Serial wound starter motor faults diagnosis using artificial neural network

Abstract

This paper presents a fault diagnosis system for a serial wound starter motor based on multilayer feed forward artificial neural network (ANN). Starter motor acts as an internal combustion (IC) engine and has a vital importance for all vehicles. That is because, if the starter motor fault occurred, the vehicle cannot be run. Especially in emergency vehicles (ambulance, fire engine, etc) starter motor faults causes the faults. This ANN based fault detection system has been developed for implementation on the emergency vehicles. Information of starter motor current is acquired and then it is practiced on a neural network fault diagnosis (NNFD) system. The multilayer feed forward neural network structures are used. Feed forward neural network is trained using the back propagation algorithm. NNFD system is effective in detection of six types of starter motor faults. NNFD system is able to diagnose the faults that can be seen in most frequencies in starter motors.