Machine fault diagnosis using industrial wireless sensor networks and support vector machine

Abstract

A machine fault diagnosis method using industrial wireless sensor networks (IWSNs) and support vector machine (SVM) is presented in this paper as a potential low-cost and effective solution for device condition monitoring and fault diagnosis. On sensor node SVM is proposed and researched to reduce the data transmission between sensor nodes, decrease node energy consumption and increase the fault diagnosis accuracy. Simulation experiments are carried out to verify the proposed method. The simulation results show SVM has strong ability to learn and recognize the machine fault pattern.