基于改进PSO-LSSVM的模拟电路诊断方法

Analog Circuit Fault Diagnosis Method Based on PSO-LSSVM

文中提出了一种基于预处理,小波包分析,归一化处理,改进粒子群算法与最小二乘支持向量机(Improved Particle Swarm Optimization-Least Squares Support Vector Machine,IPSO-LSSVM)结合的模拟电路故障诊断方法。针对待诊断的模拟电路,首先对信号进行预处理,其次进行小波包分解,通过归一化等方法进一步处理故障特征信息,作为PSO-LSSVM的输入样本。在充分考虑传统粒子群优化算法中容易陷入局部极小等缺陷的基础上,提出了利用新的模拟退火算法改进PSO-LSSVM的方法。文中优化了模拟电路故障的特征提取方法与分类效果,有效地提高了故障诊断的精度和效率。

Based on the preprocessing,wavelet packet analysis,normalization,Improved Particle Swarm Optimization （ IPSO） algorithm and Least Squares Support Vector Machine （ LSSVM） ,a new analog circuit diagnosis method is proposed. The proposed method uses the wavelet packet decomposition after the preprocessing to deal with the signals. The feature information is extracted by multi-resolution and normalization. The input patterns are satisfied when the feature information applies to the PSO-LSSVM. Under considering the character-istics of the traditional PSO algorithm,the IPSO algorithm based on Simulated Annealing （ SA） algorithm is used in least squares support vector machine. In this paper,optimize analog circuit fault feature extraction and classification results,effectively improving the accuracy and efficiency of fault diagnosis.