数据驱动的飞行器智能故障诊断系统研究

Study on a data-driven intelligent fault diagnosis system of spacecraft

摘要:

飞行器装备及试验技术的快速发展带来试验数据的指数级增长，传统的试验数据管理与分析手段已经无法满足装备研制与优化设计的需求。提出一种数据驱动的飞行器智能故障诊断技术，通过将人工智能技术与故障诊断技术相融合，能够从飞行器历史测控试验数据库中，自动获取参数的阀值、数据链特征与关联关系，形成准确的飞行器参数知识模型库。通过构建数据驱动的飞行器智能故障诊断系统，将知识模型库中的知识与实时试验数据进行比对分析，实现实时数据的自动化、智能化判读，进而发现航天飞行器可能存在的潜在故障，提前预测，将风险降低到最小。

Abstract:

The rapid development of spacecraft test technology brings exponential growth of test data. The traditional data management and analysis methods cannot meet the requirements of equipment development and optimization design. A data driven intelligent fault diagnosis technology of spacecraft is proposed, which can automatically obtain the threshold value, data link characteristics and correlations of parameters from the historical aircraft test data by integrating AI technology, and form an accurate spacecraft parameter knowledge model base. By constructing a data driven intelligent fault diagnosis technology of spacecraft, real-time data can be interpreted automatically and intelligently.