Abstract
Windmills are usually installed at an area with strong wind, e. g., large one is installed at coasts, and small ones are installed on rooftop of buildings.Then even a little wrong point on a blade of windmill is possible to be connected with a debacle.Thus, a summary diagnositic system for windmills should be developed as soon as possible in order to use the windmill system safely.In this paper we considered a diagnostic system form vibration signal for a small-scaled windmill's condition. We proposed a method of feature extraction based on Discrete Wavelet Transform (DWT), autocorrelation of the DWT's coefficients and enveloping process for autocorrelation coefficients. Then we developed automatic diagnositic system by a three-layered neural network.Through some experiments we indicated that the diagnostic accuracy of proposed system is satisfactory high.
