Based On Ga Improved LM-BP Neural Network For Analog Filter Circuit Fault Diagnosis

Analog filter circuit is an important part of the electronic system, the system’s stability and reliability is affected by its working conditions. the traditional analog filter circuit fault diagnosis technology shows much limitation in most theories and methods. Therefore, the establishment of intelligent analog filter circuit fault diagnostic is becoming increasingly important. Intelligent fault diagnosis techniques of analog filter circuit not only improve the speed, accuracy and reliability to assure the stable operation of electronic equipments and control system, but also posses very important significance to optimize the proper design of analog filter circuits.The BP neural network which is used for analog filter circuit fault diagnosis can improve the fault diagnosis reliability, while the standard BP algorithm has slow convergence and stagnation behavior. The genetic algorithm with a BP neural network can synchronously process a large number of non-linear data, pattern recognition, and diagnostic features such as forecast, especially analog circuit fault diagnosis of nonlinear complex engineering to overcome the defect of BP neural network. As a result, the use of the combination of genetic algorithm and BP neural network for analog filter circuit fault diagnosis is an important trend. This text studies technically the GA improved LM-BP optimization, the combination of GA improved LM-BP algorithm, and describes its practical application in analog filter circuit fault diagnosis.A test with a four-storey neural network model which is designed by using a Matlab simulation software, based on the GA improved LM-BP algorithmic model, together with active double-T band-stop filter circuit fault diagnosis verifies good arithmetic speed and accuracy of the GA improved LM-BP neural network model in analog filter circuit fault diagnosis, and provides a reference for complex analog circuit fault diagnosis.