Expert System Of Fault Diagnosis For Fan Based On Neural Network And Wavelet Analysis

Fan is the necessary auxiliary equipment for power, metallurgy, energy and other industrial sectors, it can run properly or not is directly related to the economic benefits of these areas, so the fault diagnosis in fan has become particularly important. Intelligent fault diagnosis is a new discipline involving multiple disciplines and fields, and it is the frontier and hot of fault diagnosis technology. With the improvement and development of neural network theory, fault diagnosis technology based on neural network provides a new way for research and development of expert system.This paper comprehensively applied the technique of feature extraction with wavelet packet and the method of fault diagnosis based on neural network, and built the basic framework of expert system of fault diagnosis for fan based on neural network and wavelet analysis. Wavelet packet analysis is more favorable than wavelet analysis in the properties of time-frequency, and it is very suitable for extracting the feature of vibration signals in fan. Expert system using neural network did not need to organize a large number of production rules, the network could self-organization, self-learning, and it could effectively solve the most difficult problem of knowledge acquisition and knowledge inference in traditional expert system. Training network using BP algorithm with the momentum and adaptive learning rate could accelerate the speed of training network, shorten the time of training network, and could inhibit the network into a local minimum. Training network with standard training samples and training samples containing white noise could make the network have certain fault tolerance; it was more conducive to identify fault type of fan in practical engineering.According to research of this paper, using the approach of Visual C++ and MATLAB programming developed the expert system of fault diagnosis for fan based on neural network and wavelet analysis, and applied it to the instances of fault diagnosis in fan. After analysis and field testing, the result of diagnosis was proved to be accurate and reliable.