基于神经网络专家系统的炮兵射击指挥系统故障诊断

基于神经网络专家系统的炮兵射击指挥系统故障诊断系统采用多层感知器组建专家知识库,建立激光仪子网、故障子模块,构建合成网络模块及训练样本.将各训练样本输入相应神经网络中训练,得到的权值和阈值存入数据文件.推理模块调用诊断知识库,输入各故障征兆,计算隐含层和输出层的神经元输出.解释模块由给定的法则判定输入神经元的输出是否有故障.

Fault Diagnosis for Artillery Firing Command System Based on NN Expert System

Fault diagnosis system for artillery fire command system based on NN expert system consists of expert knowledge base formed with multi-layer perceptions, radar subnet, fault sub-module, integrated net-module and training samples. Each training sample was inputted into corresponding NN to train, gained weights and thresholds were stored into the data file. Reasoning module is used to call diagnosis knowledge base to input fault characters, and account the outputs of hidden layer and output layer. Explaining module is used to judge if the outputs of input neutral node have faults or not according to the rules.