人工神经网络在维修性预计中的应用

Application of Neural Network in Maintainability Prediction

维修性预计运用BP神经网络,并以MatlabBP工具箱函数建立产品的维修性预计模型。其建模步骤包括:确定训练样本、建立网络并对其进行初始化、对网络进行训练及对产品维修性参数预计。并对某火炮中主要机构进行维修性分析和仿真,通过对训练样本规范化处理,再应用相应数据预计网络模型,最后检验网络训练效果。

The maintainability prediction model based on BP neural network was set up with Matlab function of BP algorithm according to main factors affecting on products maintainability. Its processes includes confirming training samples, putting up the network and initializing it, training the network and predicting maintainability parameters of the products. Maintainability analysis and simulation was made for the main mechanisms of some artillery. The trained samples were processed for standardization, and then network model was predicted with pertinent data. At last, the result of network training was checked with the corresponding prediction model.