Research On Analog Circuit Fault Diagnosis Based On Fusion Analysis Of Entropy Weight And Grey Relational Method

As an indispensable part,analog circuit has been widely used in electronic devices.However,analog circuit is prone to failure due to its inherent properties,which may lead to the paralysis of the entire circuit system.Some kinds of analog circuit failures can even pose a serious threat to people's lives and property.Therefore,it is of great significance to investigate fault diagnosis technology of analog circuit.This thesis proposes an algorithm named "Fusion analysis of Improved Entropy Weight method and Improved Grey Relational method(FIEWIGR)" for analog circuit fault diagnosis(mainly including resistance soft fault diagnosis and circuit state detection)to realize the correct fault diagnosis of analog circuit objectively through a small number of samples.The main work of this thesis is listed as below:1.After a thorough investigation,this thesis summarizes two improvement measures for the Grey Relational Method(GRM)and the Entropy Weight Method(EWM),whose necessity is elaborated by mathematical model and theoretical analysis.The first is the improvement of GRM:the method of obtaining the grey relational coefficient is improved.Instead of taking 0.5 directly according to previous experience,the optimum value of grey relational coefficient is calculated dynamically by using datasets to avoid the influence of subjective factors on the evaluation results;The second is the improvement of EWM:the method of weight calculation is improved to solve the problem that the index weight is contrary to the principle of entropy weight when the index information entropy approaches to 1.2.This thesis applies FIEWIGR algorithm to circuit state detection in fault diagnosis of analog circuits.The DC circuit,fourth-order active low-pass filter circuit and continuous-time state-variable(CTSV)filter circuit are employed as the objects to be diagnosed,and the state detection experiments are carried out based on FIEWIGR algorithm.The experimental results are consistent with the comparison diagram of the output response of fault circuit simulated by Multisim,which means that the FIEWIGR algorithm can realize correct state detection of analog circuit.3.This thesis applies FIEWIGR algorithm to resistance soft fault diagnosis of analog circuits.By using three classical testing circuits,FIEWIGR algorithm is proven to be highly effective in soft fault diagnosis of resistors.It is shown that the method of calculating entropy weight overcomes the shortcomings of traditional EWM.Finally,300 sets of entropy weights are generated randomly by MATLAB,and the diagnosis results of traditional grey correlation analysis are analyzed.It is proved that the robustness of FIEWIGR algorithm is better than that of grey correlation analysis.