造成風機噪音、磨耗、故障等損壞的主要原因通常是由於過度的振動所造成。本計畫以150kW風力發電機為標的，於風機上關鍵部位裝置振動訊號量測規(加速規)，在風機運轉情況下長時間擷取信號，加以分析整理以做為診斷故障之依據。本計畫處理訊號所依據之基礎理論為傅立葉轉換與小波轉換理論。利用振動分析軟體(如Matlab/Simulink)，進行訊號分析處理。加速規所擷取之原始訊號經過過濾、轉換、修正後之新的振動頻譜與風機正常運作時的振動頻譜進行比較，故障或損毀之零件其振動幅度會有所改變，而產生不同的訊號，依據這些異常訊號，觀測者將可以判別風機上哪些零件故障並需要維修更換，有效做到風機的維護。使用風力發電機振動量測與分析技術，進行風機之故障分析診斷，可以減少排除故障時間，防止多發性故障發生次數，減少停機時間，提高設備完好率和可利用率。由於風機位在高處，佈設線路可能有所困難，因此訊號的傳輸會需要使用無線傳輸，本計畫建議未來可考慮裝設無線傳輸的量測儀器於風力發電機上，運用無線網路監控風機的運轉狀況，建構一更易使用的觀測平台。 The main cause of wind turbine failure, noise, wear and damage is excessive vibration. This project focuses on the signal analysis of the 150kW wind turbine power generator. Vibration sensors (accelerometers) will be placed on key points of the wind turbine. The signals derived from the sensors will be kept under constant surveillance, and then processed. The processed signals will then be compared with the signals of the wind turbine running under optimized or normal conditions. An increase in a particular frequency usually indicates a failing component. The higher the amplitude of that frequency, the greater the damage at that location. By comparing the signals, we can detect malfunction and damaged parts in the wind turbine and repair it more efficiently, thus lowing the maintenance costs. The vibration analysis software we will use in this project is Matlab/Simulink. It uses Fourier transforms and Wavelet transforms to filter and process the signals derived from the accelerometers. Due to the height of the wind towers, wiring might be a problem. Thus using wireless transmission to transmit the signals from the wind turbines is an acceptable method. A more user friendly monitoring platform can be made through the execution of the project. 研究期間：10001 ~ 10012
