Analog circuit fault diagnosis based on CBEA-SVM

The problems of SVM include easy to fall into local optimum, repeated oscillation and slow convergence. Because:(1) mutation or local mutation of cloud evolutionary algorithm; (2) randomness and ambiguity of cloud model; (3) expected Ex control certainty of the optimization process, entropy En and super entropy He control uncertainty of the optimization process, these advantages make the cloud evolutionary algorithm can effectively avoid the above problems. Therefore, this paper proposes an analog circuit fault diagnosis method based on cloud evolutionary algorithm optimization support vector machine (CBEA-SVM) to solve these problem. CBEA is used to find penalty factor and kernelfunction variance of the support vector machine, SUM optimized by CBEA is used for fault chtssification. Using a Sallen-key bandpass filter and a CTSVfilter circuit as circuits to be tested, the diagnostic results show that the diagnostic accuracy of CBEA-SUM has reached more than 98%, and it proves that CBEA can avoid the problems of traditional optimization algorithms.