基于神经网络和故障字典相结合的末制导雷达故障诊断技术

采用故障字典和BP神经网络相结合的方法，利用Muhisim软件进行电路仿真，再由实测样本数据对BP神经网络进行训练，完成对网络各参数的设置。仿真结果表明，该方法可以较好地将故障定位到模块级中的元器件，也可为其他型号的雷达模块级故障诊断技术研究提供借鉴。

Fault Diagnosis Technology Based on Neural Network and Fault Dictionary of Terminal Guidance Radar

In this paper with the method of fault dictionary and BP neural network, the Muhisim software was used for cir- cuitry emulation. Then the BP neural network was trained with fact data, and the setup of parameters was accomplished. Result of the simulation indicated that this method could preferably localize fault to components of block-level, which could use as reference to other research on block-level fault diagnosis technology for radar.