Insulators ESDD Predicting Based on Wavelet Neural Network

Wavelet neural network (WNN) is a feed-forward neural network which is based on wavelet transform. The network overcomes the intrinsic shortcomings of artificial neural network, namely, slow learning speed, difficulty to determine rationally the network structure and existence of partial minimum points. Hence, WNN has more freedom degree and better adaptability than traditional multi-layer feed-forward neural network. In the interest of better reflection of the influence of meteorological factors on insulators equal salt density (ESDD) and increase of the accuracy of ESDD prediction ,the paper uses Morlet wavelet to construct WNN , adopts error backpropagation algorithm to train the network and applies the ESDD data and meteorological data of Qingshan District ,Wuhan, which were measured from April to June in 2005, and the same times in 2006 respectively, to model and forecast ESDD. The predicted results are very close to the measured ones which show the WNN model can effectively improve the speed and accuracy of the forecasting. Therefore, the model presented provides a doable thought for the computerization of pollution area map of power network.