http : //www.cigre.org B2-1030 AORC Technical meeting 2014 Implementation Approaches on Fault Information Analyzing System In Thailand s Power System N.AKEKURANANT S.CHAMNANVANICHKUL Electricity Generating Authority of Thailand THAILAND SUMMARY In order to improve the reliability and security of Thailand s transmission system, Electricity Generating Authority of Thailand (EGAT) has continuously installed Digital Fault Recorder (DFR) units at almost every substation to cover all areas of Thailand, to comply with the smart grid projects. DFRs are recording many system disturbance data such as the voltages, currents, and various statuses of digital signals pertinent to transmission system protection, when it suspects that some fault may occur within the transmission system. The recorded data from DFR will be analyzed by the system analysts in order to determine the fault type, fault location, the cause of fault event, to examine how well the protection system operates, as well as the system response characteristics when different fault types do occur. Nowadays, the information from DFRs is extremely useful for the system operators at the control center to understand the root cause problems in the disturbance event and solve the problem efficiently. Automated fault analysis system is developed as the tool to help the system operators and system analysts to access the fault event data instantaneously and conveniently via web browser. It helps to deal with the recorded data files and preliminary analyzing the fault event. The system operator will have sufficient information to decide to bring the system back as soon as possible. It is very useful for the system restoration and the reduction on the time of power failure in the system. The automated fault analysis system has the features to fulfill the system operation so-called the fault Information analyzing system. The purpose of the fault information analyzing system is to provide the information of the disturbance event for the system operator. In case of fault occurred on the transmission line, the system will be analyzed a type of fault and the location of fault event and present to the system operator automatically. The fault Information analyzing system composes of three modules. The first module, prescreening module is working to determine whether the fault occurs on the transmission line, by monitoring the condition of the relay, examining the current and voltage of the transmission line, including consideration of polarized current. The second module is the fault type classification module, using current phasor and wavelet transform technique. It will segregate the fault type in the disturbance event. The last module is the fault distance location module, relying on the concept of one terminal fault location to estimate distances from the station to the point that the short circuit are occurred. This paper is represented to the method for filtering recorded data from DFRs, classification the fault type and calculation the fault distance on the transmission line. It also proposes an approach to develop and improve the system automatically analyzes the most effective for the implementation stage in the smart grid environment. Due to the distinctive advantages of automated analysis is speed of processing, but may not be able to complexity analyze like people. Therefore, a lot of data are stored in the database should help in predicting the location of the fault or other properties more accurately. This paper will be benefit to Thailand power system as a whole to improve the system performance indices and also enhance the performance of the protection system and high voltage equipment. KEYWORDS Automated Fault Analysis, Digital Fault Recorder, Smart Grid, System Restoration, Fault Location, Fault Type Detection nitanan.a@egat.co.th
1. Introduction Electricity Generating Authority of Thailand (EGAT) has continuously installed Digital Fault Recorder (DFR) units at almost every substation to cover all areas of Thailand to record essentially the voltages, currents, and various statuses of digital signals pertinent to transmission system protection. When it suspects that some fault may occur within the transmission system, the recorded data from DFRs will then be analyzed by system analysts in order to determine the type and location of fault, the cause of fault event, to examine how well the protection system operates, as well as, the system analysts to learn about the system response characteristics when different fault types do occur. Nowadays, the information from DFRs is extremely useful for the system operators to understand the root cause problems in the disturbance event and solve the problem efficiently. The Automated Fault Analysis system has the features to fulfill the system operation socalled the fault information analyzing system. The purpose of this system is to provide the information of the disturbance event for the system operator. In case of fault occurred on the transmission line, the system will be analyzed a type of fault and the location of fault event and present to the system operator automatically, which will help the system operator to understand the situation and select the appropriate method to solve the problem. This paper is represented to the calculation method in the fault information analyzing system. The system composes of three modules. First module is filtering recorded data from DFRs, the second module is classification on the fault type and the last module is calculation the fault distance on the transmission line. Not only it also purposes an approach to develop and improve the system automatically, but also analyzes the most effective method for the implementation stage in the smart grid environment. 2. The fault information analyzing system Figure 1: Automated Fault Analysis System (AFA) The fault information analyzing system is completely installed in EGAT s system and was first used in 2012. EGAT call the system is Automated Fault Analysis System (AFA). AFA is help the system operator making the decision for restoring the system back to normal, finding information on the past fault event to use comparable with weak point in the system, improving the operating of protection equipment or other equipment in the power system. The basic concept of AFA is a centralized system to retrieve data from DFRs that are installed in the high voltage substation. It will convert the recorded data from DFRs to the standard COMTRADE format and preliminary analysis by the algorithm specified. The details of each analyzed fault event are automatically displayed via an EGAT s intranet, employing web-based technology, for all users able to access information quickly, effective and timely. 3. Application development of fault information analyzing system. AFA is identifying unusual events that are occurring in the power system. If some fault is occurred in transmission line, the system can indicate the fault type and the fault location. There are 3 modules as following. 2
3.1 Prescreening Module The purpose of the module is to check the recorded from DFRs for deciding whether the fault occurred in the transmission line. If a fault is occurred in transmission line, this module can indicate which the transmission lines have some problems and send this analysis data to another modules. It will be considered from many factors, such as - The status of the distance relay (94, 21) and/or operating of circuit breaker (52) - The voltage will drop quickly and the current is much higher than nominal value - Polarized current is greater than zero. AFA used the information from the status of the distance relay to recognize a fault pattern on a transmission line and the recorded signals in the DFRs to check the condition within the short duration. When fault occurred in the transmission lines, the circuit breaker between two substations will operate to disconnect the transmission line from the power system. The protective relay will send the command trip, therefore, AFA will using the status of the protective relay and evaluate the fault scenario. AFA will also receive the voltage and current data to analyst the event based on the principle for the fault condition. With the circuit breaker and information from the system, AFA will compute for the derivative value of RMS current. It can distinguish the fault event and the normal condition to define the starting point and ending point of the event. 3.2 Fault Type Classification Module The purpose of the module is to define the type of fault, based on the analysis of the magnitude and phase of the fault current and sequence components. Due to the classification of a fault with the different characteristics, AFA can match the fault pattern. There are 4 groups of fault types : three phase fault (A-B-C), single line-to-ground fault, SLGF (A-G, B-G, C-G), double line-to-ground fault, DLGF (A-B-G, B-C-G, C-A-G) and line-to-line fault, LLF (A-B, B-C, C-A). There are several ways to classify the fault type such as phasor current (use the magnitude of fault current, phase angle and magnitude of sequence components current), wavelet transform, fuzzy logic, neural network and so on. Phase angle and magnitude of sequence components current are selected method for categorizing the fault type in AFA system. It is general method using with the effective result without complexity of mathematics. However, these methods are modified for appropriate with the EGAT s system. Indexes are used in the classification of fault are as following: 1) The difference of the phase angle of the phasor current and the positive sequence component current with reference to the phase. ang_a arg(i afault ) arg(i a1fault ) ang_b arg(i bfault ) arg(i b1fault ) ang_c arg(i cfault ) arg(i c1fault ) 2) The difference of the phase angle of the positive sequence component current and the negative sequence component current with reference to the phase. ang_a 12 arg(i a1fault ) arg(i a2fault ) ang_b 12 arg(i b1fault ) arg(i b2fault ) ang_c 12 arg(i c1fault ) arg(i c2fault ) 3) The proportion between the size of the zero sequence component current and negative sequence component current per the positive sequence component current. By that arg(i afault ), arg(i bfault ), arg(i cfault ) are the phase angle of current phase A, B,C during fault are occurred. arg(i a1fault ), arg(i b1fault ), arg(i c1fault ), arg(i a2fault ), arg(i b2fault ), arg(i c2fault ) are the phase angle of positive and negative sequence component current phase A, B,C during fault are occure. I 0Fault, I 1Fault, I 2Faul are the zero, positive and negative sequence component current during fault are occure. In recognition of fault in 10 categories can be identified by using various criteria. In terms of unbalance fault such as A-G fault. ang_a 12 arg(i a1fault ) arg(i a2fault ) 0 ang_b 12 arg(i b1fault ) arg(i b2fault ) 120 ang_c 12 arg(i c1fault ) arg(i c2fault ) 120 3
>, > In terms of balance fault such as A-B-C fault. ang_a arg(i afault ) arg(i a1fault ) 0 ang_b arg(i bfault ) arg(i b1fault ) 0 ang_c arg(i cfault ) arg(i c1fault ) 0 Table 1: Unbalance fault type is classified by improving the deviation from theory. Fault Type ang_a 12 ang_b 12 ang_c 12 R 0Fault R 2Fault A-G 0 ± 30 120 ± 30 120 ± 30 >0.15 >0.15 B-G 120 ± 30 0 ± 30 120 ± 30 >0.15 >0.15 C-G 120 ± 30 120 ± 30 0 ± 30 >0.15 >0.15 A-B 60 ± 30 60 ± 30 180 ± 30 <0.15 >0.15 B-C 180 ± 30 60 ± 30 60 ± 30 <0.15 >0.15 C-A 60 ± 30 180 ± 30 60 ± 30 <0.15 >0.15 A-B-G 60 ± 30 60 ± 30 180 ± 30 >0.15 >0.15 B-C-G 180 ± 30 60 ± 30 60 ± 30 >0.15 >0.15 C-A-G 60 ± 30 180 ± 30 60 ± 30 >0.15 >0.15 δ and ε are the threshold that users can customize. It is an index to indicate the properties of fault in each category. Because the effect from the fault resistance has a value not equal to zero and vary by time. Cause various degrees of deviation according to the theory, as above. To appropriate the Thailand electrical system, from experimenting with electrical signals, the conditions for balanced fault resolution should be between ± 10 from theories and ± 30 for unbalance fault. In case of ground fault, δ and ε should be more than 0.15 and can be modified in the future. 3.3 Fault Distance Location Module The purpose of the module is to approximate the distance from the substation to the fault point. For example, the staff workers look forward the fault point and continue solve problem in a short time. The calculation relies on the concept of one terminal fault location, follow by IEEE Std. C37.114-2004. The formula to calculate the distance varies by the type of fault. Figure 2: The system when fault are occurring between 2 substations. Normally, when fault are occurring in the power system, circuit breakers will operate by opening the circuit on both sides of the transmission line. From figure 2, If the substation is installed DFR at DFR1 substation and DFR2 substation. The analysis for transmission line issues is based on the digital signal of circuit breaker status or relay control signal recorded by DFR. In this case, the two-terminal algorithm technique is used to find the location of fault on the transmission line. If DFR is installed only one side, the one-terminal algorithm technique is used to calculate the distance between substations to the fault point. If DFR is not installed at the end of both sides of the transmission line, the fault distance will calculate from current and voltage from DFR at next substation by multi-terminal algorithm. However, the results from this method are usually calculated by a high tolerance because this is an approximation without actual measurement data. Therefore, AFA system only uses one-terminal and two-terminal algorithm. 4
3.3.1 One-Terminal Algorithm Figure 3: Equivalence circuit of transmission lines for one-terminal fault location calculation. It is starts from a simple equivalence circuit of transmission lines caused by the fault. Write the equation represents the relationship, according to Kirchoff 's Low as follows. + (1) + (2) ( ) ($ % ) Define V g : Voltage at g side, m : Distance from V g to fault point, Z L : Line impedance, I g : Current flow from g side, R F : Fault resistance, I F : Fault current In the general case, fault impedance is only part of the resistance and there is very little value. If we neglect the effect of fault voltage (I F R F ), it possible to find the position that provides the distance from one end of a transmission line (m) according to equation (3). It is called a "Simple reactance method" and the value of the imag (V g /I g ) will vary depending on the type of fault. For example, the distance from a single phase to ground (phase A-G) fault on a transmission line (m) was calculated from equation (4) ' & ' ( ) % * +), -. ( (4) % ), ' * % In case of the double line fault (phase B-C) are occurred, m was calculated from of the station in the left end. ) (3) / 01 3 ( (5) ) 01 2 In case of double line to ground fault (phase B-C-G), if 4 5 0, using the equation (5) in order to locate a fault in the transmission line. For another fault type, use the formula from table 2 to calculate the distance. Table 2: Summary of formulas for calculated the distance in any case of fault. Fault Type Positive-Sequence impedance equation ( ) A-G V 9 /( 4 +ki ) B-G V > /(? +ki ) C-G V @ /( A +ki ) A-B or A-B-G V 9> / 4? B-C or B-C-G V >@ /?A C-A or C-A-G A-B-C V @9 / A4 V 9> / 4? or V >@ /?A or V @9 / A4 By that B ( /3 ) E 3 4 3.3.2 Two-Terminal Algorithm The recorded DFR data, are received from different substation, are different in timestamp. Some EGAT s substation could not install the GPS receiver for time synchronization. Time difference has a dramatic effect on the accuracy of the calculations. So this AFA project use one-terminal 5
algorithm to calculate the distance from each substation and calculate the average distance of fault occurred. 4. Approach to develop and improve the system In the prescreening module, the future improvement to enhance the capabilities of the AFA program is to classified event as group events. The filter should be divided the incident into the location, and significant level. This module should be the expert system using the past information and the experience. It will help the system operator to screen the event with the short duration. For the fault type classification Module, the future development is to consider for the algorithm defined the fault efficiently. The system parameter need to be changed to the adaptive parameters as the system modified. The other method such as the wavelet transform method also consider as the alternative solution. The test and data collection need to be done to prove the correction and capability of the algorithm. For the fault distance location module, the upgrade module for more precise calculation needs to be examined. The time synchronization with the GPS is the major method to perform in the first improvement stage. The new approach need to be considered, especially for the mutual line effects. The data and evaluation will be carried out to ensure the correct distance. 5. Conclusion The AFA system will manage the recorded data from DFRs and analyze for a type of fault, the location of fault event and present the information to the system operator automatically via the website. It will help the system operator to understand the situation and select the appropriate method to solve the problem, instantaneously. As the distinctive advantage of automated analysis is speed of processing, but may not be able to assist the complexity analyze. Therefore, a lot of data are stored in the database should help in predicting the location of the fault or other properties more accurately. Moreover, all data in the database can also be used to develop an expert system (pattern recognition) in the future. This paper will be benefit to Thailand power system as a whole to improve the system performance indices, enhance the performance of the protection system and high voltage equipment and also proposes an approach to develop and improve the system automatically analyzes the most effective for the implementation stage in the smart grid environment. BIBLIOGRAPHY [1] Nitanan Akekuranant, Sompol Chamnanvanichkul, Automated Fault Analysis System in Thailand Power System, CIGRE-AORC 2013 [2] Mohamed A. Lbrahim, Disturbance Analysis for Power System, John Wiley & Sons, 2011 [3] Alexander McEachern, Hand Book of Power Signatures, Basic Measuring Instruments, 1989 [4] Institute of Electrical and Electronic Engineers, Application and evaluation of automatic fault recording devices. IEEE Committee Report, IEEE Transactions on Power Apparatus & Systems, 1965 [5] Joe Perez, A GUIDE TO DIGITAL FAULT RECORDING EVENT ANALYSIS, Fault and Disturbance Analysis Conference 2010. Atlanta, Georgia, USA [6] North American Electric Reliability Corporation, NERC Standard PRC-018-1, Disturbance Monitoring Equipment Installation and Data Reporting. Author, Princeton, New Jersey, USA, 2006 [7] Center of Excellence in Electrical Power Technology, System Overview and AFA Architecture. [8] Center of Excellence in Electrical Power Technology, Admin Manual for AFA-Web 2.0 Short Bio-data of Main Author Nitanan Akekuranant graduated Master degree in Electrical Engineering from Kasetsart University, Thailand in 2000. She has join Electricity Generating Authority of Thailand (EGAT) since 2002. She is presently working in Power System Analysis Division. Her current role is to provide technical service in the fault recording system area and disturbance analysis. She has involved in various fault recording system and disturbance event analysis including the Automated Fault Analysis System project. 6