Research On Fault Diagnosis And Maintenace Decision Of Double In And Out Ball Mill In Power Plant

State-maintenance is a way, which is based on equipment's healthiness to arrange maintenance.Milling-machine is one of the most accessorial machines in power plant. Power plant have hoped to apply fault diagnosis and maintenance decision to milling-machines. With the development of fault diagnosis and application of computers, Milling-machine's diagnosis comes to be true.In this thesis, the method of combining neural network with Expert System (ES)has been used for fault diagnosis in milling-machine firstly, while wavelet analysis is also used for signal process. Milling-machine intelligent fault diagnosis and maintenance decision system is upbuilt, which include diagnosis reports and maintenance decision reports. So the power plant can reduce accidents and economic losing. This thesis mainly includes:On the bases of the ball mill fault diagnosis, the diagnosis parameters are determined.Wavelet packet decomposition technology has been used to analyze the fault signals in ball mill, a kind of fault feature extracting method based on Wavelet packets decomposition is put forward. The result have shown that the wave coefficients of different signal in different frequency domain are not the same, so that the features can be token by wave coefficients of different frequency domain.The neural network model of RBF for diagnosis has been built, the energy of signal are taken as fault feature parameters, and the method of extracting feature vector is given.Fault diagnosis system combining neural network with ES is developed, digital information is processed by RBF neural network while symbols are deal with ES.Based on result of diagnosis, the strategy of fault processing has been suggested, the consultation of fault processing is studied.To carrying out the fruit of this thesis can establish a fault diagnosis and maintenance decision system, which can offer credible gist for state-maintenance.