Research On Fan Forecasting Based On Wavelet And Support Vector Machine

Fan is widely used power equipments in modern industry and mining industry and plays an important role in the modern construction of economy .Once the facilities are breakdown, huge loss is caused for the enterprises. By applying methods of forepart prediction of fault diagnosis, fans can be stopped and examined before great faults occure to prevent sudden accidents and keep the normal order of production and reduce overplus servicing and cut down the expense of servicing. Thus the research has important application value in practice.Because common predicting methods have limitations of complicated modeling, narrow applicability, low forecasting precision and bad generalization performance, a forecasting method for fans faults based on wavelet and support vector machine was built in this paper. Original time series was firstly decomposed into many sequences according to scale by wavelet, and then every sequence was forecasted by support vector machine, the final forecasting results of the original time series can be composed by the reconstruction algorithm of wavelet. A practical application in fans of an aluminium plant was also given, results show the present methods is feasible, it can reach a better precision compared with other methods. By applying the methods, enterprise can ensure a smooth process of manufacturing...