Research On Tolerance And Multi-Fault Of Filter Circuit Based On Complex Domain Model

Analog circuit fault diagnosis has always been a hot research topic,and the current research in this field has been very deep.But the solution to the impact of component tolerance and multi-fault isolation methods are still lacking.In view of this,in order to improve the accuracy of analog circuit fault diagnosis,based on the complex domain model,this paper makes fault modeling of analog filter circuit electronic components,studies the multi-frequency diagnosis method,single fault diagnosis with tolerance,the frequency optimization of measuring points based on genetic algorithm and the problem of multi-fault diagnosis.The specific work is as follows:1)The complex domain model of fault element of analog filter circuit is established.Based on the theory of circuit analysis,the complex domain model is established.The excitation signal is a standard sine wave.The fault parameter-scanning domain is configured in the Pspice software to scan the parameters and track the real and imaginary parts of the output voltage.Moreover,the fault model on the complex domain is obtained by MATLAB data processing,and the characteristic information is stored.Components with the same output characteristics form a fuzzy group.2)Research on multi-frequency diagnosis method.The circular model established by the simulation experiment is derived from the binary quadratic equation,so the two fault circle models in the complex domain may have two intersection,one indicating no fault and the other called aliasing point.Because the characteristic function H?j??of the circuit is affected by?.The output bits at different frequencies are different,thus eliminating the aliasing problem.Based on this,the serrated wave signal is used as the excitation signal to expand the input and output signals into sine waves at multiple frequencies and the transmission characteristics at different frequencies are obtained,and the multi-frequency diagnosis is realized.An overlapping curve exists at one frequency,which can be separated at another frequency.3)Research on single fault diagnosis under the influence of tolerance.In the process of multi-frequency diagnosis,considering the influence of tolerance,the fuzziness of fault diagnosis is increased,and the result of fault classification is not easy to visualize,so the support vector machine is introduced for fault classification.Multi-frequency diagnosis can obtain the fault response characteristics of different frequencies under the measuring point at the same time,and increase the dimension of fault feature information,which can improve the accuracy of fault classification.4)Based on genetic algorithm,the optimal and multi-objective optimization of measuring points is studied.In the field of analog circuit fault diagnosis,the testability of the circuit is studied by extracting the frequency characteristics of the measuring point,and the efficiency of fault diagnosis is improved.However,not all the measuring points and frequencies are effective,and in fact,fewer points can affect the complexity of the circuit,and a selection method needs to be proposed.In this paper,single-objective optimization and multi-objective optimization based on genetic algorithm are studied,that is,joint optimization of measuring point optimization and measuring point frequency is realized,so as to ensure the extraction of optimal fault characteristics,so that the accuracy of fault classification is the highest.5)Research on multiple fault diagnosis.Existing multi-fault diagnosis methods,such as:node voltage increment method,wavelet and RBF neural network combined method,k-means fuzzy clustering and so on.Based on the complex model,the relationship between the number of fault diagnosis and the measuring point is as:n>log₂?m+1??n indicates the number of measuring points,m indicates the number of faults?,and the diagnosis of double faults requires at least 2 measuring points,which can simultaneously obtain four-dimensional fault characteristics and improve the fault resolution.