论文部分内容阅读
时间序列方法是数据挖掘技术的一个重要分支,在航天器故障诊断领域具有非常广阔的应用前景。本文在系统研究时间序列数据挖掘技术现状基础上,重点介绍了一种基于时间序列数据挖掘的航天器故障诊断方法,包括确定诊断对象、航天器遥测数据选取、航天器遥测数据再处理、航天器遥测数据时间序列特征表示、航天器故障诊断的时间序列数据挖掘方法执行、挖掘结果的解释与评估、软件系统开发共7个步骤。最后通过卫星实例描述了该方法在航天器故障诊断中的具体应用。
Time series method is an important branch of data mining technology and has a very broad application prospect in the field of spacecraft fault diagnosis. Based on the systematic study of the current status quo of time series data mining technology, this paper mainly introduces a method of spacecraft fault diagnosis based on time series data mining, including the determination of diagnostic objects, selection of spacecraft telemetering data, spacecraft telemetry data reprocessing, spacecraft Time series features of telemetry data, execution of time series data mining methods of spacecraft fault diagnosis, interpretation and evaluation of mining results, and software system development in a total of seven steps. Finally, the practical application of this method in spacecraft fault diagnosis is described by satellite examples.