風力發電在全球的發電市場中,是一個成長相當快速的能源,然而風能具有間歇性特性,對於供電穩定度及電力品質影響甚巨,穩定的控制風機輸出功率是一件極為重要的課題,本研究包含風力發電機組、儲能設備與電力電子元件之微電網併聯市電運轉分析,以提高系統穩定度與可靠度。 本文探討風速變動對於永磁同步風力機輸出功率對應關係,並利用類神經網路進而預測下一時間點風速,來推測儲能設備注入或吸收功率進行改善,利用線上學習法則來修正類神經網路輸出誤差,並在Xilinx System Generator for DSP模擬環境中實現類神經網路架構。最後使用Matlab/Simulink與OPAL-RT軟體進行平行即時運算模擬,完成儲能設備調節風力機之變流器輸出端,能以固定功率輸出與平衡負載需求之情境模擬。
In the power markets, the wind power is developed very fast in the world. However, the wind energy has an intermittent characteristic resulting in severe problems of stability and power quality. Therefore, the controlled power output of a wind turbine is an important issue. In this thesis, both of the wind turbine and energy storage are controlled via the power converters in a microgrid connected to the utility grid in order to ensure stability and power quality. In this thesis, the relation between wind speed change and output power of a wind-turbine PMSG is investigated. The artificial neural network is used to forecast the next wind speed in order to speculate the energy injection or absorption of an energy storage. The on-line learning rule, which is realized with Xilinx System Generator for DSP, is implemented in the neural network for improving the learning capability and minimizing the errors. Finally, both constant output and load balance modes of the inverter for the wind-turbine PMSG are studied using Matlab/Simulink and OPAL-RT.