離岸風場併入輸電系統之研究

A Study of Offshore Wind Farms Connected to Transmission Systems

楊任軒(Ren-HsuanYang)

指導教授 ： 李清吟

國立臺北科技大學/電資學院/電機工程系所/碩士(2009年)

摘要

台灣因天然資源短缺，故98%皆仰賴進口，加上台灣西部沿海與離島擁有相當豐富的風能資源，且為提高再生能源於我國的發電占比，台電公司計畫於西元2010年與2015年分別於彰濱、澎湖外海建構大型離岸風場。由於大規模裝置容量的風場併入台電系統，勢必將造成併接點與附近匯流排一定程度的衝擊影響，以及當系統發生不同位置、類型之故障時，風場的響應情況也將成為關心的課題。故本研究為探討大型風場併入後系統的穩態潮流、風場提供之故障電流、系統發生故障時的風場暫態響應情況與不同電壓層級之故障點造成風場切離與否之模擬等。
本研究採用電力系統模擬軟體PSS/E，以西元2010與2015年台電完整系統為研究對象，差異在於2010年108MW風場併接點為彰濱超高壓轄區的線西配電所，而2015年200MW風場併接位置為澎湖一次變電所，併接之電壓層級皆為161kV，但因澎湖由海底電纜與台灣聯結，故其成為系統之末端迴路，加上海纜過高的充電電容效應，故必須於兩端裝設電抗器降低端電壓上升的情況，對此風場併接對澎湖的影響勢必更加顯著，故文中也將提出改善方式。
模擬使用的風機為GE Wind公司所生產之GE 1500，此機型為配置雙饋感應發電機(DFIG)式風機。藉由對DFIG式風機不同的虛功率控制設定可知，西元2010年彰濱轄區系統受108MW風場於不同控制模式下併入之影響甚微；而2015年澎湖轄區系統受200MW風場不同控制模式下併入之影響明顯，需經由調整海纜兩側之電抗器方能使不同控制模式之風場併入且系統電壓符合法規。之所以兩系統對風場併入有如此差異，因併接系統強健度不同所致，但上述兩併接點之匯流排斷路器皆不會因風場併入而導致啟斷功能失效。另外，此兩個風場若皆操作於吸收虛功率的控制模式如電壓控制0.95pu與0.9功因超前時，風場輸出之實功率受到三相短路故障之影響會較其餘四種控制方式明顯。108MW風場於2010年系統尖峰時可忍受之345kV匯流排之三相短路故障範圍會較離峰時更近，但面對161kV匯流排故障則無系統尖離峰的差別。而2015年不論故障發生於345kV或是161kV匯流排，200MW風場可忍受之故障點皆為尖峰較離峰增加一個故障點。
由模擬結果可知，2010年108MW風場併入彰濱轄區系統時，若要顧及系統故障時風場的電力品質，則風場並不適合採用吸收虛功率的控制模式。而2015年200MW風場併入澎湖轄區系統則因海纜的關係，若風場須以不同控制方式併入，則必須適時的調整海纜兩側投入之補償器。

關鍵字

再生能源 ； 風力發電 ； 離岸風場 ； 雙饋式感應發電機

並列摘要

The 98% of resources are imported into Taiwan, because natural resources are in shortage. In addition, Off-shore islands and coast of western Taiwan have abundant wind power energy and for increasing renewable resource penetration of Taiwan, the TPC plan to build large wind farms in Chang-Pin and Peng-Hu in 2010 and 2015, respectively. There will be a serious impacts on the PCC of wind farms and buses nearby when large capacity wind farms interconnect to TPC power systems, also the issue of wind farm’s response behavior is concerned when faults occur in different locations and different types of TPC power system. So, the study investigates the simulation of power flow of the large wind farm which is connecting to the system, the fault current which is supplied from wind farm, transient response behaviors of wind farm when faults occur and the wind farm disconnect caused by fault occur from different voltage level.
The simulation program of this study uses power system simulation program PSS/E, the investigated systems are entire TPC power system in 2010 and 2015. The 108MW wind farm is connected to Sheng-Ci substation which is under Chang-Pin extra high voltage zone in 2010, and the 200MW wind farm is connected to Peng-Hu primary substation in 2015, both of the two wind farms are connected to 161kV side of substations. Peng-Hu will be connected to Taiwan by submarine cable in 2015, therefore Peng-Hu will be the terminal of TPC power system. The characteristic of submarine cable is over high reservoir capacitor that will cause high voltage in terminal of submarine cable. Banks of reactor should be installed on both sides of terminal of submarine cable to decrease the voltage increase. For this situation, the impact to Peng-Hu will be more serious when wind farm is connected. The improvement method to these cases are also described in this article.
The wind turbine simulated in this article is GE 1500 produced by GE Wind, and DFIG is installed in this machine. Known by different sets of reactive power of DFIG, the impact of Chang-Pin system is slight when 108MW wind farm connected Based on different control modes of reactive power of GE 1500 in 2010, but it’s opposite for Peng-Hu with 200MW wind farm in 2015. For the Peng-Hu cases, banks of reactor installed on both sides of submarine cable should be adjusted to make different reactive power control modes connected to Peng-Hu successfully, and maintain in range of TPC rules. The difference between the two cases are because the robustness of each system. Besides, the fault current provided by each wind farm won’t let breaker of PCC lose function. Two wind farms mentioned above that operate in control mode of absorbing reactive power such as voltage control 0.95pu and power factor control 0.9 leading, output active power of wind turbine is disturbed more obviously than the other four control modes during the three phase short circuit fault. The 108MW wind farm in 2010’s system with peak load under the three phase short circuit fault of 345kV bus range has a better low voltage ride through performance than light load,but there are no differences between system’s peak and light for the three phase short circuit faut of 161kV bus. However, no matter the fault occurred of 345kV or 161kV bus in 2015, the 200MW wind farm can endure more one point of fault in peak load system than in light load.
From the final results, if it needs to consider the power quality of wind farms when system faults, then it’s not suit for 108MW wind farm connected to Chang-Pin system in 2010 when using absorbed reactive power control mode. However, because the submarine cable connected between Taiwan and Peng-Hu in 2015, the compensator on the both side of submarine cable should be timely adjusted for different reactive power control mode of wind farm.

並列關鍵字

Renewable Energy ； Wind Power Generation ； Offshore Wind Farm ； Doubly Fed Induction Generator(DFIG)

參考文獻

[8]Ana Estanqueiro, “IEA Wind Energy Annual Report 2007,” International Energy Agency, July 2008.

[13]Fei Ye, Xueliang Huang, Chaoming Wang, Gan Zhou and Ping Luo, “The Impact and Simulation on Large Wind Farm Connected to Power System,” IEEE Third International Conference on Electric Utility Deregulation and Restructuring and Power Technologies, April, 2008, pp. 2608-2614.

[14]S. Towito, M. Berman, G. Yehuda, R. Rabinovici, “Distribution Generation Case Study: Electric Wind Farm with Doubly Fed Induction Generators,” 2006 IEEE 24th Convention of Electrical and Electronics Engineers in Israel, Nov. 2006, pp. 393-397.

[15]Yuriy Kazachkov, Ronald Voelzke, “Modeling Wind Farms for Power System Load Flow and Stability Studies,” 2005 IEEE Russia Power Tech, June, 2005, pp. 1-8.

[17]Jean Momeau, Chad Abbey, and Geza Joos, “Effect of Low Voltage Ride Through Technologies on Wind Farm,” IEEE Canada Electrical Power Conference, 2007, OT.L. 2007, pp. 56-61.