基于相似匹配的液体火箭发动机故障模式挖掘

Similarity-search based mining for fault mode of liquid rocket engine

将时间序列相似性匹配方法引入到液体火箭发动机故障模式挖掘中。针对发动机试车数据的特点,提出了一种基于序变换的时间序列相似匹配算法。该算法具有对时间序列幅值和持续时间不敏感、抗噪声能力强等优点。对某型液体火箭发动机故障数据的相似匹配实验表明:该算法能够为液体火箭发动机的故障检测和诊断提供较好的技术支持。

The method of time series similarity matching （TSSM） is introduced in the fault mode mining of the liquid rocket engine （LRE） . According to the feature of fire-test data, an ordinal transformation based algorithm is proposed, which is insensitive to amplitude, time span and noises of the time series. The TSSM experiments on the data of some LRE indicated that it is quite supportive to fault detection and diagnosis.