| 摘要: |
| 闭环控制系统的反馈调节机制，可能破坏开环系统的可辨识性，导致误报率和漏报率同时上升，还可能导致同一故障模式下的多个变量发生异常，给闭环控制系统故障识别带来难度。针对单输入单输出系统，分析了闭环控制对可辨识性、误报率和漏报率的影响，对多输入多输出系统，理论上推导了闭环控制系统传感器故障的传播机理，分析了闭环控制系统对系统变量的影响关系，利用神经网络构建了闭环故障识别算法。数值仿真结果验证了闭环对故障系统诊断的不利影响，而卫星姿态控制系统的仿真结果标明：与传统方法相比，提出的方法识别性能更高。 |
| 关键词: 卫星控制系统 故障识别 闭环故障 神经网络 滑动窗口 |
| DOI：10.15982/j.issn.2095-7777.2019.04.009 |
| 分类号: |
| 基金项目:国家自然科学基金资助项目（61773021）；湖南省杰出青年学者自然科学基金资助项目（2019JJ20018）；湖南省自然科学基金资助项目（2019JJ50745）；空间智能控制实验室科技基金资助项目（HTKJ2019KL502007） |
|
| Neural Network based Fault Diagnosis Methodfor Satellite Attitude Control System |
|
SUN Bowen1, HE Zhangming1,2, WANG Jiongqi1
|
|
1.College of Liberal Arts and Sciences, National University of Defense Technology, Changsha 410073, China;2.Beijing Institute of Spacecraft System Engineering, China Academy of Space Technology, Beijing 100094, China
|
| Abstract: |
| The closed-loop in control system usually brings difficulties for fault identification, because it may destroy the identification ability of the parameters andenhance the false alarm rate as well as the missed alarm rate,leading to the abnormality of multiple variables under the same faulty mode. Firstly, the effect of close loop is analyzed for the single-input single-output (SISO)system,while for the multiple-input multiple-output(MIMO)system the fault propagation is derived theoretically,and the influences on system variables are analyzed. Secondly,the neural network is used to construct the fault identification method. Finally,the mathematical simulation validates the negative effect of closed-loop on fault diagnosis,and the simulation of satellite attitude control systemverifies thattheproposed methodis more effective, comparing with the traditional ones. |
| Key words: satellite control system fault identification closed-loop fault deep learning network moving window |
