关键词: 果蝇算法, Volterra级数, 支持向量机, 模拟电路, 故障诊断

Abstract: Aiming at the problem of the localization of nonlinear components in analog circuits, an improved Drosophila algorithm is proposed to optimize the support vector machine (SVM) fault diagnosis method. Firstly, the output signal of the diagnosed circuit is sampled, the characteristics of the output signal are extracted by the Volterra series, and then the improved fruit fly algorithm is used to optimize the kernel function parameters and structural parameters of the SVM. The diagnosis model is established and the fault is established in the logarithmic amplifier circuit for diagnostic classification. Using MATLAB software to carry out simulation experiments, through experiments we can see that this method can effectively avoid the random selection of support vector machine parameters, help to improve the diagnostic accuracy and diagnostic speed.

Key words: fruit fly optimization algorithm, Volterra series, support vector machines, analog circuit, fault diagnosis