Application And Reaserch Of Wavelet Neuralon Network On Fault Diagnose Of Radiometer Circuit

With the development of integrated circuit board technology and intellectualized fault diagnose system, the application of intellectualized fault diagnose on circuit board is paid more and more attention by us. Use neural network intellectual precognitive technologies on circuit board fault diagnose which has been proved effective and has actually developed a new research direction. For the neural network has the function of processing complex multi-patterns and carrying on the association, the extrapolation and the memory, it is particularly suitable for fault diagnosis system. Fault diagnosis by applying the neural network in analog circuit has become the most recent development tendency。The paper is based on the actual project and takes the part of the radiometer system circuit as research object. The use of fault diagnose technology and MATLAB software make it possible that fuse neural network and wavelet function, found database for fault diagnose and realize the successful diagnose for digital-analog hybrid circuit.After careful analyses of circuit figuration, design some probe spot which can reflect the circuits typical fault and choose the voltage value to express the feature of the fault. Using the virtual signal generator generated signal input nodes of a circuit with virtual multimeter excitation signal, in response to the output node acquisition circuit output. In MATLAB structures using neural network platform, with good time field and frequency field of local characteristics of wavelet function to replace the traditional BP network of excitation function structure of BP network, a wavelet neural network structure and algorithm. Make a list of the typical circuit fault, in each situation we can get measured data divided to two parts, one part of the data is to train the network. Another part of the test data is used to test the network.Experimental results show that the wavelet neural network can be effective for circuit fault locating and classifying. This measerd can achieve highly accuracy, and diagnostic purpose.