Recent developments in data mining technologies make it possible to use the massive archived spacecraft telemetry data to produce the advanced system health monitoring applications for fault diagnosis. These data driven applications can get the knowledge or model from spacecraft telemetry data automatically or semi-automatically. In this paper, we first introduce the conventional approaches for fault diagnosis such as limit-check, expert system and so on, and then describe the characteristics of the spacecraft telemetry data which should be considered in the actual data mining applications. Last we propose some feasible approaches using data mining technology for spacecraft fault diagnosis and point out the future works.%近年来数据挖掘技术的快速发展使得利用航天控制中心积累的大量测控数据进行航天器在轨故障诊断成为可能.基于数据挖掘的故障诊断技术能够从历史数据中自动或半自动地获取潜在的诊断知识,从而有效解决故障诊断中知识获取困难的问题.在研究航天器实际测控数据特点的基础上,分析了传统故障诊断技术的优缺点,提出了基于数据挖掘技术进行航天器故障诊断的可行方法,指出了基于数据挖掘的航天器故障诊断技术下一步研究的主要方向.
展开▼