Analog Circuit Fault Diagnosis Based On Quantum Technology

The fault d iagnosis o f ana log c ircuits is an advanced synthes is intercrossedsubject，and the hotspot in fault diagnosis. It has deve loped a great deal of theory andapproaches. However，the approaches are limited to deal with fa ult diagnos is becauseof the variety and comp lexity of the ana log circ uits especia lly the nonlinear ana logcircuits. Due to the inherent characteristics of ana log circ uits， such as fa ilurephenome non is comp lex and diverse，continuous var iation o f the ele ment parameterand tolerances on co mponent parameter，etc. Hence，it is important to explore somemore efficient diagnosis theories and methods to solve these problem.Traditio na l methods to diagnos is of ana log circ uits needs to solve large numberof nonlinear equations, the computatio n is very large. And the diagnos is effect is notvery satisfactory. However, modern intellige nce technolo gy provides an effective wayto the fa ult d iagnosis o f ana log c irc uits. Quantum genetic a lgorithm (QGA) can so lvecomp licated proble ms whic h cannot be solved by the traditiona l technolo gy and it hasa function o f researching all k inds of the fie lds. QGA adopts some concepts andtheory of quantum computing. Artific ia l neura l network is a paralle l distr ibutedsystem, which us ing a comp letely differe nt mechanis m with the traditiona l artific ia linte llige nce and informatio n processing techno logy. And overcome someshortcomings of traditiona l methods. ANN has many unique features, such as adaptive,self-organization and real-time learning. And it has been one of the most activeresearch areas recently can be contr ibuted to solve proble ms in var ious practica lfields.This paper located the fault dia gnos is of the nonlinear ana log circuits us ingquantum genetic algor ithm and artificia l ne ura l networks according to the screentheory. At first a new approach to an optima l ana log test points se lectio n is proposed.This described method uses a mbiguity set concept and evolutionar y computations todetermine the optima l set of ana log test points. And use quantum genetic algor ithm tosolve the mathe matical model. Mainly introduces the screen theory of s ingle branchcircuit and multip le branch circuit of no nlinear analo g circuits. And found the screendiagnosis method of no nlinear ana log c ircuit. Nonlinear co mponents can be shie ldedwith a few of tests and incentive by screen method, so the fa ilure of no nlinearcompone nts can be isolated by the exter nal c ircuit，a nd it is simp le and feas ib le.Through fa ilure analys is, if the c ircuit still has fa ult, then it can be know the fa ult exists in the linear part of the c ircuit, so it is only need to diagnos is the linear ana logcircuits by artific ia l neura l networks; Othe rwise, if the circ uit is norma l, then it can beknow the fault exists in the nonlinear part of the c ircuit, so we can locate the fa ult byiso lating the nonlinear components one by one. This method can simplify thealgor ithm comp lexity o f fa ult d iagnosis of nonlinear ana log circ uits. It is respective lybased on quantum ge netic algorithm、artific ia l ne ura l network and screen techno logyof nonlinear circ uits. And simulate the exa mpled circuit based on MATLAB andOrCAD. The exper iment res ults show that the proposed methods are feas ible andsimple.