| 摘要: |
| 随着在轨航天器爆发式增长，航天测控数传任务越来越密集，同时航天测控数传一体化系统的复杂度越来越高，传统的故障诊断和周期性维护已经无法满足装备使用要求。根据航天测控数传一体化系统的特点对健康管理系统的架构及组成进行了研究，重点分析了关键设备的传感器设计、基于层次分析法和D-S证据合成法相结合的系统健康评估方法以及基于循环神经网络等机器学习算法的故障预测方法。应用健康管理系统，可有效提高航天测控数传任务的可靠性，减少设备的故障概率，降低维修成本。 |
| 关键词: 航天测控数传一体化系统 健康评估 健康管理 机器学习 |
| DOI： |
|
| 基金项目: |
|
| Design of a health management system for satellite ground stations |
| WANG Junhui,LIU Aiping,XIAO Xiaobing |
| (Southwest China Institute of Electronic Technology，Chengdu 610036，China) |
| Abstract: |
| With the explosive growth of in-orbit spacecrafts,the space mission is performed more and more frequently,and the complexity of satellite ground station is daily on the increase,the traditional fault diagnosis and periodic maintenance has been unable to meet the requirements.This paper studies the architecture and composition of health management system according to the equipment characteristics.Meanwhile,sensor design of key device,health evaluation methods based on analytic hierarchy process method combining with D-S evidence theory and fault prediction method based on recurrent neural network(RNN) are analyzed.By applying the health management system,the reliability of satellite ground station can be effectively improved,the fault probability can be decreased and maintenance cost can be reduced. |
| Key words: satellite ground station health evaluation health management machine learning |
