应用神经网络识别液体火箭发动机的故障模式

简要介绍了神经网络的基本理论和误差后向传播算法，给出了应用神经网络的液体火箭发动机故障诊断系统框图，分析了液体火箭发动机的故障模型，提出了数据训练神经网络，为网络训练提供了新的途径。最后，基于模型数据，应用神经网络识别液体火箭发动机的几种故障模式。

Recognizeing the Fault Patterns of Liquid Rocket Engine by Neural Networks

The basic theory of neural networks (NN) and BP algorithm are briefly described. A system block diagram of liquid rocket engine fault diagnosis by using the neural networks is given. By analysing the fault model of the liquid rocket engine, the method of using data from the fault model to train the neural networks is put forword, which provides a new path for training the neural networks. Finally, based on the model data,the neural networks are applied to recognize several fault patterns of the liquid rocket engine.