Signal Characteristic Analysis And Fault Diagnosis Algorithm Optimization Of Analog Circuit

With the increasingly extensive application of electronic systems,the complexity and integration of electronic circuits are deepening,and the requirements of circuit testing technology are also higher and higher.As an important part of electronic circuit system,the optimization of fault diagnosis technology is the necessary way to improve the test technology of electronic circuit.However,due to the tolerance and nonlinearity of analog circuit components,as well as the limited measurable nodes,it is difficult to obtain ideal test results in the previous fault diagnosis methods,so it is necessary to explore more efficient test and diagnosis methods.In this paper,based on neural network technology,combined with wavelet packet analysis method,aiming at the problem of soft fault in analog circuit,we introduce intelligent algorithms such as particle swarm optimization algorithm and weed algorithm to optimize BP neural network in order to improve the accuracy and efficiency of fault diagnosis.This paper mainly studies from the following aspects:1.The application of wavelet packet analysis and neural network in analog circuit fault diagnosis is explored.Wavelet packet analysis has good time-frequency analysis characteristics and can extract high-quality fault features.Neural network has excellent generalization learning ability and nonlinear mapping performance.In this paper,the fault feature extraction of analog circuit is realized by wavelet packet analysis,and then the fault location of analog circuit is realized by combining the classification and recognition ability of neural network.2.An analog circuit fault diagnosis model based on particle swarm optimization BP network parameters(PSO-BP)was constructed to improve the diagnostic accuracy and convergence speed of the BP network classification model.Secondly,aiming at the problem that PSO is easy to fall into local convergence,an improved classification model based on Improved Particle Swarm Optimization(IPSO)algorithm to optimize BP network parameters(IPSO-BP)is proposed by introducing mutation operator and nonlinear decreasing weight strategy.Sallen-key and CTSV filter circuits were used to carry out the simulation and experiment of fault diagnosis.The results show that the IPSO-BP classification model is better than BP network and PSO-BP classification model,and has faster convergence speed and higher diagnostic accuracy.3.An analog circuit fault diagnosis model based on the intrusion weed algorithm to optimize the parameters of BP network(IWO-BP)was established to further improve the diagnostic accuracy and convergence speed of the classification model based on BP network.Then,the adaptive parameter selection strategy is introduced,and the crossover,mutation and selection operators of differential evolution algorithm are integrated,an improved classification model of adaptive weed hybrid algorithm(AIWODE)to optimize BP network parameters(AIWODE-BP)is proposed.Two typical circuits were used to verify the classification effect,the results shows that AIWODE-BP classification model can effectively prevent the local convergence,shorten the classification time Fault classification time and obtain higher fault diagnosis accuracy.4.An improved weed classification model based on Hybrid Particle Swarm Optimization(HPSO)is proposed to optimize the parameters of BP network.Three typical circuits are simulated and analyzed,and compared with IPSO-BP,AIWODE-BP,CS-BP,IWO-BP and PSO-BP classification models.The results show that HPSO-BP classification model can obtain the optimal fault classification effect.The work of this paper is a beneficial exploration of intelligent fault diagnosis method,which enriches the research results of analog circuit fault diagnosis to a certain extent.