Study And Develop On Engineer Aid-analysis System In Rotate Machine Fault Diagnosis

Following the rapid development of the modern industry, which request more and more the equipments safety-run, the technology of fault diagnosis get the extensive value. However the traditional diagnosis can not meet the industrial production's need. Along with the development of computer, the technology of fault diagnosis which combined the technology of expert system and neural network science entered the intelligent stage.The state of equipments monitor technique for large-scale rotate machine was introduced in this paper. The method of vibration diagnosis which is importance to the machine's fault diagnosis was introduced in detail and introduced with its related signal processing method. The expert knowledge and measurement of fault were formed after the analysis of the ordinary fault's mechanism of rotate machine. The mathematics model of three layer BP neural network was deduced after introduced the neural network's principal, function, model etc. thus build up the model of intelligent fault diagnosis and introduced the function and model of each module in detail. Combine the actual experience of fault diagnosis, a software was developed use the development tool of VC++6.0. The algorithm of control variable steps was designed to improve the system's learning in view of the faultiness of BP algorithm.A model piece with wavelet transform was developed in this paper, in view of the many nonlinear vibration of running-machine. Because the FFT has some limits in the dispose of vibration signal, but wavelet transform which has the characteristic of time-domain and frequency-domain can guarantee the complete of information. Therefore, while at introduce of the physical model and the mathematical model of wavelet, set forth its application in fault diagnosis. Thus provide a straight show with the equipment's state for the technical personnel which coordinate the system's diagnosis.