离子推进系统仿真与故障诊断研究

Simulation and Fault Diagnosis of Ion Propulsion System

关键词:  离子电推进  系统仿真  神经网络  故障模式  故障诊断

Key words:  Ion electric propulsion  System simulation  Neural network  Failure mode  Fault diagnosis

摘要:

离子电推进系统由于组成复杂，发生故障的概率大大增加，通过系统仿真并在此基础上开展故障诊断研究，对于识别故障原因及保障离子推力器系统正常工作具有重要作用。本文结合离子电推进系统各子系统数学模型，利用Matlab/Simulink实现了离子推力器系统仿真模型，基于LIPS-300离子推力器数据，对其性能进行了仿真，仿真模型输出结果与相关文献中已有的结果一致。采用该系统仿真模型结合故障因子，对故障状态的离子推力器系统进行了仿真，利用得到的故障数据与Matlab神经网络工具箱建立了离子推力器故障诊断系统。利用仿真模型额外生成另一组已知对应故障模式的故障状态运行数据，并利用该组数据对故障诊断系统进行了诊断能力测试，系统诊断结果与其已知对应故障模式相比，正确率为93.8%。

Abstract:

Due to the complex composition of the ion thruster system， the probability of failure is greatly increased. It is very important to carry out the fault diagnosis research of ion thruster through the system simulation to ensure the normal operation of the ion thruster system. In this paper， the mathematical model of each subsystem of the ion electric propulsion system is combined with Matlab/Simulink to realize the simulation model of the ion thruster system. Based on the data of the LIP-300 ion thruster， the performance of the simulation model is simulated. The output results of the simulation model are consistent with the existing results in related literature. Using the system simulation model combined with the fault factors， the ion thruster system in the fault state was simulated. The fault data obtained and the Matlab neural network toolbox were used to establish the ion thruster fault diagnosis system. The simulation model was used to generate another set of operating data of the known fault state corresponding to the fault mode， and the diagnostic ability of the fault diagnosis system was tested by using this set of data. Compared with the actual fault mode， the accuracy of the diagnosis result was 93.8%.