Failure analysis and maintenance of a surge capacitor on the neutral bus in a ±500 kV HVDC converter station
2016; Institution of Engineering and Technology; Volume: 10; Issue: 6 Linguagem: Inglês
10.1049/iet-rpg.2015.0451
ISSN1752-1424
AutoresMing Yang, Wenxia Sima, Xue Han, Rong Wang, Chilong Jiang, Wenqi Mao, Potao Sun,
Tópico(s)Electrical Fault Detection and Protection
ResumoIET Renewable Power GenerationVolume 10, Issue 6 p. 852-860 Special Issue: DC and HVDC System TechnologiesFree Access Failure analysis and maintenance of a surge capacitor on the neutral bus in a ±500 kV HVDC converter station Ming Yang, Ming Yang State Key Laboratory of Power Transmission Equipment and System Security and New Technology,, and College of Automation, Chongqing University, Chongqing, 400044 People's Republic of ChinaSearch for more papers by this authorWenxia Sima, Corresponding Author Wenxia Sima cqsmwx@cqu.edu.cn State Key Laboratory of Power Transmission Equipment and System Security and New Technology,, and College of Automation, Chongqing University, Chongqing, 400044 People's Republic of ChinaSearch for more papers by this authorXue Han, Xue Han State Key Laboratory of Power Transmission Equipment and System Security and New Technology,, and College of Automation, Chongqing University, Chongqing, 400044 People's Republic of ChinaSearch for more papers by this authorRong Wang, Rong Wang State Grid Chengdu Power Supply Company, Chengdu,, Sichuan, 610041 People's Republic of ChinaSearch for more papers by this authorChilong Jiang, Chilong Jiang State Grid Hunan Electrical Power Corporation Maintenance Corporation, Changsha, Hunan, 410004 People's Republic of ChinaSearch for more papers by this authorWenqi Mao, Wenqi Mao State Grid Hunan Electrical Power Corporation, Changsha, Hunan, 410000 People's Republic of ChinaSearch for more papers by this authorPotao Sun, Potao Sun State Key Laboratory of Power Transmission Equipment and System Security and New Technology,, and College of Automation, Chongqing University, Chongqing, 400044 People's Republic of ChinaSearch for more papers by this author Ming Yang, Ming Yang State Key Laboratory of Power Transmission Equipment and System Security and New Technology,, and College of Automation, Chongqing University, Chongqing, 400044 People's Republic of ChinaSearch for more papers by this authorWenxia Sima, Corresponding Author Wenxia Sima cqsmwx@cqu.edu.cn State Key Laboratory of Power Transmission Equipment and System Security and New Technology,, and College of Automation, Chongqing University, Chongqing, 400044 People's Republic of ChinaSearch for more papers by this authorXue Han, Xue Han State Key Laboratory of Power Transmission Equipment and System Security and New Technology,, and College of Automation, Chongqing University, Chongqing, 400044 People's Republic of ChinaSearch for more papers by this authorRong Wang, Rong Wang State Grid Chengdu Power Supply Company, Chengdu,, Sichuan, 610041 People's Republic of ChinaSearch for more papers by this authorChilong Jiang, Chilong Jiang State Grid Hunan Electrical Power Corporation Maintenance Corporation, Changsha, Hunan, 410004 People's Republic of ChinaSearch for more papers by this authorWenqi Mao, Wenqi Mao State Grid Hunan Electrical Power Corporation, Changsha, Hunan, 410000 People's Republic of ChinaSearch for more papers by this authorPotao Sun, Potao Sun State Key Laboratory of Power Transmission Equipment and System Security and New Technology,, and College of Automation, Chongqing University, Chongqing, 400044 People's Republic of ChinaSearch for more papers by this author First published: 01 July 2016 https://doi.org/10.1049/iet-rpg.2015.0451Citations: 4AboutSectionsPDF ToolsRequest permissionExport citationAdd to favoritesTrack citation ShareShare Give accessShare full text accessShare full-text accessPlease review our Terms and Conditions of Use and check box below to share full-text version of article.I have read and accept the Wiley Online Library Terms and Conditions of UseShareable LinkUse the link below to share a full-text version of this article with your friends and colleagues. Learn more.Copy URL Share a linkShare onFacebookTwitterLinkedInRedditWechat Abstract This study analyses a surge capacitor failure on the neutral bus in a ±500 kV high voltage direct current (HVDC) converter station. Both overvoltage and insulation of the surge capacitor are investigated through simulation, practical record data analysis, and experimentation. A highly accurate HVDC model is established on the basis of the topological structure and parameters of the HVDC project. Using this model, the failure process is simulated by combining neutral bus overvoltage record analysis and voltage divider performance test. Results indicate that the 246 kV uncharacteristically high overvoltage on the neutral bus is a measurement error caused by voltage divider fault. From the field test, difference equation of the voltage divider is derived to reconstruct the actual overvoltage, the amplitude of which is 73 kV. The amplitude is significantly lower than that of the switching impulse withstand level of the surge capacitor, which should not cause capacitor failure. Therefore, the primary cause of capacitor failure is insulation damage rather than overvoltage. Finally, maintenance measures are proposed and applied in the station. 1 Introduction An increasing number of high-voltage direct current (HVDC) systems are currently being established around the world [1, 2] owing to these systems' evident advantages, such as considerably large power transmission capability, long-distance power transmission, interconnection of asynchronously operated power networks, and substantially low losses [3-5]. Overvoltage is one of the main problems threatening power system safety, particularly in HVDC transmission projects [6-10]. Thus, an analysis of the transient process and characteristics of transfer overvoltage in HVDC transmission systems would be significant for both system operation reliability and power equipment safety. Although voltage on the neutral bus is low in normal operational conditions, an increase in voltage may ensue when the system performs a certain action or when malfunction occurs [11]. Overvoltage on the neutral bus of a converter station severely threatens the insulation safety of power equipment, such as converter transformer, converter valve, smoothing reactor on the DC pole line, and DC surge capacitor [12, 13]. To limit the overvoltage to a safe value and to protect the equipment installed on the neutral bus, installing a metal oxide surge arrester on the neutral bus is essential. However, several accidents are still occurring in the HVDC system that cause high overvoltage on the neutral bus, thereby resulting in differential current protection and DC pole blocking. For example, a line-to-ground fault happened on the DC line of Pole II of a certain HVDC transmission project in China in May 2014. This fault caused an explosion of a DC surge capacitor on the neutral bus of the inverter station of Pole I. However, the responsible metal oxide surge arrester failed to limit the overvoltage. Eventually, this occurrence caused the blocking of the transmission system's Pole I. An accident investigation report indicated that the amplitude of the overvoltage on the neutral bus is as high as 246 kV. However, the switching impulse protective level (SIPL) of the responsible arrester is 75 kV, whereas the switching impulse withstand level (SIWL) is 95 kV. To determine this unusual phenomenon and the actual cause of this failure, both overvoltage and capacitor inner insulation problems are studied through simulation, record data analysis, and experimentation. 2 Jiangling–Echeng project 2.1 Project configuration and main parameters The Jiangling–Echeng ±500 kV HVDC bipolar power transmission project undertakes power delivery from Hubei Province to Guangdong Province in China, with a length of 940 km, rated capacity of 3,000 MW, rated voltage of ±500 kV, and rated DC current of 3 kA. The converter valves of the Jiangling–Echeng project are 12-pulse commutation cells in series with 1 group each pole, and the smoothing reactors are arranged on pole bus and neutral bus. The wiring diagram of the Echeng converter station of the Jiangling–Echeng ±500 kV HVDC project is shown in Fig. 1. The definition of the project's basic parameters is shown in Table 1, while the main parameters are shown in Table 2. Table 1. Definition of basic parameters of the Jiangling–Echeng ±500 kV HVDC project Parameter Description T1/T2 transformer VA/VB/VC valve bridge L1/L2 reactor C1/C2 capacitor Z1 filter T11 saturable reactor F1-3/V1-4/DB1/DB2/E1/E2/EL arrester Q11/Q12/Q13/Q21/Q22/Q23/Q24 switch Table 2. Main parameters of the Jiangling–Echeng ±500 kV HVDC project Parameter Value Parameter Value P(1,2)-WN-C1 7 μF P(1,2)-Z1-Z11-L1 11.71 mH P(1,2)-WN-L1 2 mH P(1,2)-Z1-Z11-C2 9.047 μF P(1,2)-WN-C2 16 μF P(1,2)-Z1-Z11-L2 5.84 mH WN-L1 2 mH P(1,2)-Z1-Z12-C1 2 μF WN-L2 13.64 mH P(1,2)-Z1-Z12-L1 6.46 mH WN-C1 0.065 μF P(1,2)-Z1-Z12-C2 3.725 μF WN-C2 9.54 nF P(1,2)-Z1-Z12-L2 11.35 mH WN-W1-L1 2 mH P(1,2)-WP-C1 2.52 μF WN-W1-C1 0.065 μF P(1,2)-WP-L1 290 mH P(1,2)-Z1-Z11-C1 2 μF Fig. 1Open in figure viewerPowerPoint Wiring diagram of the Echeng converter station of the Jiangling–Echeng ±500 kV HVDC project 2.2 Arrester configuration scheme in the Echeng station The detailed description of each arrester is shown in Table 3. The technical parameters of the arresters in the Jiangling–Echeng ±500 kV HVDC converter station are shown in Table 4. The arrester configuration scheme in the Jiangling–Echeng converter is shown in Table 5. Table 3. Definition of arresters in the converter station Arresters Description V1-V4 valve arrester DB1/DB2 DC line arrester E1 neutral bus arrester at the valve side E2 neutral bus arrester EL electrode line arrester F1 arrester crossing filter F2 parallel arrester of the capacitor F3 parallel arrester of the inductor Table 4. Basic parameters for the arresters in the Jiangling–Echeng HVDC converter station Arrester Lightning impulse protective level 8/20 μs, kV / kA SIPL 1 ms, kV / kA Maximum continuous operating voltage, kV peak Rated reference voltage, kV Energy capacity, MJ V 457/1 470/2 347 246 6.49 DB1 1124/20 936/1 515 631 4.92 DB2 1068/10 936/1 515 631 4.92 E1, E2 99/5 76/16.4 10 62 0.8 EL 82/10 76/16.4 10 54 26.2 Table 5. Arrester configuration scheme in the Jiangling–Echeng converter station Object Protected by Lightning impulse protect level, kV Lightning impulse withstand level, kV Margin, % Switching impulse protection level, kV SIWL, kV Margin, % DC bus line side DB1 1124 1425 27 936 1175 25 DC bus valve side 2*V + E1 1013 1550 53 1016 1300 27 2*V + EL across smoothing reactor Udmax + DB2 1583 1950 23 NA 1550 NA DC bus between 6-pulse bridges V + E1 NA NA NA NA NA NA V + EL Yy-transformer valve side 2*V + E1 1013 1675 65 1016 1300 27 2*V + EL Yd-transformer valve side V + E1 556 850 52 547 650 19 V + EL neutral bus (surge capacitors) E1,E2 99 125 26 76 95 25 EL Udmax — Maximum DC voltage, NA — not applicable 2.3 Equivalent model of the AC system In the model, the system at the AC side of the converter station is equivalent to the AC source. The maximum short-circuit current and short-circuit capacity of the AC systems of the converter stations at both ends are 63 kA and 57,288 MVA, respectively. The amplitude, phase angle, and source impedance can be calculated accordingly. The impedance angle of the power source is φ = 85°. 2.4 Model for the Jiangling–Echeng DC project On the basis of the actual parameters of the main equipment in the Jiangling–Echeng project, the simulation models for the converter station and DC transmission lines are established using PSCAD/EMTDC [14]. In equivalent models, the parameters of the converter transformer, smoothing reactor, AC filter, DC filter at both the AC and DC sides, and filter on the DC transmission lines are identical to those of the actual project. The established Jiangling–Echeng HVDC transmission system simulation model is shown in Fig. 2. Fig. 2Open in figure viewerPowerPoint Diagram of the Jiangling–Echeng HVDC transmission system simulation model To verify the validity of the simulation model, the operation steady state of the HVDC transmission system is simulated. Pole I and Pole II are running in the ground return loop with transmission power of 2109 MW as shown in Fig. 3a. From the system's recorded fault running condition, the rated line voltage is ±500 kV and the direct current is 2109 A. The simulation steps are set at 50 μs. Fig. 3Open in figure viewerPowerPoint Performance of the simulation model a Active power waveform of the rectifier/inverter side b DC voltage waveform of the DC lines In the model, the control system mainly includes a series of modules for closed-loop and open-loop control systems with negative feedback. The joint control method of the rectifier and inverter stations includes low-voltage current limiting control, constant current control, current deviation control, and constant extinction angle γ control [3, 15]. In Fig. 3b, P1 and P2 are the simulation voltages of Pole I and Pole II, respectively. The simulation results show that the DC line steady voltage reaches the rated ones of ±500 kV, and the simulated line current is 2.109 kA. These results are consistent with those of the actual project. 3 Failure simulation 3.1 Failure introduction The fault sequence diagram is shown in Fig. 4. After Pole I is blocked, the personnel who inspected the field determined that the surge capacitor on the neutral bus of Pole I was severely damaged and the capacitor oil sprayed. The fault locating system showed that the distance between the fault point and inverter station is approximately 522.6 km, which is on tower No. T1081. This fault location is shown in Fig. 2, while the failure capacitor is shown in Fig. 5. The waveforms of the DC line and neutral bus voltage obtained by the fault recorder system are shown in Fig. 6. As illustrated in Fig. 6a, when grounding fault occurred, the voltage polarity of the DC line (Pole I) reversed during the lock-up process of Pole I. Thereafter, the Pole I voltage decreases from 500 to −194 kV with a transient duration. Fig. 4Open in figure viewerPowerPoint Diagram of the fault sequence Fig. 5Open in figure viewerPowerPoint Failure capacitor a Location of the failure capacitor b Failure capacitor (outside and inside) Fig. 6Open in figure viewerPowerPoint Waveforms of the DC line and neutral bus voltage obtained by the fault recorder system a Waveform of the DC line voltage b Waveform of the neutral bus voltage 3.2 Failure simulation results To determine the process of the failure and its causes, a related simulation is conducted. The line fault is set as line-to-ground fault and the fault time is set at 0.5 s. The waveforms of the DC line and neutral bus voltage obtained through simulation are shown in Fig. 7. Compared with Fig. 6, the simulation waveform is similar with the fault recorder system, particularly the DC line voltage, which is highly similar with the recorded waveform. Fig. 7Open in figure viewerPowerPoint Waveform of the DC line and neutral bus voltage obtained through simulation a Waveform of the DC line voltage b Waveform of the neutral bus voltage Note that the neutral bus voltage amplitude is 73 kV according to the simulation, which is slightly lower than that of the responsible arrester's SIPL (76 kV). However, the voltage amplitude acquired from the record data is 246 kV, which is significantly higher than that of the arrester's SIPL (76 kV) and the surge capacitor's SIWL (95 kV). Nevertheless, the report shows that the responsible arrester took no action and the surge capacitor was severely damaged. Such a huge difference between the simulation and record should be analysed to determine the actual cause of this capacitor failure. Section 4 presents a detailed analysis of this failure. 4 Failure analysis The record indicates that the amplitude of the neutral bus overvoltage is 246 kV, which is significantly higher than the surge capacitor's SIWL on the neutral bus. However, the overvoltage amplitude of the Pole I neutral bus in the simulation of the Echeng station is merely 73 kV, which is considerably lower than the recorded value. The analysis of the distant district line-to-ground fault accident in [9] shows that the neutral bus overvoltage amplitude is no more than 200 kV in a ±500 kV converter station. Note that the recorded value is considerably high, which is uncommon in operational conditions. Further analysis should be performed to clearly understand the cause of the failure and to avoid similar failures in the future. 4.1 Neutral bus overvoltage record analysis To determine the cause of the unusual overvoltage, all records involving the distant district line-to-ground fault accident should be investigated. The overvoltage record is chronologically shown in Fig. 8. Fig. 8Open in figure viewerPowerPoint Neutral bus overvoltage on Pole I in the Echeng station caused by the line-to-ground fault Fig. 8 shows that the overvoltage amplitudes of 16 line-to-ground faults are higher than the surge capacitor's SIWL. Note that all overvoltages at the surge capacitor's SIWL occurred after April 2008. In addition, Fig. 8 shows that three overvoltages were higher than 246 kV but did not cause any failure. The maintenance record also shows that the low voltage arms of the voltage dividers on the neutral bus were repaired because they were damaged during the 'great ice storm' in southern China in January 2008. In the emergency repair process, the LV arms are replaced with spare parts. Note that the normal voltage on the neutral bus is a significantly low DC voltage and the measured value is correct because of the correct resistances of the low arms. However, the capacitances of the low arms are not adjusted by test. Thus, we can infer that the unusual overvoltage is caused by measurement error. Field tests conducted determine the low performance of the voltage divider. 4.2 Voltage divider performance test To determine whether the unusual overvoltage is caused by the voltage divider, a test on the performance of the voltage divider under different voltages (direct voltage and 100 Hz alternating voltage) was conducted. The test diagram is shown in Fig. 9, where R1 and C1 are the resistance and capacitance of the high voltage arm, respectively; and R2 and C2 are the resistance and capacitance of the low voltage arm, respectively. The test results are shown in Table 6. Table 6. Performance of the voltage divider under different voltages Source voltage Standard voltage divider, kV Voltage divider on the neutral bus of the Echeng station, kV direct current voltage 8.96 8.77 9.99 9.95 alternating current voltage (100 Hz) 5.98 24.8 6.98 28.6 Fig. 9Open in figure viewerPowerPoint Diagram of the voltage divider test Table 6 shows that the performance of the voltage divider on the neutral bus of the Echeng station is good under DC voltage, but serious error is introduced when the source is changed to 100 Hz AC voltage. This phenomenon is similar to the unusual overvoltage discussed in this paper. For further analysis, the parameters of the voltage divider is measured and shown in Table 7. Table 7. Voltage divider parameter measurement value Parameter Resistance Capacitance Resistance ratio Capacitive resistance ratio high voltage arm of the voltage divider (Pole I and Pole II are the same) 30 MΩ 300 pF / / low voltage arm of the voltage divider in Pole I 3.76 kΩ 550 nF 7978.7 1833.3 low voltage arm of the voltage divider in Pole II 3.76 kΩ 563 nF 7978.7 1876.6 Table 7 indicates that the time constants (RC) of the high and low voltage arms are not equal, thereby causing the bad ratio–frequency characteristics of the voltage divider (Fig. 10). This phenomenon is considered to be the actual cause of the unusual overvoltage discussed in this paper. Given the test results, the actual overvoltage can be reconstructed to substantially investigate the reasons for such failure. Fig. 10Open in figure viewerPowerPoint Ratio–frequency characteristics of the voltage divider on the neutral bus 4.3 Actual overvoltage reconstruction Actual overvoltage can be reconstructed on the basis of the schematic diagram and the voltage divider parameters. From Fig. 12a and Table 7, the difference equation can be obtained as follows (1)yields (2)where {uk} (k = 2, 3, …, N) is the time series of the high voltage on the primary side of the voltage divider; {yk} (k = 2, 3, …, N) is the voltage time series detected from the secondary side of the voltage divider; N is the total points number of the voltage time series; T is the sampling period of the voltage detection system; and R1, R2 and C1, C2 are the resistance and capacitance parameters, respectively, of the voltage divider. Fig. 12Open in figure viewerPowerPoint Voltage divider circuit and the reconstructed overvoltage of the neutral bus a Schematic diagram of the voltage divider on the neutral bus b Reconstructed overvoltage on the neutral bus of Pole I Using (1), the actual overvoltage can be calculated by (2). The reconstruction method is also verified by simulation as follows. In the simulation analysis, two voltage dividers can be installed on the measurement point. The parameters of one voltage divider are the same as the standard voltage divider, and the parameters of the other are the same as the incorrect voltage divider. With the known voltage divider ratio, the neutral bus voltage waveform can be obtained. Fig. 11a shows the neutral bus voltage waveform obtained by the incorrect voltage divider. The peak value and waveform are quite different from the actual ones presented in Fig. 7b. However, the true voltage waveform can be reconstructed and presented in Fig. 11b by (2). The results show that Fig. 11b is the same as Fig. 7b. Therefore, the actual overvoltage reconstruction method is correct, and the reconstructed voltage waveform is the true reflection of the system condition. Fig. 11Open in figure viewerPowerPoint Waveforms before and after reconstruction acquired by the incorrect voltage divider in simulation a Waveform before reconstruction b Waveform after reconstruction From (2), the reconstructed waveform shown in Fig. 6b is presented in Fig. 12b. Note that the amplitude of the actual overvoltage is below the arrester's SIPL and is substantially below the surge capacitor's SIWL. The record and test also show that the responsible MOA are qualified, and the overvoltage shown in Fig. 12b can explain the failure. Therefore, overvoltage is not the actual reason that causes the surge capacitor failure. From the preceding analysis, the capacitor insulation should be investigated, which may be the actual reason for the failure. The true data shown in Fig. 8 are not presented in this paper because the detailed data acquired from the fault recorder system was unavailable as of the paper's submission date. However, the true data are considerably easy to reconstruct using (2) as long as the detailed fault data are obtained. 4.4 Insulation cumulative damage The neutral bus overvoltage is considerably lower than the surge capacitor's SIWL; hence, the burning of the capacitor is quite uncommon. On the basis of the existing test phenomenon and our previous study results, we infer that the surge capacitor insulation may be damaged because the capacitor has been used for over 10 years and has repeatedly suffered voltage surges, the amplitude of which is lower than its insulation level. The capacitor core comprises several elements (its inner cross-section structure is presented in Fig. 13), which are connected in series or parallel, and insulation components. A capacitor element constantly consists of two aluminium foil electrodes separated by dielectric polypropylene film, and all the components are impregnated with impregnation liquid, such as benzyl toluene. In operational conditions, the capacitor insulation system suffers aging effect from various factors, such as chemical, thermal, and electrical aging, which may decrease the system's electrical performance [16]. From our previous study of oil paper [17], insulation failure could occur to the oil paper insulation when subjected to repeated impulses (note that the amplitude of the applied impulse voltage is considerably lower than the breakdown voltage of oil paper insulation). This accumulative breakdown phenomenon is known as accumulative effect. Moreover, the surface morphology of oil paper insulation changes substantially under the accumulation of repeated impulses: the surface roughness of oil paper insulation increases with the increase in the applied number of repeated impulses. Analysis results indicate that the accumulation of repeated impulses may cause the bond breakage of cellulose molecules, which is considered a possible reason for the accumulative breakdown of oil paper insulation. Fig. 13Open in figure viewerPowerPoint Inner cross-section structure of the capacitor element To prove whether the surge capacitor failure is caused by the accumulative effect, a similar experiment is conducted as follows. Sample preparation:Polypropylene film (with thickness of 12 μm and DC breakdown strength of 420 V/μm) and benzyl toluene impregnation fluid were generously provided by Guilin Power Capacitor Co., Ltd. All the polypropylene films were cut into disc shape of 80 mm in diameter. Pre-treatments were performed on all experimental samples. First, the polypropylene films and benzyl toluene were separately dried at 50°C for 24 h in a vacuum drying oven. Second, all polypropylene film samples are impregnated in a vacuum at 50°C for another 24 h using the benzyl toluene that had been purified and vacuum degassed in a two-stage drying unit to ensure the lowest gas and water content. After the pre-treatments, all samples were sealed in a glassware stored in a cool, dry, isolated, and well-ventilated storage area. All samples used in this test were thought to be uniform because they were similar in size, and similar pre-treatments were performed. Impulse accumulation test platform: Negative standard switching impulse voltage (250/2500 μs) was selected in this test as an example. A 30 kJ/400 kV consecutive impulse generator was used to generate the negative standard switching impulse; voltage signals measured by a high voltage divider were recorded. On the basis of the primary capacitor element structure, two plate aluminium electrodes are used to investigate the basic characteristics of capacitor insulation subjected to repeated switching impulses because the plate electrodes can produce a quasi-uniform electrical field that simulates the electrical field in surge capacitors. The diameter and edge radius of the electrodes are 25 and 3 mm, respectively. Given that polypropylene film is typically used as multilayer insulation in capacitors, two layers of the polypropylene samples were placed between the electrodes and fixed by an epoxy insulation support to avoid sloshing. The interval time for the two consecutive shots in this study was 120 s to avoid the influence of space charge. Accumulative effect test on capacitor element samples: From the preceding study, capacitor insulation failure may be caused by repeated impulses even though the amplitude of the applied impulse was considerably lower than the breakdown voltage of the capacitor insulation. This effect of repeated impulses on the insulation strength can be analysed using the V–N characteristics (breakdown voltage number of voltage applications characteristics) [18]. Fig. 14 shows that the 50% negative switching impulse voltage of the sample is 13.3 kV, and also indicates that numerous impulses below the SIWL magnitude of the capacitor element samples can result in insulation system failure. This result can be attributed to the accumulative effect of repeated switching impulses. Evidently, the applied times of impulses until the breakdown of the capacitor insulation samples decreased rapidly with the increase in test voltage, thereby exhibiting an exponential relationship. The amplitude of the applied impulses considerably influences the number of applied impulse times until the breakdown of the capacitor insulation samples. Although the experiment is exploratory in nature, the idea that insulation system failure can be caused by repeated impulses with lower amplitude than the withstand voltage of the insulation system is convincing. Fig. 14Open in figure viewerPowerPoint V–N characteristics of the capacitor element sample under standard switching impulses The preceding failure analysis indicates that the main reason that caused the capacitor failure discussed in this paper is capacitor insulation damage rather than overvoltage. Meanwhile, the bad performance of the voltage divider on the neutral bus of the Echeng station is also a serious problem, which may confuse the field operator regarding the unusually high overvoltage and cause a misunderstanding of the failure. 5 Maintenance and discussion Two maintenance measures are proposed and applied on the basis of the current study. The capacitor on the low voltage arm is replaced and verified to make the ratio of the voltage divider stable with the change of overvoltage waveform. Experiment shows that the most possible reason of this failure is the insulation cumulative damage in the surge capacitor during long-time running. Multiple impulse voltage, even lower than insulation levels, makes the capacitor inner insulation cracking, thereby leading to the neutral bus voltage exceeding the actual withstand voltage of the capacitor and causing the capacitor to explode eventually. For this reason, all of the surge capacitors in the same batch on the neutral bus have been changed entirely. With the development of the HVDC transmission system, several converter stations have been running for several years. Thus, studying the aging and accumulative effects of capacitors is increasingly necessary because the HVDC station (particularly in the filter field) is equipped with numerous capacitors. Meanwhile, an online monitoring system for capacitor insulation state, such as dielectric loss and leakage current, should be developed and installed in the converter station to monitor the capacitor state. With the expansion and deepening of the research, efficient strategies will be proposed to avoid future failures from cumulative damage from repeated surges. 6 Conclusions This study analyses the failure of a surge capacitor on the neutral bus in a ±500 kV HVDC converter station from the point view of overvoltage and insulation through simulation and practical record information analysis. Combining the neutral bus overvoltage record and voltage divider performance test and simulation, the result obtained indicates that the unusually high overvoltage, the amplitude of which is 246 kV, on the neutral bus was caused by the wrong capacitance of the low voltage arm of the voltage divider on the neutral bus. From the field test and measured parameters, the actual overvoltage on the neutral bus can be calculated, the amplitude of which is 75 kV and is considerably lower than the surge capacitor's SIWL (95 kV). Therefore, overvoltage is not the actual cause of the failure. From the experimental results, the main cause of the capacitor failure is the capacitor insulation damage rather than the overvoltage. On the basis of this study's results, maintenance measures are proposed and applied in this converter station. Further study of the aging and accumulative effects of the capacitor insulation should be conducted and an online monitoring system for capacitor insulation state should be developed and installed in the converter station. 7 Acknowledgments This study was supported by the National Basic Research Program of China (973 program) (grant no. 2015CB251003), Fundamental Research Funds for the Central Universities (grant no. 106112015CDJRC151212), Key Research Project of the State Key Laboratory of Power Transmission Equipment & System Security and New Technology (Chongqing University) (2007DA10512714104), and Funds for Innovative Research Groups of China (grant no. 51321063). Thanks are due to Guilin Power Capacitor Co., Ltd. for their generous support for the experiment on Insulation Cumulative Damage. 8 References 1Taggart, S., James, G., Dong, Z.Y., et al: 'The future of renewables linked by a transnational Asian grid', Proc. IEEE, 2012, 100, (2), pp. 348– 359 (doi: https://doi.org/10.1109/JPROC.2011.2159089) 2Larruskain, D.M., Zamora, I., Abarrategui, O., et al: 'VSC-HVDC configurations for converting AC distribution lines into DC lines', Int. J. Electr. Power Energy Syst., 2014, 54, pp. 589– 597 (doi: https://doi.org/10.1016/j.ijepes.2013.08.005) 3Hammons, T.J., Woodford, D., Loughtan, J., et al: 'Role of HVDC transmission in future energy development', IEEE Power Eng. Rev., 2000, 20, (2), pp. 10– 25 (doi: https://doi.org/10.1109/MPER.2000.819913) 4He, Z.Y., Liao, K., Li, X.P., et al: 'Natural frequency-based line fault location in HVDC lines', IEEE Trans. Power Deliv., 2014, 29, (2), pp. 851– 859 (doi: https://doi.org/10.1109/TPWRD.2013.2269769) 5Yong, L., Zhang, Z.W., Rehtanz, C., et al: 'Study on steady- and transient-state characteristics of a new HVDC transmission system based on an inductive filtering method', IEEE Trans. Power Deliv., 2011, 26, (7), pp. 1976– 1986 (doi: https://doi.org/10.1109/TPEL.2010.2098888) 6Zhang, X., Yu, Z.Q., He, J.L., et al : 'Mechanism of +/− 800 kV HVDC converter abnormal block fault caused by lightning transient', Electr. Power Syst. Res., 2014, 113, pp. 157– 164 (doi: https://doi.org/10.1016/j.epsr.2014.03.023) 7Chen, X.L., Zhou, H., Wang, D.J., et al: 'Study on transient overvoltage of Zhexi converter station of ±800 kV DC power transmission project', Power Syst. Technol., 2012, 36, (3), pp. 22– 27 8Hang, Y.X., Li, L.C., Chen, H.X., et al: 'Influence of modeling methods on the calculated lightning surge overvoltages at a UHVDC converter station due to back flashover', IEEE Trans. Power Deliv., 2012, 27, (3), pp. 1090– 1095 9Jing, Y., Ren, Z., Yang, J.B., et al: 'Study on overvoltage of Tian–Guang HVDC transmission', High Volt. Technol., 2002, 28, (4), pp. 1– 3 10Lu, W.X., Ooi, B.T.: 'DC overvoltage control during loss of converter in multiterminal voltage-source converter-based HVDC (M-VSC-HVDC)', IEEE Trans. Power Deliv., 2003, 18, (3), pp. 915– 920 11Wang, D.J., Zhou, H., Deng, X.: 'Switching overvoltage characteristics of ±1100 kV UHVDC converter station', IEEE Trans. Power Deliv., 2015, 30, (3), pp. 1205– 1212 (doi: https://doi.org/10.1109/TPWRD.2014.2368598) 12Hingorani, N.G.: 'Transient overvoltage on a bipolar HVDC overhead line caused by DC line faults', IEEE Trans. Power Appar. Syst., 1970, PA89, (4), pp. 592– 610 (doi: https://doi.org/10.1109/TPAS.1970.292606) 13Zhong, J.Y., Guo, Y.J., Zhang, H., et al: ' Simulation research on DC field over-voltage caused by the open-circuit fault in neutral busbar and metallic return'. 2014 Int. Conf. on Information Science, Electronics and Electrical Engineering (ISEEE), 2014, 1–3, pp. 735– 739 14Faruque, M.O., Zhang, Y.Y., Dinavahi, V.: 'Detailed modeling of CIGRE HVDC benchmark system using PSCAD/EMTDC and PSB/SIMULINK', IEEE Trans. Power Deliv., 2006, 21, (1), pp. 378– 387 (doi: https://doi.org/10.1109/TPWRD.2005.852376) 15Kim, C.K., Jang, G.: 'Development of Jeju–Haenam HVDC system model for dynamic performance study', Int. J. Electr. Power Energy Syst., 2006, 28, (8), pp. 570– 580 (doi: https://doi.org/10.1016/j.ijepes.2006.03.015) 16Zhu, M.X., Li, Y., Wei, Y.H., et al: 'Influence of thermal degradation on surface discharge characteristics for oil-polypropylene insulation', IEEE Trans. Dielectr. Electr. Insul., 2015, 22, (5), pp. 2661– 2668 17Sima, W.X., Sun, P.T., Yang, Q., et al: 'Study on the accumulative effect of repeated lightning impulses on insulation characteristics of transformer oil impregnated paper', IEEE Trans. Dielectr. Electr. Insul., 2014, 21, (4), pp. 1933– 1941 (doi: https://doi.org/10.1109/TDEI.2014.004298) 18Sima, W.X., Sun, P.T., Yang, M., et al: 'Effect of space charge on the accumulative characteristics of oil paper insulation under repeated lightning impulses', IEEE Trans. Dielectr. Electr. Insul., 2015, 22, (5), pp. 2483– 2490 Citing Literature Volume10, Issue6July 2016Pages 852-860 FiguresReferencesRelatedInformation
Referência(s)