Research On Electromotor Fault Noise Diagnosis System

With science and technology continually developing, improvement of electrization, electrical equipment is used in industry and life more and more popular. It is easy to be failure when electromotor is working. In case electromotor go wrong, it influences normal work of whole system, indeed endangers people's life, and takes great expense or tragedy. So, it is important to diagnose electromotor fault, it ensures producing safe, avoid people and wealth losing, if electromotor is diagnosed in time and right. Voice signal can react to electromotor running state and variety ruler rightly. Because when electromotor is in fault, it produces different voice with normal electromotor, it is feasible to diagnose electromotor by researching the voice of electromotor. The object of electromotor noise fault diagnosis research is electromotor voice.The paper use blind source separate technology (BBS),Wavelet transform and probability theory in the electromotor fault noise diagnosis system. When electromotor is in fault, it is probable that some breakdown happen at the same time, and electromotor voice signal populate collected by sensor must be populated by other voice signal spread in air. So, sensor collect mix signal. The premise of diagnosing fault rightly is separating mix signal collected by sensor to single signal. We use one algorithm of BBS technology can separate mix signal successfully.The fault signal is non-stationary, transient one, FT just adapt to analyze stationary signal, and it is an analytic method used in frequency field but not in time field. Wavelet transform have excellent local character both in frequency and time field, it is suitable for check normal signal that contain abnormal appearance and show its composition. So we use Wavelet technology to analyze single signal separated from mix signal.We can judge what faults are in electromotor by compare the character of signal and faults data base. It is a huge data statistics experience to build fault data base. The paper use parameter estimation and combine samples of known fault to build data base. At last, give an example of electromotor fault diagnosis system to prove the system is effective and practical.