摘要
装甲车辆上电气系统电路板功能日趋多样与完善,但同时其复杂程度也日益提高,故障层次越来越多,故障现象与故障原因的映射关系更加复杂,组合故障频发,传统的故障诊断方法已不能满足其故障诊断的要求;针对此,设计了基于免疫遗传算法优化的BP神经网络对电路板进行故障诊断,并在免疫和遗传过程中保留了部分训练最优解;实现了神经网络收敛速度的提高,使用Matlab编程优化算法并完成了电路板仿真故障的诊断;通过实验验证了该诊断模型的准确性和可靠性,为电气系统通用检测设备的神经网络诊断方法实现提供了理论支撑。
In view of the armored vehicle circuit board size is bigger, function has become increasingly diverse and perfect at the same time, and the complexity is improved, fault levels become more complex, the relationship of the fault phenomenon and the cause of the prob- lem becomes more complex, combination malfunction appears frequently, the traditional fault diagnosis methods cannot meet the require- ments of circuit board fault diagnosis. The thesis designs immune genetic algorithm to optimize the BP neural network fault diagnosis to cir- cuit board, and retained part of the training optimal solution in the process of immune and genetic. The algorithm realize the improvement of the neural network convergence speed, using Matlab programming optimization algorithm and circuit board fault diagnosis simulation is com- pleted. The accuracy and reliability of the diagnosis model is verified by experiment, and the neural network diagnosis method provides a the- oretical support for the general testing equipment of electrical systems.
出处
《计算机测量与控制》
2017年第6期9-13,共5页
Computer Measurement &Control
关键词
电路板
神经网络
故障诊断
circuit board
neural network
fault diagnosis
