Research On Motor Fault Diagnosis By Estimating Of Noise Sources

With the greatly improved automation technology and computing capacity, modern technology allows that motor fault could get a more accurate diagnosis.Temperature monitoring, vibration monitoring and current monitoring are the most important parameters from diverse monitoring when motors are running. However, it is unrealistic to use a large and detection system on one motor and it is also impossible to put all detecting projects for one device. Monitoring programs selection should be taking into account the cost of motor monitoring, specific circumstances and requirements.Since the sound absorption rate is low, the noise from motor vibration will be external radiate unreservedly when the device is running. This paper mainly studied the type of noise in motor fault and analysed the way of motor noise production. The noise data has motor structure informations and the running status information, so we can use this characteristic to research the fault on motor in-depth. This paper proposes to monitor the sound when the fault motor is running, and this diagnosis method have two advantage point of non-contact way and collection time shortly which the record in a limited time are sufficed the analytical requirements.Completed using a microphone array sensor to collect the motor noise, the sound datas processing has been departed three ways to diagnose the fault motor which contain the frequency of noise, the numbers of signal source and the direction of noise. The collected sound signal will be amplified and digital filtering from Matlab software. Used the gerschgorin algorithm to determine the number of sources and analysed by the multiple signal classification algorithm (MUSIC algorithm), the motor fault has been diagnosis correctly finally.In this paper, the parameters of signal processing algorithms including the noise source digital filtering, gerschgorin algorithm and MUSIC algorithm have been one by one simulated. All values have been calculated to satisfy the algorithm requirement. Finally, we diagnose a fault motor successfully and effectively through on-line Finally, we diagnose a fault motor successfully and effectively through on-line monitoring the motor and the noise sound data analysis.