電能擷取優化之風力發電機即時偏航動態模擬分析

Abstract

風力發電的經濟效益取決於發電效率，而風機發電量是與風速立方成正比，最直接影響到風機接收風能的多寡即取決於其迎風轉向偏航系統追風的性能。因此，本研究針對兆瓦級風力機做為研究載具，並由長期監控的台電四湖發電廠風力發電系統資料庫，進行短期風速與風向預測，設計風機電能擷取優化迎風轉向偏航系統。此系統以類神經網路預測短期風速/風向做為初步基準，使迎風轉向偏航控制系統能對預測之風速/風向資料進行運算，藉由在風機擷取電能及偏航轉動耗能兩者間進行爬山控制(HCC)轉向偏航決策，研擬最佳的迎風轉向偏航機制，進而控制風機偏航角度使其維持在最佳淨獲能，並同時以爬山演算法控制風機轉子輸出使風機系統維持在其最佳風能利用係數。最後為驗證本研究建構之優化偏航系統能穩定的使風機維持在其最佳迎風轉向偏航狀況，以Labview人機介面與風機發電資料庫實際模擬風機運作的即時情況，動態模擬驗證此迎風轉向偏航控制的功率優化效果及其穩定性。

The economic benefits of wind turbine system depend on efficiency of power generation. The generating capacity is proportional to the cube of wind speed, and the amount of wind energy directly impacted on the wind turbine is controlled by yaw system. This study based on megawatt wind turbine operation to design a wind turbine yaw control system to optimize energy capture. A neural network process predicted short-term wind speed/direction as a preliminary basis, and the yaw control system calculated energy captured by the blade and the power consumption of yaw rotation. The optimization of yaw angle control mechanism and a simultaneously mountain climbing algorithm to control the blade rotor system are applied to maintain the wind energy utilization at its best factor. A Labview HMI operation of real-time simulation of wind turbine dynamic verifies the effectiveness of yaw control and its stability of power optimization. In additions, the prediction methods based on feed-forward neural network from the Taipower’s meteorological data of wind turbine is described in this work.