摘要
为实现装甲车辆电源系统故障诊断的智能化,在Simulink中建立爪极同步发电机数学模型和桥式不可控整流电路来构建装甲车辆电源系统基础模型,通过设置各类故障,采集不同的电压故障波形中的数据并进行归一化处理;在Matlab中编写、运行BP神经网络训练程序,得到网络模型,并取测试样本进行验证;结果证明了该训练后的神经网络的正确性,表明了BP神经网络在装甲车辆电源系统故障诊断中具有可行性,达到了对整流器故障诊断的目的。
To achieve the intelligence of fault diagnosis in armored vehicle power system, the paper establishs the mathematical model of claw-- pole synchronous generator in Simulink and the full-- bridge and non-- controlled rectified circuit, which sets up the basic model of armored vehicle power system. Setting kinds of diagnosis, the data of various voltage diagnosis waveforms is collected and dealt with the unitary measure. Writing and running the training program of BP neural network in Matlab, acquires and validates the net model. As a result, the neural network trained is proved be excellent, BP neural network is feasible greatly in fault diagnosis of armored vehicle power system, which achieves the aim at fault diagnosis of rectifier.
出处
《计算机测量与控制》
CSCD
北大核心
2012年第10期2719-2721,共3页
Computer Measurement &Control
