS RELAY_/ DIGITAL I PROGRAM CONV. FAULT REC. LFigure 1.2: Major System Components of the Simulator This diagram shows the digital input data from EMTP simulations or DFRs; the control of the digital data by the simulator program; the conversion of digital data to analog signals via the on-board D/A converters; the amplification of low voltage analog signals by the 5

PAGE 17

current voltage amplifiers; the transmission of the transient waveforms to the relays; and the recording of the relays' performance by a dedicated DFR. The following chapters will discuss all of the components shown in Figure 1.1 in detail. Topics will include a brief history of protective relay development, relay testing techniques with an emphasis on transient testing, EMTP modeling of instrument transformers, details regarding actual implementation of the simulator, and applications of Vestern's simulator. 6

PAGE 18

CHAPTER II HISTORICAL REVIEV OF RELAY ADVANCEMENTS The definition of a protective relay referred to in this thesis will be a device that detects abnormal power system conditions, such as short circuits, and initiates actions to isolate these conditions. In practice, the protective relay works in combination with circuit breakers that physically isolate the short circuit or fault from the rest of the system. There are three major classifications of protective relays which refer to the level of technology at the time of their development: the electromechanical relay, the solid state relay, and the microprocessor (or digital) relay. At first, all relays were electromechanical relays. Two basic operating principles were used for this type of relay: electromagnetic attraction and electromagnetic induction. Most electromagnetic attraction relays operated by means of a plunger being drawn into a solenoid or an armature being attracted to the poles of an electromagnet [1]. Instantaneous overcurrent relays often used this method. Electromagnetic induction relays followed the principle of the induction motor and created operating torque by induction in a rotor [1]. Inverse time overcurrent relays often used electromagnetic induction. By the 1930s, electromechanical inverse time overcurrent relays with instantaneous elements were used to 7

PAGE 19

coordinate relays with downstream power fuses [2]. In that time period, the two main relay manufacturers in the United States were the General Electric Company and Westinghouse Electric Corporation. Two major performance issues affecting electromechanical relays were their tendency to overreach and overtravel. On an electromagnetic attraction relay, the direct current (de) component of the fault waveform significantly affected its operation. Vith severe and long-lasting de components, an overcurrent relay could pick up even though the steady-state value of the offset wave was less than the pickup setting [1]. This overreaching tendency had to be considered when setting the relay. Since the electromechanical relays had moving parts, relay overtravel was an important issue. For example, in an inductiontype overcurrent relay, an induction disk would operate in proportion to the current provided. If the current was suddenly removed, the induction disk would continue to move due to inertia. In closely coordinated applications, the engineer had to account for relay overtravel. The introduction of the transistor by Bell Laboratories in 1948 eventually made a significant impact on the relaying industry [2]. By the late 1950s, the transistor had successfully been used to replace some of the moving parts in protective relays. However, due to the robustness of electromechanical relays and their long history of use, relays using transistor technology have not completely replaced electromechanical relays. In fact, most U.S. 8

PAGE 20

utilities still have electromechanical relays on their systems in 1994. The solid state relays are characterized by diode rectified bridges, level detectors, comparators, and timers. The power system current and voltages are often full-wave rectified to de values which can be compared to threshold levels used for tripping decisions. Comparator circuits are also used to calculate the degree of coincidence between polarizing and operating quantities. Due to the use of transistors, extremely high speed operating times less than one cycle ---can be attained in some relay designs. As the solid state technology progressed, relays began appearing with microprocessors. These relays are characterized by analog-to-digital (A/D) converters which sample and digitize the power system currents and voltages. The microprocessor functions as a signal processor and ultimately makes the operating decisions. Features of the microprocessor based relays include lower cost, multiple functions, self-checking and monitoring diagnostics, and ability to change settings remotely via computer modems. These relays can often implement all of the commonly used pilot schemes used by relay engineers; the logic is inherent to the relay and the engineer simply chooses the scheme through dip switches or menus on a monitor. As opposed to electromechanical relays where relay integrity is tested during maintenance intervals or, undesirably, during an actual fault, the microprocessor based relay can automatically run self-checking tests during its idle time. In 9

PAGE 21

most designs, a program routinely runs every few minutes to check the power supply, EPROMs, settings integrity, and other hardware. Another major advantage of the microprocessor relay is the ability to call the relay by modem and remotely change the relay settings based on power system conditions. Most of the microprocessor relays, however, cannot match the operating times of the fastest solid state, analog relays. There may be a downside to the high technology relays. Typically, electromechanical relays were left in operation for thirty year lifetimes. Years of consistent operation by specific types of relays were attained. The same relays were applied in so many places that the relay engineer gained a high level of confidence in the way the relay performed. In some ways, the microprocessor relay has a very short lifetime. This is not to say that these relays have a higher failure rate, but the technology changes so quickly that the relays become obsolete in just a few years. Thus, the years of consistent use and operation do not exist as compared to the electromechanical relays. Today, the relay engineer is bombarded by manufacturers' claims of faster operating times and more advanced relay features. In many instances, the engineer is learning new setting procedures every few years. The fact that relay technology is rapidly changing may reinforce the need for relay testing. To be sure that a relay package will meet the system requirements for speed and security, 10

PAGE 22

testing the relay in a laboratory environment before it is placed on the power grid may become a common practice. To be effective, the relay engineer will require the ability to subject the relays to numerous faults in a short period of time to evaluate relay performance since identical relays are not purchased year after year. 11

PAGE 23

CHAPTER III PROTECTIVE RELAY TESTING This chapter highlights the differences between the types of relay tests and provides details about the evolution of transient testing. Types of Relay Testing There are three main types of relay testing in common use: steady-state tests, dynamic (or pseudo-transient) tests, and transient tests. Each test has unique properties and specific applications. In steady-state tests, the operating quantity (usually current or voltage) is changed very slowly until the relay just operates [3]. The test signal cannot be changed faster than the operating time of the relay in order to allow the relay a chance to operate. For relays which depend on more than one input to operate, only one quantity is varied at a time while the others are held constant. For example, since distance relays depend on both voltage and current to make an operating decision, both quantities would be connected to the relay. The voltage would be decreased while the current is held constant until the relay operates. The test signals for this example would be similar to those shown in Figure 3.1. The use of voltage as the operating quantity in this 12

PAGE 24

example is known as non-destructive testing since the short-time ratings of the relay are not exceeded. lEST VOLTAGE 100.0 v 50.00 0.0 -50.00 -100.0 lEST CURRENT 5.000 A 2.500 0.0 t t -2.500 -5.000 Figure 3.1: Steady-State Test Signals Due to the procedure of the steady-state test, the relay operating time cannot be evaluated. It should be noted that this test does not resemble the actual conditions on the power system. Faults normally are associated with a decrease in voltage and a corresponding increase in current. Even with this limitation, the steady-state test is useful. Many utilities use the steady-state test to check for relay integrity on relays installed on the power system. During periodic maintenance intervals, relay technicians 13

PAGE 25

carry portable test sets into remote substations to perform steadystate tests. These tests can be compared to previous years' results to ensure the relays are still operating correctly and within specifications. In dynamic testing, operating quantities are changed simultaneously. These tests are characterized by abrupt switching of pre-fault, fault, and post-fault states and do not simulate the normal transition between fault states [3]. As a result, the de offset associated with fault initiation at a point other than a voltage peak is not present. Since the power system is a threephase system, multi-phase faults will always produce a de component in one of the phases in the actual power system. Ignoring this component is an obvious shortcoming of this type of testing. The fact that the operating quantities consist only of the fundamental frequency is another limitation of dynamic testing. In the example of testing a distance relay, a dynamic test would apply 60 Hz nominal voltage and current to the relay during the pre-fault condition. The fault would be initiated by providing the relay diminished voltage and elevated current. Next, the relay would be given post-fault voltage and current. Since the fault states are merely switched in, however, the post-fault conditions may not always begin at current zeros as in actual breaker openings. The test signals for this example would be similar to those shown in Figure 3.2. 14

PAGE 26

v A 100.0 50.00 0.0 -50.00 -100.0 60.00 -60.00 TEST VOLTAGE t TEST CURRENT Figure 3.2: Dynamic Test Signals Transient tests use voltage and current waveforms that closely resemble actual fault conditions on the power system. They often look like the waveforms on digital fault recorders (DFRs) that capture actual faults. The tests consist of pre-fault, fault, and post-fault conditions; they contain the fundamental frequency as well as harmonics, including the de-component associated with the fault condition. Today, analog simulators using a model power system and digital simulators using EMTP generated waveforms can perform transient testing. Figure 3.3 shows the voltage and current waveforms for typical transient testing. 15

PAGE 27

v kA 120.0 60.00 0.0 -60.00 -120.0 16.00 8.000 0.0 ..s.ooo -16.00 TEST VOLTAGE t TEST CURRENT Figure 3.3: Transient Test Signals Several papers have discussed the application of dynamic testing and transient testing. Renville et al. [3) discuss several cases where dynamic testing identified potential problems with protective relays. One example involved the well-known expansion of the mho characteristic of many distance relays. Depending on the power system's source impedance ratio and the method that the distance relay uses healthy phase voltages for polarization, the relay can expand the dynamic characteristic in comparison to the steady-state characteristic. This is demonstrated in Figure 3.4. 16

PAGE 28

Characteristic X I I I I I / Zreoch I I I lsource \ Dynamic Characteristic R Figure 3.4: Expansion of the Mho Characteristic The dynamic tests were used to determine the amount of mho expansion and resulted in redesign of the relay by the manufacturer to prevent unwanted operation beyond the steady-state reach. Henville et al. [3] also list several of the limitations of dynamic testing. They concede that some very high speed relays may misoperate during dynamic testing due to the unrealistic transition between pre-fault and fault conditions that would otherwise be secure on the actual power system and during full transient testing. Also, line relays being tested for series compensated applications may show inaccurate responses since actual power system signals tend to display higher frequency components due to 17

PAGE 29

the series capacitors. Alexander et al. [4] discuss the differences between dynamic testing using Doble amplifiers and transient testing using an analog model power system. Their conclusion states that to fully evaluate the performance of a protective relay, realistic waveforms of voltages and currents must be used. These waveforms should consist of the transient components that affect relay performance including the de offsets, high frequency components, and instrument transformer transients. The decision to use dynamic testing or full transient testing should depend on the degree in which the engineer must know the relay's behavior on the actual power system. Vhen the engineer is attempting to evaluate the performance of a number of different relay types for a specific application and intends to purchase relays based on this evaluation, full transient testing seems to be warranted. Historical Review of Transient Testing Prior to the development of power system models in the laboratory environment, protective relay performance on a transient basis had to come from actual operation on the real power system. Relay manufacturers were able to develop the electromechanical relays and perform steady-state tests on them using scaled resistors, inductors, and capacitors; then, the relays were released for production and installed on the system. It was not 18

PAGE 30

until the relays actually operated on the utility's power grid that the transient performance could truly be evaluated. In fact, the manufacturers relied on the fact that the longer the relay stayed on the market, the more data of actual system faults could be documented. This live testing allowed improvements to be made by the manufacturer based on field experience [5]. As time progressed, the manufacturers and utilities realized the advantages of being able to model the power system and test the relays in the laboratory. The development of the Transient Network Analyzer (TNA) which is capable of general transient studies was quickly adapted to the task of investigating the transient performance of protective relays. The TNA was able to model shunt and series capacitors, reactors, loads, surge arresters, circuit breakers, current transformers, and potential transformers as well as the distributed effects of transmission lines by using discrete resistors, inductors, and capacitors to represent the real power system. Components were sized according to the actual transmission line characteristics and were rated for the secondary current and voltage values expected during fault conditions. Once the transient was initiated, the currents and voltages propagated through the model system in real time just as they would on the actual system. However, the physical size and cost of the TNA limited its use [5,6]. The TNA study consists of three distinct phases: interconnection of the individual elements, running of the tests, and 19

PAGE 31

evaluation of the results. The set-up phase can be very time and labor intensive which explains some of the high costs of a TNA study. Running the tests can be very efficient once the model is set up. Changing certain parameters such as the type of fault (i.e. single-line-to-ground or three-phase) or the value of fault resistance could often be done extremely easily by changing a potentiometer on the TNA panel. However, modifications to the system configuration such as adding another transmission line or changing parameters of a transformer could take considerable effort. As in most cases when interpreting results, the analysis can be extremely tedious but is often eased by the use of computers to organize the data [6]. Vith the advent of the Electromagnetic Transients Program (EMTP), digital modeling of the power system became possible. EMTP was developed by Dr. Hermann Dommel in the early 1960s at the Munich Institute of Technology. He continued advancing the program with Dr. V. Scott Meyer while at the Bonneville Power Administration (BPA) [6]. The EMTP Development Coordination Group (DCG) was formed in 1982 to further develop the program. The Electrical Power Research Institute's (EPRI) participation with the DCG was formalized in 1984 [7]. The EMTP is a computer program designed to solve the mathematical equations derived from the lumped and distributed parameter circuits which model an electric power system. EMTP can output the transient current and voltage waveforms of interest to 20

PAGE 32

the relay engineer. It has only been in the last five years that a growing interest in digital simulators by relay manufacturers, utilities, and universities has occurred [5,8,9]. BPA can be credited as having one of the first digital simulators which took output from digital studies and applied the waveforms to relays through amplifiers. BPA's first generation simulator, however, took up considerable physical space mainly due to the implementation of the high power amplifiers. Today's digital simulators are characterized by a compact physical design at sUbstantially less cost than their TNA counterparts. Computer workstations generally store the digital data and are responsible for controlling the digital-to-analog converters. From that basic system, the relay manufacturers and utilities have customized their simulators according to their needs. Relay manufacturers have opted to develop automation systems to make type testing more efficient. Type testing involves creating literally hundreds of EMTP cases with varying fault initiation angles, fault resistance, and initial load conditions. To validate new relay ideas and concepts as well as test software changes in a developing digital relay prototype, the relay is subjected to these cases. The relay manufacturers need the ability to automatically load the fault cases, play the waveforms through the relay, store the results, and generate report summaries. One manufacturer estimates that its simulator can run a fault case 21

PAGE 33

every five seconds, allowing up to 17,000 cases a day [5]. Some utilities have decided to develop elaborate user interfaces and post-processors for their simulators. BPA's digital simulator records the output waveforms using analog-to-digital (A/D) converters and displays them for the user. BPA is also pursuing the development of an expert system to analyze the enormous amount of data created from relay testing. As more relay engineers become familiar with the simulators, features affecting ease of use and improving analysis will undoubtedly develop. Recent computer hardware and software advances have promoted interest in digital simulator testing by utilities who previously thought transient testing to be economically prohibitive. 22

PAGE 34

CHAPTER IV INSTRUMENT TRANSFORMER MODELING IN EMTP In actual power system applications, the instrument transformers perform the vital task of reproducing the primary current and voltage waveforms for the protective relays in secondary quantities. For most of the relays used by Western, the rated input current is five amperes and rated line-to-neutral input voltage is 66.4 volts. Instrument transformers consist of current transformers (CTs), potential transformers (PTs), and coupling capacitor voltage transformers (CCVTs). Due to the choice of the instrument transformers' ratio and their basic design, the secondary waveforms are not always accurate representations of the primary waveforms ---even to the extent that relay operation is affected. If the relay engineer sizes the CT ratio too small for the available fault current, saturation of the CT may occur during fault conditions. This will produce distorted secondary current waveforms which will affect relay performance. Due to the inherent design of a CT, the presence of de in the current waveform will affect the CT's ability to reproduce this waveform to varying degrees. The capacitors in CCVTs have a natural response to sudden changes in primary voltage which tends to smooth the changes in the secondary waveforms. When using EMTP as the source of data, the effects of 23

PAGE 35

saturation and transient response of the instrument transformers should be modeled. The modeling of CTs and CCVTs is not a trivial task. Although EMTP makes available all of the components needed for modeling the instrument transformers, the actual values of transformers, resistors, capacitors, and inductors are not always readily available. Also, the effect of stray capacitance is not a component value listed on a parts list. To model these effects, knowledge of the physical design of the instrument transformer is required. York has been done by various sources to model CTs and CCVTs in EMTP and verify the response of the models with actual field values in the frequencies important to protective relays. The instrument transformers should be accurate in the frequency range of 0-3 kHz since most relays have input filters with cutoff frequency at 3kHz [10]. Western has not investigated the modeling effects of PTs because the installations requiring simulator testing to date have used CCVTs. The simulator testing has been done on systems at voltage ratings of 230 kV and above. Due to economic factors at these voltages, CCVTs have been used to provide voltages to the protective relays. Current Transformer (CT) Models The CT model used by Western in relay testing was provided by the Bonneville Power Administration. The topology is similar to 24

PAGE 36

the models described in recent literature [10,11] and is shown in Figure 4.1. ,--------------------------------------, I I I I I Rsurge 0.10 Xp j.00033 Zburden 8.3E-6 Rsurge=surge resistance leakage reactance Rmag=core loss Rs=secqndary winding resistance Zburden=cable lead resistance and relay burden all values in ohms Figure 4.1: EMTP CT Model As seen from Figure 4.1, the CT can be represented by the saturable transformer model in EMTP. The primary leakage reactance (XP) will be negligible in CTs with fully distributed windings. The parameters are usually attainable from the CT secondary excitation curves provided by the transformer and breaker manufacturers who 25

PAGE 37

install the bushing CTs. Impedances associated with the relay burden and cable leads need to be explicitly defined on the secondary of the transformer models. The burden is normally available in relay manufacturer specifications. Resistance values for the cable leads can be determined by the cable size and length. These impedances affect the error in the CTs when combined with substantial fault current. In Western's experience, modeling an unreasonably high relay burden has resulted in saturated current waveforms that were not indicative of the real system. To validate the CT model, the system configuration used by Kezunovic et al. [11] was duplicated in EMTP using the Western model. A 1200:5 CT was used in their case study. Since Western's model uses a 5:5 ratio, the relay burden of 4.0 + j6.93 ohms used by Kezunovic et al. [11] had to be reflected to the high side of the CT. This results in the apparent impedance of 0.000069 + j0.00012 ohms. The test circuit modeled in EMTP and the resulting waveforms are shown in Figures 4.2 and 4.3, respectively. The waveforms compare well with those documented by Kezunovic et al. [11]. 26

PAGE 38

35.00 30.00 kA lS.OO 20.00 15.00 10.00 5.000 -5.000 -10.00 -15.00 Rsc Xsc 0.038 j1.1425 r-------1 I I I I I 1 Rsurge I 0.10 I I I I I I Rid 40.0 Xld j10.0 -------------------, Xp j .00033 CT Model l I I I I Zburden .000069 + j .00012 Rsc=source resistance Xsc=source reactance Rld=system load resistance Xld=SStem load reactance all values Tn ohms Figure 4.2: Test Circuit for CT Saturation Evaluation .... :, .... + ......... .) ............. ... ; ........... . . : : : ... : ............. : : .:: ....... : ............ ; . . : : r -< '''; ............ ... . ........ : .. 0.0 20.00 -40.00 liO.OO 80.00 100.0 Tune in MiDi>oc
PAGE 39

Coupling Capacitor Voltage Transformer (CCVT) Models Western uses the CCVT model developed and validated by Kezunovic et al. [12]. The circuit model for a 345 kV Trench Electric CCVT similar to those used on Western's system is shown in Figure 4.4. C11_ .0082 uF C2 .1351 uF Rc 127. Xc j 16267 Rp 11730:66 Xp Cc 1.0012 uf Rb 1000. C1.C2=coupling capacitors Rc.Xc.Cc=compensating inductor parameters Rp.Xp.Cp=step-down transformer parameters Rb=relay burden resistnr and inductor values in ohms Figure 4.4: EMTP CCVT Model [12] The work by Kezunovic et al. [12] has shown that the stray capacitance (Cc) from the compensating inductor has significant effect at higher frequencies. The stray capacitance (CP) of the 28

PAGE 40

step-down transformer primary winding also has an effect but to a somewhat lesser degree. These parameters should be derived as accurately as possible when using CCVTs of different manufacturers and design. 29

PAGE 41

CHAPTER V DEVELOPMENT OF WESTERN'S SIMULATOR Many factors influenced Western's decision to develop the digital simulator for relay testing. Since environmental concerns have made constructing new transmission lines more difficult, existing transmission must handle increasing levels of power transfer. Also, existing generation is transported over large distances from areas with excess generation to the heavily populated load areas. This is especially noticeable in the Western United States where many projects have strengthened the ties into California. Due to the operation of the power system closer to its stability limits, protective relays must operate faster and yet remain secure for faults outside their zones of protection. Today's fastest relays claim operating times of less than one cycle. Utilities are also placing demands that protective line relays have phase selection logic to determine the faulted phase on single-line-to-ground faults; this logic is used on single-pole tripping an9 reclosing schemes to help maintain stability. To evaluate advertising claims and to place relays on an equal basis, Western found that transient relay testing was necessary. An additional force driving simulator development was Western's involvement in the California-Oregon Transmission Project (COTP). 30

PAGE 42

This project involved construction of a 500 kV transmission line from the southern Oregon border to Redding, California. The line relay specification required all potential bidders to submit to transient relay testing before award of the relay contract. Yestern was responsible for performing the testing. At the time of COTP, digital simulators capable of full transient relay testing were not commercially available. There were computer workstations on the market that allowed digital signals to be converted to analog signals. However, an integrated system that would allow digital waveforms to be played back and amplified to levels found during fault conditions was not commercially produced. Bonneville Power Administration (BPA) developed an in-house digital simulator in the mid-1980's. However, Yestern had several constraints that would have made using BPA's simulator very difficult. In determining the direction of Yestern's simulator, the following implementation goals were set: 1) Yestern had very limited lab space. The simulator was to reside in a room approximately 12 feet by 12 feet. 2) Yestern wanted to use commercially available components in creating the integrated system. This requirement referred to the computer hardware as well as the current and voltage amplifiers. 3) The current and voltage amplifiers had to have maximum output ratings to model faults on the majority of 31

PAGE 43

Western's lines. This meant that the current amplifiers had to be able to output 180 amperes peak momentarily and the voltage amplifiers had to have a maximum range of volts. After substantial investigation, hardware meeting our constraints was found and procured. The computer was made by the Concurrent Computer Corporation. The voltage amplifiers were manufactured by the D.I.R. Corporation. The current amplifiers were made by the Techron Corporation. The description of the equipment specification can be found in Chapter VIII "Hardware Description." A team of engineers and technicians was formed to integrate the components into a working simulator. The team resolved hardware-software interface issues, developed the simulator program code, constructed cabinets to house the amplifiers, and completed acceptance testing on the amplifiers. There were a number of difficulties in this project. One of the first hurdles was to configure the computer to most efficiently output and synchronize fourteen D/A signals. The UNIX operating system had to also be learned. The voltage amplifiers had significant design problems that were fixed primarily by Western personnel. Western's simulator is showing its age. Since the time of the initial hardware purchase nearly five years ago, significant advancements in computer capability have emerged. Today, cards are available with D/A outputs that can be used with personal computers 32

PAGE 44

to do what the Concurrent computer does at less cost. The increased clock speeds and available random access memory (RAM) of today's computers would enhance the performance of the simulator. Also, the larger hard disks today would make storing and retrieval of the large EMTP files much easier; this would decrease the amount of file transfer time that exists with the present simulator. 33

PAGE 45

CHAPTER VI PREPARATION OF SIMULATOR DATA An effective procedure for preparing EMTP generated data is usually developed and tailored by the engineer who is creating the simulator test cases. However, a basic outline of the process is generally helpful to prevent omitting critical items. In addition, an understanding of the input format used by the simulator and basic concepts in digital sampling theory can prove quite useful to the engineer obtaining questionable results from the simulator. Data Preparation The process of generating data, transferring the data to the simulator computer, and organizing the data on the simulator is relatively straightforward. Figure 6.1 shows the digital simulator block diagram from the sources of data to the ultimate recording of the output waveforms and relay contact operation. At the present time, EMTP resides on the VAX mainframe computer. Regardless of the source, the data input to the simulator must be formatted according to the IEEE Standard Common Format for Transient Data Exchange (COMTRADE). 34

PAGE 46

CONCURRENT COMPUTER DIGITAL. FAULT RECORDER Figure 6.1: Digital Simulator Block Diagram As shown in Figure 6.1, one source of data is the actual recording of a disturbance on the power system using a DFR. As long as the DFR adheres to the COMTRADE standard, the digital data can be ported to the simulator computer and used directly by the simulator program. For a valid run, however, the data must be properly filtered for the desired sampling rate. 35

PAGE 47

There are a number of steps to prepare the data for the simulator from the second source, EMTP. First, and of utmost importance, is to accurately model the power system of interest in EMTP. All pertinent system components should be included in this model. In general, the model will include frequency dependent line models, constant parameter line models, series capacitors including metal oxide varistor (MOV) protection with or without triggered gaps, shunt reactors, voltage sources with appropriate load flow, Thevenin equivalent impedances, current transformer models, and coupling capacitor voltage transformer models. Several other constraints are placed on the EMTP user when setting up the data for the simulator, as described below. The simulator's optimal clock speed to output waveform points is 70 microseconds. Western has found the following process to work successfully: The EMTP case should be executed with a time of 10 microseconds. The simulator program requires at least three cycles of steady-state from the EMTP model before the event (i.e. the fault) is initiated. If the filter program on the VAX is used, the EMTP data case should be set to provide at least four cycles of steady-state. The filter program deletes the first cycle as invalid data. Every computed point should be saved to the EMTP plot (.PL4) file. This can be done by setting the "IPLOT" 36

PAGE 48

parameter in the EMTP miscellaneous card to 1 [7]. The maximum values for the current and voltage waveforms should be saved to the EMTP output (.OUT) file. To accomplish this, the "MAXOUT" parameter in the EMTP miscellaneous card should be set to 1 [7). A Fortran program has been written to read the EMTP output file and print the maximum and minimum values for the requested nodes and branches. The program can be executed by typing "VARMAX" at the VAX prompt. This information is useful in setting the variable gain on the current and voltage amplifiers. The .PL4 file should be low pass filtered with a cutoff frequency of 5 kHz and decimated to every seventh point. This will result in a 70 microsecond time step for the simulator waveforms. Frequency content will be reduced but the low pass digital filtering prior to decimation will prevent aliasing. A routine has been written to run the EMTP case in batch mode by typing "EBAT" at the VAX prompt. From the "EBAT" menu, the user may request the data to be low pass filtered, decimated, and saved to a file. The new file will have an identical case name except the new time step will be appended to the file name. As noted above, the program deletes the first cycle of filtered data due to initialization errors in the digital filter. 37

PAGE 49

A step signal is required at the time of fault initiation. The simulator expects a node voltage variable set to zero volts during the steady-state period; at the time of the fault, the node voltage should be switched to five volts. An EMTP data file which can be used to accomplish this is provided in Appendix A. The EMTP case file names should not be longer than ten characters (excluding the extension name; i.e . PL4). This limitation ensures that file names are unique on the simulator computer. After the EMTP .PL4 file has been created, filtered, and decimated, the binary file must be translated into the ASCII format of the COMTRADE standard. This can easily be done using a plotting program developed by BPA. An example of the profile file used by the Plot Program for this purpose is shown in Appendix B. Once the ASCII formatted COMTRADE file is created, the data is ready to be ported to the simulator computer. The transfer can be easily and quickly done using the Ethernet connection between the VAX and simulator computers. Standard Ethernet commands are used to connect the computers and transfer data. After the file is transferred, the waveform data can remain in ASCII format or can be translated into binary format using a Fortran program residing on the simulator computer named "conv bin." The advantages of using a binary format are a reduction in file size and faster reading time by the simulator program. The simulator program, however, will 38

PAGE 50

accept either format. The user may invoke the simulator program by entering "relaytst" at the computer prompt. IEEE Standard Common Format for Transient Data Exchange (COMTRADE) A brief description of the COMTRADE standard may provide insight to possible format problems in the simulator data. Many power system devices have the ability of storing and using digital data: digital fault recorders, real-time oscillography, computer based protective relays which digitally store fault events, and digital simulators. Initially, many of the manufacturers used proprietary methods of formatting the digital data and marketed software for data display and analysis. To manipulate the digital data beyond the standard programs, power utility programmers were forced to comply with all of the different formats for digital storage in order to write their own code. As the name implies, COMTRADE defines a common format for storing and exchanging digital data of transient events. Although different devices may use the data differently, the standard makes it possible for these devices to physically exchange data. COMTRADE defines three files which completely describe the digital data [13]. The header file which has a .HDR suffix is similar to a comment file where textual descriptions of the event and device recording the data can be written. The header file must be an ASCII file and can often be created using a word processor. 39

PAGE 51

The configuration file with a .CFG suffix contains information for a computer program to correctly read the actual transient data. The configuration file must follow a strict format so that a program can read this file. The specific format is described in the standard. The last file is the data file which has a .DAT suffix. This file contains the actual values of the sampled transient data. The values are scaled by factors listed in the configuration file. The data file also has a strict format which is described completely in the standard. The data values may be stored in either ASCII or binary. However, data transferred between devices should be in the ASCII format. Data transferred in the binary format may be unreadable by computers that represent numbers differently than the host device. Digital Processing Theoretical Considerations Since the simulator uses digital data, digital signal processing techniques must be considered. The Nyquist criteria must be followed when the EMTP data is low pass filtered and decimated. Recall that the optimal time step for the simulator is 70 microseconds. EMTP modeling requires that the time step be less than the travelling wave time of the shortest frequency dependent line. A 70 microsecond time corresponds to the travel time of a thirteen mile line segment. Since the user often models transpositions in the transmission line of interest, the EMTP time 40

PAGE 52

step chosen is usually less than 70 .microseconds. The EMTP waveforms, then, must be manipulated to arrive at a final sampling rate of 70 microseconds for the simulator. The user must be careful not to alias high frequency signals into the final data when decimating. Aliasing refers to the undesirable process of allowing high frequency components in the time function to impersonate low frequencies because the sampling rate is too low [14]. An example of aliasing is shown in Figures 6.2a through 6.2c. Figure 6.2a shows a 10 kHz voltage waveform created from an EMTP case. The time step used was 5 microseconds which corresponds to a sampling frequency of 200 kHz. Every calculated point was saved and plotted. The identical EMTP case was run again except the program was instructed to save every twenty-first data point to the plot file (EMTP variable !PLOT = 21), giving an effective sampling frequency of 9.5 kHz. Figure 6.2b shows the resulting voltage waveform. By overlaying the two waveforms, Figure 6.2c shows how the sampling rate "created" the aliased waveform. 41

PAGE 53

600.0 v 300.0 0.0 -300.0 -600.0 Figure 6.2a: 10 kHz Waveform Sampled at 200 kHz 600.0 v 300.0 0.0 t -300.0 -600.0 Figure 6.2b: 10 kHz Waveform Sampled at 9.5 kHz 600.0 v ...... 300.0 ' / / 0.0 ' / t -300.0 / / ..... -600.0 Figure 6.2c: 9.5 kHz Sampled Waveform Overlaid on the 200 kHz Sampled Waveform 42

PAGE 54

Sampling at a rate at least twice as high as the highest frequency present in the waveform will prevent aliasing. If the device storing or replaying the data cannot meet the required sampling rate, the data must be low pass filtered to attenuate the high frequencies. In the case of the simulator which has a minimum sampling rate of 70 microseconds, frequencies above 7 kHz must be removed prior to any decimating. Although some of the original data may be lost, this process will prevent the high frequency components from being.folded into the lower frequency data. As discussed in the first section of this chapter, a low pass filter routine can be called automatically after running the EMTP case. A second order Butterworth digital filter is used. The filter equation is shown below: G(s) = w 2 c where we = cutoff frequency Using the bi-linear transform: w 2 c G(z) = --------------wc2 T 2 (1 + 2z-1 + z-2 ) = --------------------------------------------------(4+2{2wcT+wc2T2) + (2wc2 T 2-8)z-1 + (4-2{2wcT+wc 2 T 2)z-2 where T = sampling interval z-1 = unit delay operator 43

PAGE 55

The digital filter routine was written in a stand-alone program in Fortran. 44

PAGE 56

CHAPTER VII USER INTERFACE DESCRIPTION This chapter describes the man-machine interface of the simulator program that allows the user to set all parameters and initiate replaying of the transient waveforms. There are two primary menus that are used to (1) select the waveforms of interest and (2) begin the simulation. The Main Menu and Run Menu are described in the following sections. Explanation of the Main Menu The Main Menu directs the user to enter the COMTRADE input file of interest; to use the default D/A port assignment or to assign a waveform to each D/A port; to use the port assignment of the previous run; to modify or create the scaling factors; or to exit the program. The main menu text is shown below: 1) Select Simulator File and Use the Default D/A Port Assignment 2) Select Simulator File and Define D/A Port Assignment 3) Use the Previous D/A Port Assignment with a New Simulator File 10) Set Program Parameters 99) Exit the Program 45

PAGE 57

Default D/A Port Assignment Option The Default D/A Port Assignment option frees the user from having to assign each port individually every time the program is run. The program assumes the following port assignments: D/A Output Port 1: D/A Output Port 2,3,4: D/A Output Port 5,6,7: D/A Output Port 8: D/A Output Port 9,10,11: D/A Output Port 12,13,14: Trigger Output Ph. A, Ph. B, Ph. C Terminal #1 Currents Ph. A, Ph. B, Ph. C Terminal #1 Voltages Trigger Output Ph. A, Ph. B, Ph. C Terminal #2 Currents Ph. A, Ph. B, Ph. C Terminal #2 Voltages Each output port refers to a D/A channel as explained in Chapter VIII "Hardware Description." The program prompts the user for the simulator case name and then lists the available branch currents from the COMTRADE input file. The user is asked to enter the branch name for each phase current at terminal #1. The user then sees the available node voltages from the COMTRADE input file and enters the bus name for each phase voltage. The user must enter appropriate current transformer (CT) and potential transformer (PT) ratios. It is recommended that the user simply input unity values for the CT and PT ratios and set the scaling factors according to the section "Set 46

PAGE 58

Program Parameter Option." The user is then prompted for the same information for terminal #2. Next, the user is prompted for event trigger information. The D/A port with the trigger output waveform is intended to be connected to the trigger input of a digital fault recorder. The trigger output waveform is expected to be a step change waveform with the rising edge at the time of the event. This allows the user to time stamp the event initiation. If the simulator file does not contain a trigger waveform, the user will be prompted for the number of time steps at which the rising edge should be placed. For example, if the event occurred three cycles into the event file and the time step was 70 microseconds, the user should input a value of 714 ([3 cyc/60 eye] *70E-6 = 714) time steps for the rising edge of the trigger. Then, the D/A port with the trigger waveform will send out the rising edge three cycles after the user requests the entire event waveform to be replayed. A typical input session using the default port assignment option is shown in Appendix C. Diagrams showing the connections between the computer and the amplifiers are shown in the Chapter VIII "Hardware Description." Manual D/A Port Assignment Option The user has the flexibility to assign the waveform to each D/A port. After the simulator case name is entered, the user is 47

PAGE 59

provided the following menu to define the type of waveform: Enter the Type of Vaveform for D/A Output Port "N" BC: Branch Currents NV: Node Voltages TR: Trigger Signal NU: D/A Output Port Not Used RU: Remaining D/A Output Ports Not Used EX: Abort Input Session; Return to Main Menu The branch currents and node voltages are taken from the COMTRADE input file. The trigger signal is typically a node voltage which has a rising edge at the event initiation. As in the Default Port Assignment Option, if a trigger signal is not available, the user may input the number of time steps where the rising edge of the fabricated signal should be placed. The user can assign a port to not be used which will simply output zeros during the simulator run. The user can also force the present D/A port assignment and all remaining ports to have "Not Used" status. An abbreviated input session using this option is shown in Appendix D. Previous Case Option This option allows the user to select another simulator case name and use the same port assignments of the previous run. The user is allowed to use this option only after a successful run has 48

PAGE 60

taken place. The program will check that the present COMTRADE input file and previous COMTRADE input file order the branch currents and node voltages in exactly the same way before allowing this option to proceed. Set Program Parameter Option In this option, the user is prompted to change the scaling factors for the current waveforms, voltage waveforms, and trigger signals. This information is stored in an auxiliary file called AUXDRFL.DAT in the user's working directory. The user should avoid creating or editing this file. This option prompts the user for important parameters needed for proper scaling of the digital data. First, the routine asks the user to enter the actual current transformer (CT) ratio used on the power system and the current amplifier gain. If the current data is represented in secondary amperes, however, the user should enter a value of unity for the CT ratio. The program calculates a maximum current value and checks that all current waveform data is equal to or less than this maximum value. This value is calculated as follows: CURRMAX CTR CURR AMP GAIN 5 (211-1) (2 ) where CTR = Current Transformer Ratio CURR AMP GAIN = Current Amplifier Gain 49

PAGE 61

Similarly, the user is asked to enter the actual potential transformer (PT) ratio used on the power system and the voltage amplifier gain. Again, if the voltage data is represented in secondary volts, the user should enter a value of unity for the PT ratio. The program checks that all the voltage waveform data is less than or equal to the calculated maximum voltage value. This value is calculated as shown below: VOLTMAX = PTR VOLT AMP GAIN 5 (211-1) (2 ) where PTR = Potential Transformer Ratio VOLT_AMP_GAIN = Voltage Amplifier Gain If the data values exceed the calculated maximum values, the program will not continue. If the above procedure is followed, the instrument transformer ratios asked for in the input sessions should be entered as unity values. If the trigger signal is a node voltage with a maximum value of five volts as suggested in Chapter VI "Preparation of Simulator Data," the user should enter a value of six for the maximum scaling factor. A slightly larger value is required due to the overshoot caused by the digital filter. 50

PAGE 62

Explanation of the Run Menu After Option 1, 2, or 3 of the Main Menu is selected, a listing of the port assignments that the user has selected is provided. The user is asked if any modifications are necessary. At this point, the user has the ability to change one or all of the port assignments. Once all of the port assignments are correct, the following Run Menu is displayed: SIMULATOR RUN MENU: 1) CYCLE ZERO OUTPUT ENTER "CTRL C" TO CYCLE STEADY-STATE . ENTER "CTRL C" TO CYCLE ZERO OUTPUT. 2) CYCLE ZERO OUTPUT ENTER "CTRL C" TO CYCLE STEADY-STATE .. ENTER "CTRL C" TO EXECUTE THE ENTIRE INPUT FILE . AUTOMATICALLY CYCLE ZERO OUTPUT AT THE END OF RUN. 10) VIEV THE D/A OUTPUT PORT ASSIGNMENTS 90) RETURN TO THE MAIN MENU During actual replaying of the waveforms, the user controls the state of the program using a "Ctrl C" interrupt. Whenever this interrupt occurs, the program execution is altered and the waveform data is modified as described in Chapter IX "Details of Software Implementation." 51

PAGE 63

Cycle Steady-State Option This option is primarily used during the initial setup of the amplifiers and protective relays. The 60 Hz steady-state waveforms are continuously cycled to allow the user to verify correct gains on the amplifiers and calibrate other recording equipment. This option is not used for transient testing since the fault event is not sent out through the D/A converters. When this option is first entered, the D/A converters are set to output zero values. Upon the first "Ctrl C" interrupt, the D/A converters are instructed to constantly repeat the first cycle of the transient waveform which should be the prefault, steady-state condition. At this time, the user can check the steady-state currents and voltages using ammeters and voltmeters. A second "Ctrl C" interrupt instructs the D/A converters to output zero values again. A third "Ctrl C" interrupt is required to exit the simulation and return the user to the Run Menu. Cycle Steady-State and Execute Input File Option This option is used to cycle the steady-state waveforms for a user-defined time and then replay the transient waveforms in their entirety. This is the option to choose for transient testing. As in the Cycle Steady-State Option, the D/A converters are initially set to output zero values. This allows the user to make sure that the equipment is ready for a test run. Upon the first "Ctrl C" interrupt, the D/A converters are instructed to constantly 52

PAGE 64

repeat the first cycle of the transient waveform which should be the steady-state condition. This provides the protective relays with prefault currents and voltages, resembling the actual power system environment. A second "Ctrl C" interrupt instructs the D/A converters to output the entire input file which usually consists of a fault simulation. After completion of the input file, the D/A converters are automatically set to cycle zeros again. A third "Ctrl C" interrupt is required to exit the simulation and return the user to the Run Menu. View the Port Assignments Option This option lists the D/A port assignments for information purposes only. The user is not allowed to modify these assignments from this option. 53

PAGE 65

CHAPTER VIII HARDWARE DESCRIPTION This chapter describes the hardware performance of the simulator computer, current amplifiers, and voltage amplifiers. In addition, explanations and diagrams are provided to show the proper connections between the computer and amplifiers. Simulator Computer The simulator computer is responsible for organizing and controlling the digital data to the power amplifiers. To accomplish these tasks, the MC5450 computer, manufactured by the Concurrent Computer Corporation, uses a 32-bit Motorola MC68020 CPU with a math co-processor running at 20 MHz. Programs can access four megabytes of RAM; an additional two megabytes reside on a separate board for graphics support. The 142 megabyte hard drive is used to store the COMTRADE data files. Figure 8.1 shows a basic drawing of the Concurrent computer. The major devices are shown. 54

PAGE 66

0 EF'12M MODULE DAOSF MODULE 1 DA08F MODULE 2 142 MB HARD DRIVE D c: 0 [E OL-4 I Ls 1/4" F SWITCH DISK DRI II 114 TAPE DRIVE Figure 8.1: Front View of the Concurrent MC5450 The synchronizing clocks and digital-to-analog (D/A) LOPPY VE converters reside on separate boards installed in the MC5450. The EF12M Extended Function Data Acquisition Module contains sixteen 12-bit analog inputs, two 12-bit D/A outputs, a bi-directional 16-bit I/0 port, and a clock module with five programmable counters. At the present time, only the clock functions are used on this module. The simulator program requires fourteen D/A converters. This requirement was met by using two DA08F modules which contain eight 12-bit D/A converters on each board. The program uses the volt 55

PAGE 67

output range on the D/A modules. The resolution of the least significant bit is 10 VI 212, or 2.44 millivolts. Table 8.1 shows the conventions used to map the channels on the DA08F modules to the D/A assignments established by the program. Table 8.1. Convention for Mapping DAOBF Channels to Program Ports Program D/A Assignment Concurrent Module D/A Output Port 1 Channel 0, Top DAOBF Module 1 D/A Output Port 2 Channel 1, Top DA08F Module 1 D/A Output Port 3 Channel 2, Top DA08F Module 1 D/A Output Port 4 Channel 3, Top DA08F Module 1 D/A Output Port 5 Channel 4, Top DA08F Module 1 D/A Output Port 6 Channel 5, Top DA08F Module 1 D/A Output Port 7 Channel 6, Top DA08F Module 1 D/A Output Port 8 Channel 0, Bottom DAOBF Module 2 D/A Output Port 9 Channel 1, Bottom DA08F Module 2 D/A Output Port 10 Channel 2, Bottom DA08F Module 2 D/A Output Port 11 Channel 3, Bottom DA08F Module 2 D/A Output Port 12 Channel 4, Bottom DA08F Module 2 D/A Output Port 13 Channel 5, Bottom DA08F Module 2 D/A Output Port 14 Channel 6, Bottom DA08F Module 2 Figure 8.2 shows the proper connections for the simulator's D/A ports to the DFR and the amplifiers; it also shows the connections between the simulator modules. 56

PAGE 68

DIGITAL FAULT RECORDER TRIGGER r--------------------------------, I CONCURRENT COMPUTER I I I I TERMINAL +1 I ::IEJ EF12M VOLTAGE AMPLIFIERS I IG04 IG03 ll';o2 ABC I : 2 I I I TERMINAL :t1 I I CURRENT AMPLIFIERS I ABC I I I I I I I I ODAOSF I ,0 20 fJ TERMINAL :i2 I VOLTAGE AMPLIFIERS I 0 012J 587 ABC I I I I TERMINAL +2 CURRENT AMPLIFIERS ABC I I oo .o 2o Jo rR o DAOSF 012J456 Figure 8.2: Connections between the Computer and Amplifiers 57

PAGE 69

Current Amplifiers For the current amplifiers, six Techron 7700 Series power amplifiers were used. Table 8.2 lists the important specifications of the current amplifiers. Table 8.2. Current Amplifier Specifications Parameters Performance Supply Voltage 240 volt, 60 Hz AC, 3-Phase Maximum Output Current 180 amperes peak RMS Output Current 45 amperes, averaged over one hour Gain 20 amperes/volt, nominal Input Impedance 20K ohms, differential Output Impedance Greater than 500 ohms Frequency Response DC to 30 kHz Voltage Amplifiers For the voltage amplifiers, six D.I.R. Model 9272-1 power amplifiers were used. Table 8.3 lists the important specifications of the voltage amplifiers. 58

PAGE 70

Table 8.3. Voltage Amplifier Specifications Parameters Performance Supply Voltage 120 volt, 60 Hz AC Maximum Output Voltage volts Gain 45 volts/volt, nominal Frequency Response DC to 10 kHz New voltage amplifiers have been ordered to replace the existing amplifiers. The new amplifiers will be produced by the Techron Corporation which is the same manufacturer as the current amplifiers. Vestern expects that the new amplifiers will exceed all of the present amplifiers' rating specifications. 59

PAGE 71

CHAPTER IX DETAILS OF SOFTWARE IMPLEMENTATION This chapter describes the software implementation of the digital simulator. The simulator program is written in Fortran 77. Explanations of the simulator features and the associated code are provided in detail; they are intended to assist someone wanting to enhance the simulator program. Many of the variables and arrays are defined for the user in this chapter. The simulator computer was provided with data acquisition software already installed by the manufacturer. The program uses the resident library routines which require the data to be organized in structured data arrays. Soft Start Feature As discussed in Chapter III, full transient testing requires steady-state, 60 Hz current and voltage waveforms to be applied to the relay prior to a fault test. Then, the transient waveforms associated with a fault can be sent to the relay and the relay performance can be evaluated. The simulator program continually repeats the first cycle of the waveform to simulate prefault conditions. This is the reason for requiring the EMTP simulations to provide three cycles of steady-state before an event or fault. A sketch labeling the important time frames of a waveform is shown in Figure 9.1. 60

PAGE 72

!--Steady-I Voltage + Post--4 State During Fault Voltage Fault Voltage 160.0 v 80.00 0.0 -80.00 -160.0 0.0 74.00 148.0 Time in Figure 9.1: Time Frames of a Yaveform EMTP uses a phasor steady-state solution to determine the initial value of each current and voltage waveform when starting the transient simulation [7]. Since the power system is a three-phase system, not all of the current and voltage waveforms will turn on at a zero-crossing. Given three voltage waveforms in steady-state, for example, there is a good chance that one waveform will start with a near-zero magnitude and another will start with a near-peak magnitude. In the waveform above, the voltage magnitude is near its peak value at time equal zero. That is, at time t=O-, the voltage is zero; at time t=O+, the voltage is at its peak value. There was concern that the extremely high rate of rise due 61

PAGE 73

to this step change may cause the amplifiers to overshoot, causing high magnitude transients to be applied to the relays being tested. In addition, these transients would require adequate settling time before the amplifiers would accurately track the input. To avoid these potential problems, a soft start feature was implemented. The D/A converters are configured to access the data in specific arrays. By manipulating this data, the soft start technique ensures a smooth transition for each amplifier as it turns on. In the steady-state portion of the waveform, each line current is 120 degrees out-of-phase with respect to one another. Each phase of the voltage is similarly displaced. A scanning technique was implemented to store the time value that each waveform first crossed zero. The slow, gradual slope of the steady-state 60 Hz sine wave provided an excellent turn-on waveform for the amplifiers. For a voltage waveform, a zero value was stored in the data array until the first zero crossing. From that point, the actual waveform data points were stored in the arrays. This was done for each voltage waveform. A sketch showing the original voltage waveform and the corresponding soft-start feature is shown in Figure 9.2. 62

PAGE 74

100.0 Original Voltage Waveform v 50.00 0.0 -50.00 t -100.0 100.0 Modified Voltage Waveform with Soft Start v 50.00 0.0 -50.00 t -100.0 Figure 9.2: Voltage Waveform Showing Soft Start Feature For a current waveform, a zero value was stored in the data array for the entire first cycle and continued until the first zero crossing. Then, the actual waveform data points were stored in the arrays. A typical sketch of the original current waveform and a waveform having the soft start feature is shown in Figure 9.3. 63

PAGE 75

10.00 Original Current Waveform A 5.000 0.0 -5.000 t -10.00 10.00 Modified Current WavefoiDl with Soft Start A 5.000 o.o -5.000 t -10.00 Figure 9.3: Current Waveform Showing Soft Start Feature The difference in the "zeroing" of the voltages and currents is due to the fact that most protective relays use voltage as a restraint quantity and current as an operate quantity. By allowing the voltage waveforms to turn on before the current waveforms, the soft start prevented nuisance operation of the relays during testing. Array Manipulation As mentioned in the introduction of this chapter, several arrays are used to organize the data from the COMTRADE files and to prepare the waveforms for output to the D/A converters. The waveform data is represented digitally. So, for each time step of, say, 70 microseconds, there is an instantaneous magnitude for each 64

PAGE 76

signal that the user is interested in ---namely, the relay currents and voltages. A list of all of the important arrays, their dimensions, and a brief description is shown in Table 9.1. Table 9.1. Array Description and Function ARRAY(DIMENSION) ARRAY DESCRIPTION-AND FUNCTION DAPORT(14,20000) EXEPT1T06(1000000) EXEPT7T012(1000000) SSTATE1T06(300000) SSTATE7T012(300000) SYAP1ARR(300000) SYAP7ARR(300000) TEMP1EXE(300000) TEMP7EXE(300000) Input Data Array Contains data points for the four-teen channels. Main buffer for the top DA08F. Main buffer for the bottom DA08F. Contains steady-state waveforms for top DA08F. Contains steady-state waveforms for bottom DA08F. Contains two cycles of steady-state waveforms for top DA08F. Contains two cycles of steady-state waveforms for bottom DA08F. Contains the entire event waveforms for top DA08F. Contains the entire event waveforms for bottom DA08F. One of the arrays used to organize the simulator input data is DAPORT which is a two dimensional array; the first index is dimensioned to 14 and the second index is dimensioned to 20,000. After the user has assigned the waveforms to each D/A port using the Main Menu options, the data points are read from the COMTRADE data file and stored in this array. The first index of DAPORT 65

PAGE 77

corresponds to each of the fourteen ports. The second index is the time step associated with the first index's waveform. Note that the program can store a maximum of 20,000 time steps. For example, if a time step of 70 microseconds is used, the program cannot accept waveforms longer than 1.4 seconds. The data is normalized using the maximum current and voltage values described in Chapter VII as it is stored in the DAPORT array. This step is necessary to prepare the data for the 12-bit D/A converters. The output voltage of each channel of the D/A converters is given by: vout = 10 (code/212) volts where code is the data in the DAPORT array. The structure of the DAPORT array allows easy, straight-forward access when creating the arrays for the D/A converters. Digital-to-Analog Converter Arrays The remaining arrays organize the waveform data for easy retrieval by the D/A converters. Since there are two D/A converter cards, two instances of each type of array are required. The program assigns a circular buffer to each converter card. The buffer is defined by an array name. It is considered circular because the D/A converters continuously cycle through the buffer to access data. Each circular buffer is partitioned into five sub-66

PAGE 78

buffers. At any given time, two sub-buffers are allocated to the D/A converters. The third sub-buffer is idle, ready to be sent to the D/A converters on demand; another is dedicated for the program to modify; and the fifth sub-buffer is a spare to provide a margin of safety [15]. The sub-buffer dedicated to the program will be referred to as the "free" sub-buffer. By using the circular buffer technique, the D/A converters can output the values in one sub-buffer while the program changes the data in the free sub-buffer. Vhen a change is required, the program keeps track of the number of sub-buffers it modifies until each one has the correct data. This process is used to swap in the proper waveforms when the user enters a "Ctrl C" interrupt. Some experimentation was needed to derive the most efficient size of each sub-buffer. It was found that swapping waveform data is optimal if each sub-buffer contains approximately 40 cycles of data. Since the time step (defined as Fortran variable DELTAT) can be different for each simulation, the following variables were defined: PTSINCYC Number of points in a cycle 1.0 = !NT 60.0 DELTAT CYCPERCHAN == Number of cycles per channel 50000.0 == !NT + 2 Number of Channels PTSINCYC 67

PAGE 79

TOTAL Total array elements in one sub-buffer = Number of Channels CYCPERCHAN PTSINCYC So, the Fortran variable TOTAL defines the size of each sub-buffer. The circular buffers are designated EXEPT1T06 and EXEPT7T012; each is dimensioned for 1,000,000 indices. EXEPT1T06 contains the digital data that will be sent to the top D/A card; EXEPT7T012 contains the digital data that will be sent to the bottom D/A card. These buffers will be directly accessed by the computer's data acquisition applications routines. The one-dimensional arrays SSTATE1T06 and SSTATE7T012 contain the steady-state 60_Hz waveforms. By accessing the DAPORT array, the first cycle of the transient waveform is repeatedly stored in each of these arrays to the size of one sub-buffer. As becomes necessary, the steady-state waveforms will be stored into the circular buffers. The first two cycles of the SSTATE1T06 and SSTATE7T012 arrays contain the soft start feature. The one-dimensional arrays SVAPlARR and SVAP7ARR contain two cycles of steady-state waveforms to replace the soft start feature during execution of the simulation. The data is taken from the first cycle of the EMTP simulated data. The one-dimensional arrays TEMPlEXE and TEMP7EXE contain the entire fault waveforms consisting of the three cycles of steadystate followed by the transient waveforms. As a software implementation requirement, the event waveform data must not exceed 68

PAGE 80

the length of two sub-buffers. In the simulations to date, this requirement has not been a limitation. Typically, the waveforms will not completely fill the size of two sub-buffers. Any excess array space in TEMPlEXE and TEMP7EXE will be filled with zero values. Figure 9.4 shows the flow of data through the circular buffer for one of the D/A converter modules. The technique is identical for the other D/A module except the array names are different. As an example, imagine that the user chooses the Run Menu option that replays the steady-state waveforms followed by the event waveforms. For simplicity, consider the data swapping of only one D/A module. When the user enters the first "Ctrl C," the program copies the values from the steady-state waveform array, SSTATE1T06, into the free sub-buffer. Recall that this array contains the soft start feature. As each of the remaining sub-buffers becomes free, the program stores the data from the steady-state array into the subbuffer but replaces the soft start waveforms with the contents of the SWAPlARR array. Eventually, the D/A converters access the first modified sub-buffer and begin to cycle the steady-state waveforms. The program replaces the soft start waveforms in the first sub-buffer with the steady-state waveforms in the SWAPlARR array after replaying them once. A similar swap of data occurs when the user requests the fault waveforms. After the transient waveforms are sent out, the program automatically fills the subbuffers with zero values. 69

PAGE 81

-....J 0 SSTATE1T06 @] SWAP1ARR ARRAY TI:MP1EXE ARRAY [!] '\, ', Store SSTATE1T06 Upon First ', Interrupt '',,,,, Store TEMP1EXE Upon Second Interrupt ______ .,. ... -------, ,, ,, ,, ,," ,-" Store Zeros ,, After TEMP1EXE I I I l I I I I I I EXEPT1T06 SUHUFFER 4 \ \ \ \ \ \ \ \ \ \ I EXEPT1T06 SUB-BUFFER 3 EXEPT1T06 SUB-BUFFER 5 ................... .................... ......... ......... ......... ......... OA08F 0/A Converter Module ......... ...................... EXEPT1T06 SUHUFFER 2 EXEPT1T08 SUB-BUFFER ... -----------------------------Figure 9.4: Flow of Data Through Circular Buffer Analog Output

PAGE 82

D/A Correction Routines A small amount of drifting by the D/A converters was noticed when the digital input values were set to zero. That is, during the simulation run when the converters were sending out the zero values stored in the sub-buffers, actual measured voltages ranged from 2 millivolts to 15 millivolts. Software correction routines were written to add a fixed integer to each array value to correct the de bias from the D/A hardware. This fixed integer value can be easily changed as needed in the future. 71

PAGE 83

CHAPTER X DIGITAL SIMULATOR APPLICATIONS Vestern's digital simulator has been used in several applications since its initial development in 1990. The projects and a brief description of the simulator's use are shown in Table 10.1. Table 10.1. Summary of Simulator Applications Year System Simulator Function 1991 California-Oregon Pre-qualification of 500 kV Transmission Project Line Relays 1991 Glen Canyon-Flagstaff Analysis of 345 kV Line Line Relay Misoperation on Series Compensated Line 1992 California-Oregon Verification of Line Relay Transmission Project Settings 1993 Kayenta Advanced Verification of Line Relay Series Compensation Settings 1993 Thane-Snettisham Verification of Fault Line Locating Capability on (Alaska Power Transmission Line/Cable Administration) Application All of the projects listed above enabled Vestern engineers to gain further insight into the transient response of the line relays tested. Although all of the projects are good candidates for a 72

PAGE 84

case study, the testing associated with the California-Oregon Transmission Project (COTP) line relays will be discussed in detail. The following COTP topics will be addressed in this chapter: the protective relay schemes, the relay specifications, and the transient testing used for evaluation and relay settings verification. California-Oregon Transmission Project The California-Oregon Transmission Project involved the design and construction of theThird Pacific Alternating Current (AC) Intertie. Operated at 500 kV, the transmission line spans over 340 miles starting in southern Oregon and terminating in the vicinity of Redding, California. Figure 10.1 shows a map of the area. The project participants consisted of over thirty municipal utilities, investor-owned utilities, and state and federal agencies. The Transmission Agency of Northern California (TANC) was the project manager. 73

PAGE 85

Proposed single circuit 500-kV Proposed single circuit 500-kV AC line (upgrade) Sacramento Proposed 500-kV AC line (new) Figure 10.1: Vicinity Map for COTP [16] The intertie is sectionalized into four line segments: Captain Jack-Olinda line, Olinda-Tracy line, Tracy-Tesla line, and Tracy-Los Banos line. The Captain Jack-Olinda line is 66% series compensated with capacitor banks at Captain Jack and Olinda. The Olinda-Tracy line is 70% series compensated by means of three series capacitor banks, including one midpoint bank. Prior to COTP, the intertie between the Pacific Northwest and California consisted of two 500 kV ac transmission lines and one 74

PAGE 86

kV de line. The three lines had a total capacity of 5,200 MW. The addition of COTP increases the transfer capability by 1,600 MW [16]. Construction of the Third Pacific AC Intertie has also provided system stability benefits. Prior to COTP, tripping of one of the 500 kV ac lines during a fault initiated a remedial action scheme that dropped generation and applied dynamic braking resistors to arrest system acceleration. Under certain conditions, 1,400 MY of generation could be dropped. Although braking resistors are still used, the addition of the third ac line allows one of the lines to be tripped without dropping any generation. Protective Relay Schemes Due to the importance of COTP, fully redundant protection schemes were implemented. Initially, each transmission line was designed to have two primary relay sets. The relay sets were provided with separate microwave channels over geographically different paths. This design allowed failure of one relay set or communication path without affecting the operation of the other relay set in any way. Ultimately, a third primary relay set was installed which utilized the microwave channels on both paths. To improve system stability on the Pacific AC Intertie, the protection schemes on the Captain Jack-Olinda and Olinda-Tracy lines were required to implement single-pole switching. In this type of switching, only the faulted phase is tripped for single-75

PAGE 87

line-to-ground faults. The other two phases remain intact which allow electrical power to flow through the unfaulted phases. This effectively lowers the accelerating power during the fault and improves the chances of a stable system. Automatic reclosing occurs approximately thirty cycles after tripping the breakers. If the fault is still present, the relays must initiate three-pole tripping. Relay Specifications Vestern was given the task of issuing the line relay specifications for COTP and evaluating the proposals. Final selection was approved by TANC. Relay terminals for all four line sections were included in the specifications. The Captain Jack Olinda line, Olinda-Tracy line, and Tracy-Los Banos line had basically identical relaying requirements. The eight-mile TracyTesla line had requirements associated with a short line. Several relay types were likely candidates for this project: traveling wave relays, distance relays, and current differential relays. Traveling wave relays use the changes in the current and voltage waveforms at the fault inception to make tripping decisions. Using this technique, the relays can easily operate under one cycle. Distance relays implementing pilot schemes base their operation on overreaching zones of protection and fault information from the remote end of the line. Current differential relays compare the direction of current at both terminals of the 76

PAGE 88

line to make operating decisions. Current differential relays are typically applied on short transmission lines. The relaying requirements were very stringent based on the importance of the intertie and the strict stability margins. The major requirements outlined in the relay specification for the Captain Jack-Olinda, Olinda-Tracy, and Tracy-Los Banos lines are given below [17]: Two primary relay sets (sets A and B) and at least one backup relay type (set C) were to be installed at each terminal. It was also stated that an additional step distance relay was to be provided as part of the primary relay set if (1) the primary set could not provide high speed protection for zone 1 faults upon loss of the communication channel or (2) the entire primary set had to be taken out of service during relay maintenance. A directional ground overcurrent relay was required at each terminal for additional backup protection. The two primary relay sets either had to be made by different manufacturers or, if made by the same manufacturer, use different measuring principles and conceptional designs. Western's intent for this requirement was to allow the differences in line relay designs to complement one another. For example, one relay design might favor detecting high resistance faults while the other relay design might provide faster operating times 77

PAGE 89

for severe faults. During single-line-to-ground faults, the pilot relay sets were to initiate single-pole tripping and reclosing or, optionally, three-pole tripping and reclosing. The latter scheme was needed to allow reclosing during outages of the line reactors which are required for arc suppression in single-pole applications. It was indicated that the TracyLos Banos line was to be operated with three-pole tripping and reclosing at the present time but the option was available to operate the line single-pole at a later date. Faults other than single-line-to-ground were to result in three-pole tripping and blocking of reclosing. Relay operation was to be very high-speed (typical 1 cycle or less) with the pilot channel in service; this did not include channel delay. Figure 10.2 is a block diagram of the major relay requirements for these lines. 78

PAGE 90

SET 'A' (PILOT) TRAVELING WAVE OR SEGREGATED PHASE COMPARISON STEP DISTANCE SET 'B' (PILOT) STEP DISTANCE SET 'C' (NON-PILOT) STEP DISTANCE BACKUP DIRECTIONAL GROUND OVERCURRENT (67G) TRANSMITTER 1 r---TRANSMITTER f----+ 2 PILOT TRIP SIGNAL VERY HIGH SPEED (TYPICALLY ONE CYCLE OR LESS) LOCAL VERY HIGH SPEED TRIP (TYPICALLY ONE CYCLE OR LESS FOR ZONE 1) PILOT TRIP SIGNAL VERY HIGH SPEED (TYPICALLY ONE CYCLE OR LESS) LOCAL NON-PILOT TRIP HIGH SPEED (MAX. 2 CYCLES FOR ZONE 1) TIME DELAYED TRIP FOR HIGH RESISTIVE FAULT NOTE: RECLOSER, BREAKER FAILURE AND POLE DISAGREEMENT RELAYS NOT SHOWN Figure 10.2: COTP Line Relaying for the Captain Jack-Olinda, Olinda-Tracy, and Tracy-Los Banos Lines [17] 79

PAGE 91

The major requirements outlined in the relay specification for the Tracy-Tesla line are given below [17)!Two primary current differential relay sets (sets A and B) with step distance subsystems were required. As with the other lines, an additional step distance relay set (set C) and directional ground overcurrent relay were required. The relay sets were to use three-pole tripping and initiate automatic reclosing only for single-line-to-ground faults. Relay operation was to be very high-speed (typical 1 cycle or less) with the pilot channel in service; this did not include channel delay. Figure 10.3 is a block diagram of the major relay requirements for this line. 80

PAGE 92

SET 'A' (PILOT) CURRENT DIFFERENTIAL STEP DISTANCE SET 'B' (PILOT) CURRENT DIFFERENTIAL STEP DISTANCE SET 'C' (NON-PILOT) STEP DISTANCE BACKUP TRANSMITTER 1 r----TRANSMITTER 2 PILOT TRIP SIGNAL VERY HIGH SPEED (TYPICALLY ONE CYCLE OR LESS) LOCAL VERY HIGH SPEED TRIP (TYPICALLY ONE CYCLE OR LESS FOR ZONE 1) PILOT TRIP SIGNAL VERY HIGH SPEED (TYPICALLY ONE CYCLE OR LESS) LOCAL NON-PILOT TRIP HIGH SPEED t-------------- (MAX. 2 CYCLES FOR ZONE 1) NOTE: RECLOSER, BREAKER FAILURE AND POLE DISAGREEMENT RELAYS NOT SHOWN Figure 10.3: COTP Line Relaying for the Tracy-Tesla Line [17] 81

PAGE 93

For all terminals except the Captain Jack terminal, reclosing relays, breaker failure relays, and pole disagreement relays were to be provided for each relay set. Transient Testing for Line Relay Evaluation The Captain Jack-Olinda, Olinda-Tracy, and Tracy-Tesla line relay sets were required to submit to an evaluation test using Western's digital simulator as a basis for award. The relay manufacturer was required to provide and deliver two sets of their proposed relay terminals to Western's Digital Test Facility in Golden, Colorado, where back-to-hack line model testing would be performed. In addition, the manufacturer was to submit optimal relay settings for each line and provide a representative to participate and supervise the transient relay testing. Western selected the fault events and performed the EMTP studies to create the current and voltage waveforms. In preparing the fault cases, Western engineers looked for likely places for relay misoperations. One particular location involved faults near the series capacitors due to the effect of the compensation. The engineers also wanted to simulate fault conditions that fully exercised the single-pole tripping logic. Western provided the personnel to set up and run the simulator testing. The following types of faults were created for each line section: Internal faults. These are faults within the relay's high 82

PAGE 94

speed zone of protection; the relay is expected to operate on internal faults. External faults. These are faults outside of the relay's high speed zone of protection; during the evaluation testing, the relay is not expected to operate on external faults. Evolving faults. These are faults which initially involve a single-phase-to-ground and then shortly after develop into a multi-phase fault at the same location. Cross-country faults. These are faults involving two separate fault locations that occur shortly after one another. In Western's cases, the first fault was an external fault that was followed by an internal fault. It was expected that the relay would operate slower on this type of fault compared with an internal fault alone since the relay typically sets up restraint tendencies during external faults. Switch-onto-faults. These are faults that simulate a power circuit breaker closing into a de-energized line with a pre-existing three-phase-to-ground fault. This condition occurs, for example, when the personal grounds are inadvertently left on a line after line maintenance. These faults are generally a problem for distance-type relays with line connected CCVTs since the relays are not given pre-fault voltages. 83

PAGE 95

Figure 10.4 shows a system diagram showing all of the fault locations. For each type of fault, the following parameters were varied: Faulted phase. Faults involving single phases to ground (AG, BG, CG), two phases (AB, BC, CA), two phases to ground, three phases (ABC), and three phases to ground were modeled. The faulted phases were systematically varied along the line so that only a subset of these faults were placed at any one location. Fault resistance. For single-phase-to-ground faults, the resistance of the fault was either zero ohms or thirty ohms. The thirty ohm faults gave an indication of the relay's ability to detect high resistance faults. Fault incidence angle. By changing this angle, the fault was initiated at a different point on wave. Prefault loading. The megawatt loading was changed to provide the relay with different initial conditions. Temporary or permanent faults. On single-line-to-ground faults with automatic reclosing, some faults were removed for the reclosing cycle to see if the relay correctly refrained from tripping the second time. Tables 10.2 and 10.3 summarize the types of external faults and internal faults that were required for the Captain Jack-Olinda line. 84

PAGE 96

OJ lJI NERlDIAN GRinLY GRIZZLY [CAPTAINJACK) IMALINI FX1.2 FX1. 1 TABlE DIXON IICMfTAlH IIEfCALF OATES IIID'IIAY IIOSS LAIllNG IH ROUND IIOUIITAIN ILOSBAN]Sl 47 IIILES rn.J 47 IIILES FI1.4 47 MILES fl1.5; FX2.1 FlJ.I t i FXJ.J TO OLINDA 230 -J' TO 11IACY 230 !MAXWELL! I TRACY I Figure 10.4: COTP System Diagram Showing Fault Locations

PAGE 97

Table 10.2 [17]. Captain Jack Olinda 500 kV Line External Fault Summary Table Fault Fault Incident Load Reclosing Test No. Type Resis. Angle Level Type FX1.1A+ CG 0. 90 1900 MW None FX1.1B AG 0. 90 1900 MW Single-Pole FX1.1C ABCG o. 90 1900 MW None FX1.1D AC 0. 150 1900 MW None FX1.2A BG 0. 150 1900 MW None FX1.2B ABCG o. oo 3500 MW None FX1.2C BC 0. 150 1900 MW None FXl. 20* AG+BG 0. 150+10ms 1900 MW None FX1.3A AG 0. 90 3500 MW None FX1.3B BCG 0. oo 1900 MW None FX1.4A AG 0. goo 1900 MW Single-Pole FX1.4B ABCG 0. oo 3500 MlJ None FX1.4C ACG 0. 150 260 MW' None + Test FX1.1A simulates double line outage of the existing two Malin-Round Mountain lines under maximum loading with the resulting large power swing through the Captain Jack-Olinda line Test FX1.2D consists of an external AG fault at location FX1.2 followed 10 msec later by an internal BG fault at location FIL 2. 86

PAGE 98

Table 10.3 [17]. Captain Jack Olinda 500 kV Line Internal Fault Summary Table Fault Fault Incident Load Test No. Type Resis. Angle Level FI1.1A CG o. oo 1900 MW FI1.1B* AG+AB o. oo 1900 MW FI1.1C ABCG o. oo 1900 MW FI1.2A AG o. oo 1900 MW FI1.2B CG 30. oo 1900 MW FI1.2C AG 30. 150 3500 MW Fil. 2D* AG 30. 150 1900 MY Fil. 2E ABCG o. 150 1900 MW FI1.3A AG o. oo 1900 MY FI1.3B** CG 30. oo 1900 MW FI1.3C AG 30. 150 3500 MW FI1.3D ABCG o. 150 1900 MY Fil. 3E AC o. oo 1900 MY FI1.3F ACG o. oo 1900 MY Fil. 3G* AG+AC o. oo 1900 MW FI1.4A BG o. 150 260 MY FI1.4B* CG 30. oo 260 MY FI1.4C AG 30. 150 3500 MY FI1.4D ABCG o. 150 260 MY FI1.4E AC o. oo 1900 MY FI1.4F ACG 30. 150 1900 MY FI1.4G* BG+AB 30. 150 1900 MY FI1.5A BC o. oo 260 MY Fil. 5B CG 30. oo 260 MY FI1.5C AG 30. 150 3500 MY FI1.5D* AG 30. 150 1900 MY FI1.5E ABCG o. 150 260 MY FI1.6A BG o. oo 1900 MY Fil. 6B ABCG o. oo 1900 MY FI1.6C AB o. 150 1900 MY Single line-to-ground fault that evolves into line-to-line-to-ground fault. * Full single-pole trip and reclose simulated, shunt reactors with neutral reactor switched in after the faulted phase opens. 87

PAGE 99

In all, 122 faults were listed in the specification. Each relay set was scored based on its secure operation on external faults and its dependable and accurate operation on internal faults. Figure 10.5 shows a typical case of the currents and voltages sent to the protective relays during testing. This testing procedure was extremely valuable for the evaluation process. Each proposed relay set was tested with the same waveforms and on the same simulator. Although the proposed relay sets were not subjected to every conceivable fault on the transmission line, the relay engineers were given equal performance information for each relay for a given fault. By using the same simulator during the evaluation tests, the performance of the D/A converters and amplifiers was identical for each relay manufacturer. Western believed in controlling as many variables as possible for a fair evaluation. The actual relay sets procured exceeded the requirements in the specification. Based partly on the results of the evaluation tests, each step distance backup relay set (set C) on every line was changed to a primary pilot relay set. Due to the presence of the series capacitors on this system, reliable operation of the third relay set required a communication channel. Transient Testing for Relay Settings Verification After delivery of the primary relay sets and after completion of the control design, Western performed another series of 88

PAGE 100

Case: OTI-38 Olinda Phase A Volrage d -80.00 v v v v -160.0 I60.D Olinda Phase B Vollage r D. (j (\ 6 (\ 1\ 1\ --80:v v if\rvv vu v -160.0 p"""" '""' c -' v ":: 4P D 4J\:P 6 1\ 0 Aficf\P 1\ -80.00 v \(; v \[v I[V v\J -160.0 Olinda Phase A Q.urent Olinda Phase C Oureot Olinda Neutral Current Fault Iudicator v 6.DOO I +-----+...J...I_.......__-+----+-----+--+----+-----t-----4 TUIIC in Milli"""""" COlPS:[EMIP.ODA....TCY]OTI-3B70.PL4:1 Figure 10.5: Typical Currents and Voltages for COTP Relay Testing 89

PAGE 101

transient testing to validate the relay settings. At this time, parameters for the autotransformers, series capacitors, shunt reactors, and transmission lines had been finalized. In particular, the series capacitor varistor sizing and thresholds for bypass using the triggered gap (on the Tracy bank) were provided by the manufacturer. The EMTP models were updated with the latest parameters, the cases rerun, and the relay sets tested by Western personnel. A few setting changes were needed to improve operation especially due to the single-pole tripping and reclosing schemes. The protection engineers requested several additional simulator cases with faults placed adjacent to the series capacitors to test relay operation. The large amount of series capacitance on the lines required some of the relay settings to be placed at their maximum values. The additional cases showed that the relays would remain secure for external faults near the capacitor banks for the conditions tested. 90

PAGE 102

CHAPTER XI CONCLUSION The digital simulator for relay testing has significant benefits over previous methods of testing. Relays can be subjected to transient waveforms representative of actual and simulated fault conditions to evaluate performance and investigate misoperations. The digital simulator can test relays with faults of various type (phase faults, ground faults), location (internal faults, external faults), fault resistance, and incidence angle. The digital simulator has become economically attractive due to the recent advances in computer hardware and software. When considering the cost of the digital simulator, two areas should be addressed: hardware costs and software development costs. The major components of the digital simulator are the dedicated computer, D/A converters, and high power amplifiers. Although a minimum of twelve D/A converters is needed, additional converters may be desired for special applications. To simulate a two terminal transmission line, six current amplifiers and six voltage amplifiers are needed. Cost estimates can easily be done using today's prices since all of the components are commercially available. The software development costs, however, are not so easily predicted. The factors that may add cost to the simulator 91

PAGE 103

implementation are the software developer's familiarity with the operating system of the simulator computer, the extent of high level or low level programming needed to control the D/A converters, the need for an elaborate user interface, and the amount of desired automation in relay testing and reporting. This thesis explored the use of transient testing during the procurement process of the COTP line relays. To determine the need for evaluation testing, the costs and benefits should be considered by the utility. The costs of performing evaluation testing include the EMTP modeling time, simulator time, and analysis time. The modeling time is directly related to the system complexity and the availability of information relating to the system equipment. For example, a simple system consisting of a few transmission lines and series capacitors may take a couple of weeks to model. On the other hand, a complex system consisting of multiple lines, series capacitors, power transformers, and shunt reactors may take significantly longer to model. The additional time is usually needed for model validation since each power system model must be verified with known performance, such as equipment test reports. During the simulator testing, each relay set should be scheduled for a week in the laboratory to allow time to set up and run the tests. Additional time should be allocated to allow for contingencies such as component failures in the relays, amplifiers, or digital fault recorder. Since relay design is proprietary, the 92

PAGE 104

testing schedule must ensure that different relay manufacturers are not at the test site simultaneously. Once the tests are completed, time must be set aside to analyze all of the results. This process will become more efficient as the utility gains experience in performing the evaluation testing. In addition, computer programs will undoubtedly develop to streamline this effort. The primary benefit of using evaluation testing is to develop common performance data for all of the proposed relays. Although the test results should not be the sole means of the purchasing decision, they can be used in conjunction with the manufacturers' overall adherence to the relay specifications. The evaluation is fair and the utility has some assurance that the relays will perform as expected. Future York Further use of the digital simulator by Western will direct future developments. A graphical user interface can already be envisioned. All of the program prompts could be menu driven. A single line diagram could be drawn on the monitor to reference the system being tested. This interface would be especially useful for the casual user so that program parameters would always be properly set. In addition, development of a high speed data acquisition system to store test results would be useful. A summary report of 93

PAGE 105

the test results tabulating trip times and relay operations could be written by this system. As mentioned in the thesis, BPA's postprocessing expert system is being developed for this purpose. In looking at the overall direction for simulators, a growing interest in real-time digital simulators is developing [18,19]. These simulators run a modified EMTP in real-time, thereby allowing the operation of the relays to interact with the EMTP models. Vestern's Experience Vestern has found the digital simulator to be extremely useful in the applications discussed in this thesis: evaluation testing and validating relay settings on lines with series compensation and single-pole switching. The simulator has become a powerful tool in understanding and evaluating relay transient performance for Ves tern's relay engineers. 94

PAGE 106

APPENDIX A EMTP MODEL USED FOR SIMULATOR FAULT INITIATION SIGNAL The following EMTP "include" file can be used to create the step signal needed at the time of fault initiation. BEGIN NEV DATA CASE MODULE FLT SIGNAL.INC ARG-<> NUM Tclose>> ; Fault initiation time ================================================================= C File: FLT SIGNAL.DIN C Fault Initiation Switch C Parameters: Fault Initiate C Usage: $INCLUDE FLT_SIGNAL.INC 0.5000 ================================================================= /BRANCH C Fault timing Circuit C Bus-->Bus-->Bus-->Bus--><-R--><-L--><-C-->===================== FTIME1 1.0 GENDC1 100.0 c ================================================================= /SVITCH C Fault Timirig Switch C Bus-->Bus--><-Tclose-><--Topen-> GENDC1FTIME1<> 10.00 c ================================================================= /SOURCE C Fault Timing Source; Low Frequency Source to Simulate DC C Bus--> 14GENDC1 5.0000 0.0001 0.0 c ================================================================= $EOF END OF DATA 95

PAGE 107

APPENDIX B BPA PLOT PROGRAM "PROFILE" FILE The following profile file can be used in the Bonneville Power Administration's (BPA) Plot Program to convert an EMTP binary plot file (.PL4) to a COMTRADE ASCII file. FORMAT: COMTRADE BITS: 12 SIGNALS: 13 1,QOMSCA l,QOMSCB 1,QOMSCC 1,QTMSCA 1,QTMSCB 1,QTMSCC 1,FTIME1 3,IOMCTA/IOMCTN 3,IOMCTB/IOMCTN 3,IOMCTC/IOMCTN 3,ITMCTA/ITMCTN 3,ITMCTB/ITMCTN 3,ITMCTC/ITMCTN STATION: Olinda-Tracy Line, 1 ANALOG: 13 Olinda Ph A Voltage,A Olinda Ph B Voltage,B Olinda Ph C Voltage,C Tracy Ph A Voltage,A Tracy Ph B Voltage,B Tracy Ph C Voltage,C Fault Indicator,N Olinda Ph A Current,A Olinda Ph B Current,B Olinda Ph C Current,C Tracy Ph A Current,A Tracy Ph B Current,B Tracy Ph C Current,C DATA TYPE: ASCII 96

PAGE 108

APPENDIX C TYPICAL INPUT SESSION USING THE DEFAULT PARAMETER MENU OPTION *********************************************** **** SIMULATOR PROGRAM FOR RELAY TESTING **** **** VERSION 2.0 **** *********************************************** MAIN MENU: 1) SELECT SIMULATOR FILE AND USE THE DEFAULT D/A PORT ASSIGNMENT 2) SELECT SIMULATOR FILE AND DEFINE D/A PORT ASSIGNMENT 10) SET PROGRAM PARAMETERS 99) EXIT THE PROGRAM ENTER THE MENU OPTION: 1 PLEASE ENTER THE NAME OF THE CONFIGURATION FILE TO BE USED FOR THE RUN. THE ".cfg" SUFFIX IS ASSUMED UNLESS THE USER SUPPLIES ANOTHER ONE. (TYPE "END" TO EXIT OR "DIR" FOR DIRECTORY) OTI-1A PROCESSING DATA .. PLEASE VAIT. 97

PAGE 109

THE DEFAULT PARAMETER ASSUMPTIONS ARE: D/A OUTPUT PORT 1: TRIGGER SIGNAL D/A OUTPUT PORTS 2,3,4: CURRENTS FOR TERMINAL #1 D/A OUTPUT PORTS 5,6,7: VOLTAGES FOR TERMINAL #1 D/A OUTPUT PORT 8: TRIGGER SIGNAL D/A OUTPUT PORTS 9,10,11: CURRENTS FOR TERMINAL #2 D/A OUTPUT PORTS 12,13,14: VOLTAGES FOR TERMINAL #2 ASSIGNMENT OF THE CURRENTS AND VOLTAGES TO SPECIFIC D/A OUTPUT PORTS: THE BRANCH CURRENT BUS NAMES ARE: 1) Olinda Ph A Current 2) Olinda Ph B Current 3) Olinda Ph C Current 4) Tracy Ph A Current 5) Tracy Ph B Current 6) Tracy Ph C Current ENTER THE NUMBER FOR THE CURRENTS AT TERMINAL #1 PHASE A: 1 PHASE B: 2 PHASE C: 3 THE NODE VOLTAGE BUS NAMES ARE: 1) Olinda Ph A Voltage 2) Olinda Ph B Voltage 3) Olinda Ph C Voltage 4) Tracy Ph A Voltage 5) Tracy Ph B Voltage 6) Tracy Ph C Voltage 7) Fault Indicator ENTER THE NUMBER FOR THE VOLTAGE AT TERMINAL #1 PHASE A: 1 PHASE B: 2 PHASE C: 3 98

PAGE 110

ENTER THE CURRENT TRANSFORMER RATIO AT TERMINAL #1 1 ENTER THE POTENTIAL TRANSFORMER RATIO AT TERMINAL #1 1 THE BRANCH CURRENT BUS NAMES ARE: 1) Olinda Ph A Current 2) Olinda Ph B Current 3) Olinda Ph C Current 4) Tracy Ph A Current 5) Tracy Ph B Current 6) Tracy Ph C Current ENTER THE NUMBER FOR THE CURRENTS AT TERMINAL #2 PHASE A: 4 PHASE B: 5 PHASE C: 6 THE NODE VOLTAGE BUS NAMES ARE: 1) Olinda Ph A Voltage 2) Olinda Ph B Voltage 3) Olinda Ph C Voltage 4) Tracy Ph A Voltage 5) Tracy Ph B Voltage 6) Tracy Ph C Voltage 7) Fault Indicator ENTER THE NUMBER FOR THE VOLTAGE AT TERMINAL #2 PHASE A: 4 PHASE B: 5 PHASE C: 6 ENTER THE CURRENT TRANSFORMER RATIO AT TERMINAL #2 1 ENTER THE POTENTIAL TRANSFORMER RATIO AT TERMINAL #2 1 99

PAGE 111

IS THERE AN EVENT TRIGGER SIGNAL (Y OR N)? y THE POSSIBLE EVENT TRIGGERS ARE: 1) Olinda Ph A Voltage 2) Olinda Ph B Voltage 3) Olinda Ph C Voltage 4) Tracy Ph A Voltage 5) Tracy Ph B Voltage 6) Tracy Ph C Voltage 7) Fault Indicator ENTER THE NUMBER FOR THE EVENT TRIGGER SIGNAL: 7 LISTING OF THE D/A PORT ASSIGNMENTS, C.T. RATIO AND P.T. RATIO D/A PORT 1: OTI-1A.dat TR Fault Indicator D/A PORT 2: OTI-1A.dat BC PH. A CTR= 1. Olinda Ph A Current D/A PORT 3: OTI-1A.dat BC PH. B CTR= 1. Olinda Ph B Current D/A PORT 4: OTI-1A.dat BC PH. C CTR= 1. Olinda Ph C Current D/A PORT 5: OTI-1A.dat NV PH. A PTR= 1. Olinda Ph A Voltage D/A PORT 6: OTI-1A.dat NV PH. B PTR= 1. Olinda Ph B Voltage D/A PORT 7: OTI-1A.dat NV PH. C PTR= 1. Olinda Ph C Voltage D/A PORT 8: OTI-1A.dat TR Fault Indicator D/A PORT 9: OTI-1A.dat BC PH. A CTR= 1. Tracy Ph A Current D/A PORT 10: OTI-1A.dat BC PH. B CTR= 1. Tracy Ph B Current D/A PORT 11: OTI-1A.dat BC PH. C CTR= 1. Tracy Ph C Current D/A PORT 12: OTI-lA.dat NV PH. A PTR= 1. Tracy Ph A Voltage D/A PORT 13: OTI-1A.dat NV PH. B PTR= 1. Tracy Ph B Voltage D/A PORT 14: OTI-1A.dat NV PH. C PTR= 1. Tracy Ph C Voltage DO YOU WANT TO MODIFY THE TESTING PARAMETERS? (Y OR N) N PROCESSING DATA PLEASE WAIT. 100

PAGE 112

SIMULATOR RUN MENU: 1) CYCLE ZERO OUTPUT . ENTER "CTRL C" TO CYCLE STEADY STATE .. ENTER "CTRL C" TO CYCLE ZERO OUTPUT. 2) CYCLE ZERO OUTPUT ENTER "CTRL C" TO CYCLE STEADY STATE .. ENTER "CTRL C" TO EXECUTE THE ENTIRE INPUT FILE . AUTOMATICALLY CYCLE ZERO OUTPUT AT THE END OF RUN. 10) VIEW THE D/A OUTPUT PORT ASSIGNMENTS 90) RETURN TO THE MAIN MENU ENTER THE MENU OPTION: 90 MAIN MENU: 1) SELECT SIMULATOR FILE AND USE THE DEFAULT D/A PORT ASSIGNMENT 2) SELECT SIMULATOR FILE AND DEFINE D/A PORT ASSIGNMENT 3) USE PREVIOUS D/A PORT ASSIGNMENT WITH A NEW SIMULATOR FILE '10) SET PROGRAM PARAMETERS 99) EXIT THE PROGRAM ENTER THE MENU OPTION: 99 101

PAGE 113

APPENDIX D ABBREVIATED INPUT SESSION FOR DEFINING EACH D/A PORT *********************************************** **** SIMULATOR PROGRAM FOR RELAY TESTING **** **** VERSION 2.0 **** *********************************************** MAIN MENU: 1) SELECT SIMULATOR FILE AND USE THE DEFAULT D/A PORT ASSIGNMENT 2) SELECT SIMULATOR FILE AND DEFINE D/A PORT ASSIGNMENT 10) SET PROGRAM PARAMETERS 99) EXIT THE PROGRAM ENTER THE MENU OPTION: 2 PLEASE ENTER THE NAME OF THE CONFIGURATION FILE TO BE USED FOR THE RUN. THE cfg'.' SUFFIX IS ASSUMED UNLESS THE USER SUPPLIES ANOTHER ONE. (TYPE "END" TO EXIT OR "DIR" FOR DIRECTORY) OTI-1A PROCESSING DATA PLEASE YAIT. 102

PAGE 114

ASSIGNMENT OF THE CURRENTS AND VOLTAGES TO SPECIFIC D/A OUTPUT PORTS: ENTER THE TYPE OF WAVEFORM FOR D/A OUTPUT PORT 1: BC BRANCH CURRENTS NV NODE VOLTAGES TR TRIGGER SIGNAL (NODE VOLTAGE) NU D/A OUTPUT PORT NOT USED RU REMAINING D/A OUTPUT PORT NOT USED EX ABORT INPUT SESSION; RETURN TO MAIN MENU TR IS THERE AN EVENT TRIGGER SIGNAL (Y OR N)? y THE POSSIBLE EVENT TRIGGERS ARE: 1) Olinda Ph A Voltage 2) Olinda Ph B Voltage 3) Olinda Ph C Voltage 4) Tracy Ph A Voltage 5) Tracy Ph B Voltage 6) Tracy Ph C Voltage 7) Fault Indicator ENTER THE NUMBER FOR D/A OUTPUT PORT 1: 7 ENTER THE TYPE OF WAVEFORM FOR D/A OUTPUT PORT 2: BC BRANCH CURRENTS NV NODE VOLTAGES TR TRIGGER SIGNAL (NODE VOLTAGE) NU D/A OUTPUT PORT NOT USED RU REMAINING D/A OUTPUT PORT NOT USED EX ABORT INPUT SESSION; RETURN TO MAIN MENU BC 103

PAGE 115

THE BRANCH CURRENT BUS NAMES ARE: 1) Olinda Ph A Current 2) Olinda Ph B Current 3) Olinda Ph C Current 4) Tracy Ph A Current 5) Tracy Ph B Current 6) Tracy Ph C Current ENTER THE NUMBER FOR D/A OUTPUT PORT 2: 1 ENTER THE PHASE ASSOCIATED D/A OUTPUT PORT 2: A ENTER THE CURRENT TRANSFORMER RATIO 1 ENTER THE TYPE OF FOR D/A OUTPUT PORT 3: BC BRANCH CURRENTS NV NODE VOLTAGES TR TRIGGER SIGNAL (NODE VOLTAGE) NU D/A OUTPUT PQRT NOT USED RU REMAINING D/A OUTPUT PORT NOT USED EX ABORT INPUT SESSION; RETURN TO MAIN MENU NV THE NODE VOLTAGE BUS NAMES ARE: 1) Olinda Ph A Voltage 2) Olinda Ph B Voltage 3) Olinda Ph C Voltage 4) Tracy Ph A Voltage 5) Tracy Ph B Voltage 6) Tracy Ph C Voltage 7) Fault Indicator ENTER THE NUMBER FOR D/A OUTPUT PORT 3: 1 ENTER THE PHASE ASSOCIATED VITH D/A OUTPUT PORT 3: A 104

PAGE 116

ENTER THE POTENTIAL TRANSFORMER RATIO 1 ENTER THE TYPE OF VAVEFORM FOR D/A OUTPUT PORT 4: BC BRANCH CURRENTS NV NODE VOLTAGES TR TRIGGER SIGNAL (NODE VOLTAGE) NU D/A OUTPUT PORT NOT USED RU REMAINING D/A OUTPUT PORT NOT USED EX ABORT INPUT SESSION; RETURN TO MAIN MENU RU LISTING OF THE D/A OUTPUT PORT ASSIGNMENTS, RATIO D/A PORT 1: D/A PORT 2: D/A PORT 3: D/A PORT 4: D/A PORT 5: D/A PORT 6: D/A PORT 7: D/A PORT 8: D/A PORT 9: D/A PORT 10: D/A PORT 11: DIA PORT 12: D/A PORT 13: D/A PORT 14: OTI-lA.dat OTI-lA.dat OTI-lA.dat OTI-lA.dat OTI-lA.dat OTI-lA.dat OTI-1A.dat OTI-1A.dat OTI-lA.dat OTI-lA.dat OTI-lA.dat OTI-1A.dat OTI-lA.dat OTI-lA.dat TR BC PH. NV PH. NU NU NU NU NU NU NU NU NU NU NU A CTR"' 1. A PTR= 1. C.T. RATIO AND P.T. Fault Indicator Olinda Ph A Current Olinda Ph A Voltage DO YOU VANT TO MODIFY THE TESTING PARAMETERS? (Y OR N) N PROCESSING DATA ... PLEASE VAIT. 105

PAGE 117

SIMULATOR RUN MENU: 1) CYCLE ZERO OUTPUT ... ENTER "CTRL C" TO CYCLE STEADY STATE ... ENTER "CTRL C" TO CYCLE ZERO OUTPUT. 2) CYCLE ZERO OUTPUT .. ENTER "CTRL C" TO CYCLE STEADY STATE ENTER "CTRL C" TO EXECUTE THE ENTIRE INPUT FILE .. AUTOMATICALLY CYCLE ZERO OUTPUT AT THE END OF RUN. 10) VIEY THE D/A OUTPUT PORT ASSIGNMENTS 90) RETURN TO THE MAIN MENU ENTER THE MENU OPTION: 90 MAIN MENU: 1) SELECT SIMULATOR FILE AND USE THE DEFAULT D/A PORT ASSIGNMENT 2) SELECT SIMULATOR FILE AND DEFINE D/A PORT ASSIGNMENT 3) USE PREVIOUS D/A PORT ASSIGNMENT YITH A NEY SIMULATOR FILE 10) SET PROGRAM PARAMETERS 99) EXIT THE PROGRAM ENTER THE MENU OPTION: 99 106

PAGE 118

BIBLIOGRAPHY [1] C. R. Mason, The Art and Science of Protective Relaying, John Wiley and Sons, T:nc., New York, NY, 1956. [2] R. P. DePuy, L. E. Goff, and W. C. New, "Reclosing Practices -Past and Present," Proceedings, 1983 Western Protective Relay Conference, Spokane. Washington State University. [3] C. F. Henville and J. A. Jodice, "Discover Relay Design and Application Problems Using Pseudo-Transient Tests," Proceedings, 45th Annual Georgia Tech Protective Relaying Conference, Atlanta, 1991. Georgia Institute of Technology. [4] G. E. Alexander, P. J. Lerley, and R. Ryan, "Comparative Testing Using Digital Simulation and an Analog Model Power System," Proceedings, 1990 Western Protective Relay Conference, Spokane. Washington State University. [5] M. G. Adamiak, G. E. Alexander, and J. G. Andrichak, "Power System Simulation: A High Power Amplifier Approach," Proceedings, 1992 Western Protective Relay Conference, Spokane. Washington State University. [6] A. Greenwood, Electrical Transients in Power Systems, Second Edition, John Wiley and Sons, Inc., New York, NY, 1991. [7] "Electromagnetic Transients Program (EMTP) Revised Rule Book Version 2.0," Report EL-6421-L, EPRI, Palo Alto, CA. [8] J. Esztergalyos, J. Nordstrom, T. H. Short, and K. Martin, "Digital Model Power System," IEEE Computer Applications in Power, Vol. 3, No. 3, pp. 19-24, July 1990. [9] M. Kezunovic, A. Abur, Lj. Kojovic, V. Skendzic, H. Singh, C. W. Fromen, and D. R. Sevcik, "DYNA-TEST Simulator for Relay Testing, Part I: Design Characteristics," IEEE Transactions on Power Delivery, Vol. 6, No. 4, pp:-1423-1429, October 1991. [10] A. K. Chaudhary, J. B. Anich, and J. B. Wisniewski, "Influence of Transient Response of Instrument Transformers on Protection Systems," Proceedings, Twelfth Biennial Transmission and Substation Conference, November 1992. Sargent and Lundy. 107

PAGE 119

[11] M. Kezunovic, A. Abur, Lj. Kojovic, V. Skendzic, H. Singh, C. Til. Fromen, and D. R. Sevcik, "DYNA-TEST Simulator for Relay Testing, Part II: Performance Evaluation," IEEE Transactions on Power Delivery, Vol. 7, No. 3, pp:-I097-1103, July 1992. [12] M. Kezunovic, Lj. Kojovic, V. Skendzic, C. Y. Fromen, D. R. Sevcik, and S. L. Nilsson, "Digital Models of Coupling Capacitor Voltage Transformers for Protective Relay Transient Studies," IEEE Transactions on Power Delivery, Vol. 7, No. 4, pp. 1927-1935, October 1992. [13] IEEE Standard, "IEEE Standard Common Format for Transient Data Exchange (COMTRADE) for Power Systems," IEEE C37.1111991, IEEE Inc., New York, NY, 1991. [14] G. D. Bergland, "A Guided Tour of the Fast Fourier Transform," Digital Signal Processing, Edited by L. R. Rabiner and C. M. Rader, pp. 228-239, IEEE Press, New York, NY, 1972. [15] "Data Acquisition Application Programming Manual," Order No. M-SP50-AP, Concurrent Computer Corporation, 1990. [16] K. D. Simpson, "California-Oregon 500 kV Transmission Line Development of Design Criteria," Proceedings, Tenth Biennial Transmission and Substation Conference, November 1988. Sargent and Lundy. [17] "COTP 500 kV Protective Relays Request for Proposals," Solicitation Number DE-RP65-90YN06808, United States Department of Energy, 1990. [18] M. Kezunovic, M. Aganagic, V. Skendzic, J. Domaszewicz, J. K. Bladow, D.M. Hamai, S.M. McKenna, "Transients Computation for Relay Testing in Real-Time," IEEE PES Summer Meeting, Paper No. 93SM 383-0-PWRD, Vancouver, B.C., Canada, July 1993. [19] P. G. McLaren, R. Kuffel, R. Yierckx, J. Giesbrecht, L. Arendt, "A Real Time Digital Simulator for Testing Relays," IEEE Transactions on Power Delivery, Vol. 7, No. 1, pp. 207213, Jan. 1992. 108