Mine ventilator fault diagnosis based on information fusion technique

Abstract

Abstract A fault diagnosis method of multi-fault-featured information fusion is proposed to improve accuracy of fault diagnosis. The multi information of this method includes stator current signal, axial vibration signal, and radial vibration signal. These collected signals are processed by wavelet analysis to extract the fault feather. Based on each type of information, primary conclusion is achieved by neural networks. In order to achieve the finally conclusion, Dempster combination rule is used to realize information fusion. The experiment result shows that the reliability of fault diagnosis with the multi-fault characteristic information fusion is improved evidently and its uncertainty decreases remarkably. It proves that the proposed method can improve the accuracy and reliability of fault diagnosis.