基于多模型组合的模拟电路故障诊断方法

针对传统模拟电路故障诊断方法识别准确率低及耗时较长的问题,提出一种基于改进的二进制粒子群优化算法(IBPSO)与隐马尔科夫模型(HMM)的综合诊断方法。该方法利用IBPSO对故障特征进行提取优化,降低故障特征维度,进一步利用HMM对提取的故障特征进行预处理,排除不可能类故障特征,提高了LSSVM的分类准确率。经过仿真结果分析验证,该方法较现有的BP神经网络诊断方法,能够在确保正确率得到提升的基础上,进一步提高故障诊断速度,具有更强的战场环境适用性。

A Method of Analog Circuit Based on Multimodel Combination

In view of the traditional analogy circuit fault diagnosis method with the problem that the low accuracy and the long time-consuming. This paper puts forward an integrated diagnostic method based on the improved IBPSO and the hidden Markov model (HMM). The method extracts and optimizes the fault feature and reduces the fault dimension by using IBPSO,and then uses the HMM to rule out the impossible faulty feature in LSSVM,which increase the accuracy of the classification of LSSVM. Proven by the simulation results analysis,the method,comparing with the existing BP neural network diagnosis one, improves the speed of fault diagnosis on the basis of ensuring the accuracy and has the stronger applicability of the battlefield environment.