Font Size: a A A

Research On Analog Circuit Fault Diagnosis Method Based On Optimization Technique

Posted on:2017-10-29Degree:MasterType:Thesis
Country:ChinaCandidate:T Y LiuFull Text:PDF
GTID:2348330485997299Subject:Control theory and control engineering
Abstract/Summary:PDF Full Text Request
With the continuous development of science and technology,complexity of manufacturing technique is increasing.The requirements of electronic equipment reliability are higher,and the fault diagnosis of analog circuit has great significance for the stable operation of electronic systems.However,the progress of analog circuit fault diagnosis has been hampered by many factors such as fault diversity,nonlinear,component tolerance and so on.The fault diagnosis of analog circuit is a challenging research issue.The shortcomings of traditional diagnostic method are low accuracy and poor performance.As a result,it becomes an important task which the faults of analog circuit are needed to be found quickly and accurately.In the process of analog circuit fault diagnosis,feature extract and parameters optimize of fault diagnosis algorithm are the key to improve the performance of fault diagnosis.Wavelet packet algorithm is used to extract the information of analog circuit.Use IPSO algorithm optimize the parameters of BP diagnosis algorithm.Traditional wavelet transformation in the process of recursive decomposition is not completed,which only retained the approximate coefficient,not retained reservation details,lead to the low frequency resolution.Although the neural networks provide a way to the fault diagnosis of the analog circuit,every neural network has its own limitations.To solve the above problem,a new fault diagnosis model based on the combination of the IPSO and BP neural network is put forward for the analog circuit.Firstly,to solve the problem of standard PSO falling into local optimal easily,an improved PSO based on the time-varying weights and asynchronous learning factor is proposed.The IPSO algorithm is used to optimize weight and threshold of BP neural network to improve the performance of diagnosis method.Wavelet packet not only decompose approximate coefficient of signal decomposition,but also decompose the detail coefficients.And it realized adjust the signal resolution automatically,which has characteristics such as flexible structure and operate timely.WPT-IPSO-BP is proposed to diagnose the 25 KHZ band-pass filter circuit fault,compared with the traditional method.Theoretical research and experimental results show that WPT-IPSO-BP diagnosis performance is better than the traditional method on the diagnosis accuracy and quickness.
Keywords/Search Tags:Analog circuit, Fault diagnosis, Fault feature extraction, Artificial neural network, Intelligent optimization method
PDF Full Text Request
Related items