Experimental Verification of Advanced Voltage Control for Penetration of PV in Distribution System with IT Sectionalizing Switches

Transcription

1 21, rue d Artois, F PARIS C6 _113 _2012 CIGRE 2012 http : // Experimental Verification of Advanced Voltage Control for Penetration of PV in Distribution System with IT Sectionalizing Switches Y. HAYASHI N. TAKAHASHI Y. HANAI Waseda University Central Research Institute of Electric Power Industry E. KAMIYA T. WATANABE H. ISHII Tokyo Electric Power Co. Japan SUMMARY In Japan, penetration of photovoltaic generation (PV) systems has been strongly promoted and the accumulated amount of the installation by 2009 is approximately 2.6 GW. Further, the national target of PV installation capacity is 28 GW in 2020 and 53 GW in 2030, respectively. Although most of PV systems are connected into low voltage distribution networks, it brings power flow with greater uncertainty and causes some operational and control problems. In particular, the voltage control problem is one of the foremost concerns and the dominant constraint of the penetration of PV systems into distribution networks. In order to mitigate the impact of future large PV penetration on distribution network voltages, we have proposed a distribution voltage control method that combines a centralized voltage control by tap change transformers (LRT: load ratio control transformer, SVR: step voltage regulator) and a coordinated voltage control by reactive power compensators (STATCOM: static synchronous compensator). The proposed voltage control has been verified by computer simulations. Meanwhile, we have constructed an analog type distribution system simulator for solving various control and operational problems that environmental changes such as an increase of distributed generations cause in 6.6kV distribution system. The simulator named ANSWER (Active Network Simulator With Energy Resources) is a 200V experiment system that is designed by three-phase, three-wire, and nongrounded as well as 6.6kV distribution system in Japan. This paper describes the experimental results on the analog type simulator in order to verify the feasibility and availability of the proposed voltage control. The demonstration experiments on a simple test models were conducted and the voltage control behavior with LRT, SVR, and STATCOM was evaluated. As the experimental results, it was verified that the proposed voltage control is effective and implementable to maintain line voltages in distribution networks with high penetration of PV systems. Moreover, it was confirmed that the coordinated voltage control with STATCOM have effects of reducing the voltage fluctuations and avoiding frequent tap changes of LRT and SVR. KEYWORDS Distribution network - Voltage control - Photovoltaic generation - Analogue type simulator - IT sectionalizing switch hayashi@waseda.jp

2 1. Introduction Penetration of photovoltaic generation (PV) systems has been strongly promoted in order to reduce CO2 emissions in Japan. The governmental targeted values of PV installation are 28 GW in 2020 and 53 GW in 2030, respectively. Since most of PV systems will be connected into low voltage distribution networks, there is a possibility that the reversed power flow from PV systems causes not only voltage rise but also voltage fluctuations in distribution feeders. Conventional autonomous voltage control scheme has limitations of voltage control ability for the penetration of PV systems. On the other hand, in Japan, IT sectionalizing switches, which are remotely-operable sectionalizing switch with voltage and current sensors [1], are beginning to be installed in high voltage distribution networks in order to realize more accurate distribution systems planning, operation and control based on data integration of measured from these switches. The authors have focused on utilization of measured data from IT sectionalizing switches and proposed a centralized and coordinated voltage control method by voltage control devices such as load ratio control transformer (LRT), step voltage regulators (SVR) and static synchronous compensator (STATCOM) for distribution networks with high penetration of PV systems. This paper shows that the validity of the proposed voltage control approach was experimentally verified by using an advanced analog type distribution system simulator. 2. Proposed Centralized and Coordinated Voltage Control for penetration of PV The penetration of PV systems causes voltage fluctuation and wide voltage gap between voltage rise and voltage drop. These influences cause excess of the allowable voltage range in distribution network. In order to resolve the excess of the allowable voltage range based on the voltage gap and fluctuation, the authors propose a centralized and coordinated voltage control with LRT, SVR and STATCOM. The proposed method has two functions; the first function (voltage gap reducing function) is to reduce the voltage gap by centralizing control of LRT and SVR, and the other (voltage smoothing function) is to smooth rapid voltage change by autonomous control of STATCOM. In the proposed voltage control method, the autonomous control of STATCOM and centralized control of LRT/SVR are coordinated through available use of all measured data from IT sectionalizing switches. The outline is shown in Figure 1. The central control system for LRT and SVRs conducts four procedures ((1) data acquisition, (2) state estimation, (3) voltage gap evaluation, and (4) tap Figure 1. Outline of the centralized and coordinated voltage control 1

3 change instruction) in 1-5 minute cycle. The parameters of STATCOM such as PI control gain, smoothing time constant, and dead band are pre-determined and updated from the central control system. Figure 2 shows control logic of centralized voltage control of LRT and SVR. As shown the figure, the tap positions of LRT and SVR are controlled to reduce the voltage gap evaluated by the measured voltage data from IT sectionalizing switches. On the other hand, STATCOM applies an autonomous reactive power control based on its receiving voltage data to take advantage of its highspeed performance. STATCOM compensates both instantaneous voltage exceedance and voltage fluctuations without causing interference with LRT and SVR. Reduction of total number of tap changes of LRT and SVR by compensating voltage fluctuation by STATCOM can expand the mechanical life of them. 3. Advanced Distribution System Simulator An advanced analog type distribution system simulator has been constructed for solving various control and operational problems that environmental changes such as an increase of distributed generations cause in Japanese distribution system. This simulator named ANSWER (Active Network Simulator With Energy Resources) is a 200V experiment system which is designed by three-phase, three-wire, and non-grounded, as well as Japanese 6.6kV distribution system. The appearance of the simulator is shown in Figure 3. The experiment system has distribution system components such as 15 distribution line devices with switches and sensors, tap change transformers (2 LRTs, 2 SVRs), 21 single-phase RLC loads, and 9 three-phase inverters with bidirectional power flow. The distribution systems configuration can be made by freely choosing and assembling suitable component devices for its purpose of research. The transformer's tap position, the switches states (opened/closed), loads and PVs profiles can be freely controlled by external control order from an online control computer system. 4. Experimental Results Figure 2. Control logic of centralized tap control of LRT and SVR using measured data from IT switches In order to verify the validity of the proposed centralized and coordinated voltage control, demonstration experiments have been conducted on two test models shown in Figure 4 by using the ANSWER. 2

4 Figure 3. The appearance of advanced distribution system simulator (ANSWER). (a) Single feeder model 4.1 Behavior of the proposed voltage control (b) Two feeders model Figure 4. Test models assembled in the advanced distribution system simulator. The behavior of the proposed voltage control is examined in following three cases conducted on a single feeder model shown in Figure 4(a). (Case 1) Voltage control by LRT and SVR on a sunny day (Case 2) Voltage control by LRT and SVR on a cloudy day (Case 3) Voltage control by LRT, SVR, and STATCOM on a cloudy day The daily variations of distribution network state are created by the daily power curves of electric loads and PV systems shown in Figure 5. For simplicity, the electric loads have an even distribution in a feeder. The total load capacity is 3.0 kw (equivalent of 2475 kw), and each load device (Load 1 ~ Load 5) behaves as the demand profiles of approximately 800 households. The total output of PV 3

5 (a) Load power (b) PV systems power Figure 5. Daily profiles of electric loads and PV systems. Tap position Line Voltage [kv] Load1 6.7 Load2 6.6 Load3 6.5 Load4 6.4 Load (a) Voltage profiles (a) Voltage profiles LRT 5 4 SVR (b) Tap profiles of LRT and SVR (b) Tap profiles of LRT and SVR Figure 6. Experimental results in Case 1. Figure 7. Experimental results in Case 2. Line Voltage [kv] Tap position Load1 Load2 Load3 Load4 Load5 LRT SVR systems is also 3.0 kw (1.0 kw*3) at a maximum, therefore, the penetration level of PV system in daytime is over 100 % against the total load power in a residential area. The cloudy-day output profile in Fig.5(b) was created by using actual solar radiation data from NEDO's demonstrative PV project in Ota city, Japan [2]; the daily profile has the largest variation in the observing period. The allowable voltage ranges are from 0.95 pu to 1.05 pu. The impedances of distribution line devices are set as the total line length is 2.5 km. The capacity of STATCOM is 0.6 kva (equivalent of 500 kva) corresponding to one sixth of the distribution feeder capacity. Measured line voltages and tap profiles of LRT & SVR once every minute for Case 1, Case 2 and Case 3 are shown in Figure 6, Figure 7 and Figure 8, respectively. Furthermore, number of LRT & SVR tap changes for each experimental case is compared as shown in Table 1. (Case 1) Results of voltage control by LRT/SVR on a sunny day As shown in Figure 6 (a), voltage profiles by LRT/SVR on a sunny day indicate that each node voltage is controlled within the proper voltage range from 0.95 pu to 1.05 pu. Since PV output power gradually increases and decreases for daytime period on a sunny day, the tap positions of LRT and SVR move to the lower position in order to compensate the voltage rise by reversed power flow from PV system, and then they move to the upper position to compensate the voltage drop as shown in Figure 6(b). 4

6 Line Voltage [kv] Tap position (a) Voltage profiles (b) Tap profiles of LRT and SVR Reactive Power [kvar] (c) Output pofile of STATCOM Figure 8. Experimental results in Case 3. Load1 Load2 Load3 Load4 Load5 LRT SVR Q(SVC) Table 1. Comparison of number of LRT and SVR tap changes. Voltage control devices Weather Number of tap changes LRT SVR LRT and SVR Case1 LRT + SVR Sunny Case2 LRT + SVR Cloudy Case3 LRT + SVR + STATCOM Cloudy (Case 2) Results of voltage control by LRT/SVR on a cloudy day Although the line voltages are rapidly fluctuated by unstable reverse power flow from PV systems on a cloudy day, the line voltages are almost maintained within the allowable voltage range by LRT/SVR as shown in Figure 7(a). However, the tap position of LRT/SVR is frequently changed to compensate the sharp fluctuation of the line voltage. Table 1 shows that the number of both LRTs and SVRs tap change on a cloudy day of Case 2 is two times that of both LRTs and SVRs tap change on a sunny day of Case 1. 5

7 Penetration rate of PV systems (Residential area feeder ) Table 2. Voltage control capabilities to increase of PV systems. Proposed voltage control LRT + SVR + STATCOM LRT + SVR Conventional control LRT + SVR 100 % Accepted Not accepted Not accepted 90 % Accepted Accepted Not accepted 80 % Accepted Accepted Not accepted 70 % Accepted Accepted Not accepted 60 % Accepted Accepted Not accepted 50 % Accepted Accepted Accepted (Case 3) Results of the proposed voltage control by LRT/SVR/STATCOM on a cloudy day Compared with the results of Case 2 shown in Figure 7, Figure 8 shows that the line voltages fluctuation by reverse power flow from PV systems on a cloudy day is mitigated by STATCOM with the smoothing function, and each number of tap changes of LRT and SVR is drastically reduced by the reactive power control. The total number of LRTs and SVRs tap changes obtained by the proposed voltage control with LRT/SVR/STATCOM of Case 3 is about a half times less than that of Case 2 as shown in Table Validation of the proposed voltage control In order to conduct more realistic experimental simulations, the validation on two distribution feeders model shown in Figure 4(b) have been carried out. The test model has a large gap of voltage distribution between feeders due to differences of line lengths, electric loads, and PV systems capacities. In the upper feeder (industrial area), the load capacity equivalent of 2556 kva and the line impedance equivalent of 2.5 km (0.5 km * 5 lines) are given. On the other hand, in the lower feeder (residential area), the load capacity and line impedance are equivalent of 1500 kva and 3.5 km (0.7 km * 5 lines), respectively. PV systems are connected to the lower feeder and the output profiles in cloudy day are applied. The capacity of STATCOM is 0.4 kva (approximately equivalent of 300kVA) corresponding to 10% of the feeder capacity. The voltage profiles and tap changes (LRT, SVR) are evaluated while the pentration rate of PV systems is increased by up to 100%; the penetration rate is defined as a rate of peak output power of PV systems to peak load power of a feeder with PV. The experimental results on two feeders model are shown in Table 2. For comparison, the results of conventional voltage control by locally-controlled tap change transformer (LRT, SVR) are also shown. When a 50 % pentration rate of PV systems, the conventional voltage control reaches its voltage limitation. The limitation is due to voltage gap between feeders. Increase of output power of PV systems declines line currents in the residential area. In response to the power flow decline, LRT reduces its secondary voltage to lower level. In the industrial area, line voltages reduced by LRT eventually deviate from the lower voltage limits. As the results, PV systems are forced to suppress thier output power in order to maintain network voltages. However, such voltage violations are detectable and aviodable by acquiring several line voltages from IT sectionalizing switches. As shown in Table 2, the centralized voltage control by LRT and SVR raise the acceptable penetration rate of PV systems to 90%. Moreover, the acceptable penetration rate is up over 100 % when appling the coordinated voltage control using STATCOM. The tap changes of LRT and SVR are significantly reduced; the total number of their tap changes is 23 times (LRT: 16, SVR: 7) in case without STATCOM but only 10 times (LRT: 5, SVR: 5) in case of the coordinated voltage control. The above 6

8 experimental results show that the proposed method can control line voltages without exceeding the allowable voltage range for the wide voltage gap and rapid voltage fluctuation caused by the fluctuated output of PV systems on a cloudy day. 5. Conclusions In order to mitigate the impact of future large PV penetration on distribution network voltages, we have proposed a distribution voltage control method that combines a centralized voltage control by tap change transformers (LRT, SVR) and a coordinated voltage control by reactive power compensators (STATCOM). Meanwhile, we have constructed an analog type distribution system simulator for solving various control and operational problems that environmental changes such as an increase of distributed generations cause to 6.6kV distribution system. The simulator named ANSWER (Active Network Simulator With Energy Resources) is a 200V experiment system that is designed by threephase, three-wire, and non-grounded as well as 6.6kV distribution system in Japan. The demonstration experiments on a simple test models (single feeder model and two feeders model) were conducted and the voltage control behavior with LRT, SVR, and STATCOM was evaluated. The experimental results have indicated that the centralized voltage control yields more reliable voltage management and maximize potential capacity for accepting high penetration of PV systems. The experimental results also show that the coordinated voltage control with STATCOM has significant effects on reducing the voltage fluctuations and avoiding frequent tap changes of LRT and SVR. This research has been performed as a part of Japan Next-generation Transmission and Distribution Systems Optimal Control Technologies Demonstration Project carried out by 28 entities. This project is partly funded by Ministry of Economy, Trade and Industry (METI). BIBLIOGRAPHY [1] T. Kasajima, R. Endo, Y. Wada, Y. Kudo and H. Kanawa The Development of the Advanced Distribution Automation System with Optical Fiber Network of Tokyo Electric Power Co., Inc., IEEE Power Engineering Society General Meeting, (2004-6) [2] 7