摘要
针对当前模拟电路故障特征和参数选择的不匹配难题,提出了一种特征和神经网络参数同步选择的模拟电路故障模型。首先提取模拟电路的信号特征,并进行归一化处理,然后将特征和神经网络参数编码成一个粒子,每一个粒子根据自身和粒子群的最优位置寻找最优特征和神经网络参数,最后建立模拟电路故障诊断模型,并将其应用具体模拟电路故障诊断应用中。结果表明,模型的模拟电路故障率达到97%以上,而且模拟电路故障结果要优于当前经典模型,具有更高的实际应用价值。
In view of the mismatch between analog circuit fault features and selected parameters,this paper puts forward a fault diagnosis of analog circuit based on improved particle swarm optimization algorithm selecting features and neural network parameters. Firstly,the signal features of the analog circuit are extracted and normalized,and secondly the features and parameters of the neural network are coded into a particle, and the optimal features and the neural network parameters are obtained according to the particle itself and the particle swarm optimal position, and finally, the fault diagnosis model is established and applied into the analog circuit fault diagnosis. Results show that analog circuit fault correct rate of the proposed model is above 91.6%,and the analog circuit fault result is superior to the classic analog circuit fault models,and the model has higher practical application value.
出处
《电网与清洁能源》
北大核心
2015年第7期44-48 53,53,共6页
Power System and Clean Energy
基金
江苏省淮安市农业科技支撑计划项目(SN12058)~~
关键词
模拟电路
故障诊断
粒子群优化算法
特征选择
analog circuit
fault diagnosis
particle swarm optimization algorithm
feature selection
