Motor Fault Diagnosis Based On Data Fusion

This paper is mainly about motor fault diagnostic method and discusses the characteristics of induction motor fault diagnosis and its achievements, which is based on a large collection of the abroad and domestic scientific and technical documents. Then it gives a new diagnostic method based on BP neural net and D-S inference, and through experiments it is verified.In the process of research, data fusion technology are firstly introduced,which includes principle, model, structure and algorithm of data fusion. Then against electrical fault and mechanical fault of motor fault, the fault mechanism of induction motor are analyzed by stator current detection method, internal relations between motor faults and typical frequencies is presented.Secondly, we discuss the characteristic of BP neural network, and it shows that neural network can be used for motor fault diagnosis. We study motor fault diagnosis means based BP neural network, and put forward improvement algorithm integrating alter-learn rate and appending momentum, which quicken the convergence speed. To the uncertainty of fault diagnosis, the decision-making fusion based on D-S evidence theory is put forward. The basic conception of evidence theory is introduced, and combined with the example it is analyzed.Finally, electrical fault diagnosis system is brought forward, experimental detection system of induction motor is introduced, and the noise and vibration signal of experimental motor is measured. The measured data is local diagnosed by improved BP neural network, and it is together with by D-S inference, and the satisfaction experimental results are got. The accuracy of motor fault diagnosis is improved that in order to prove the effectiveness this diagnostic method based on BP neural network and D-S inference are verified.