论文部分内容阅读
针对现有监测方法对时变过程易产生误警且对微弱故障的检测能力不足等问题,提出一种基于可变遗忘因子的改进递归主元分析(recursive principal component analysis,RPCA)方法用于自适应故障监测。在主元模型的在线更新中引入一种可变遗忘因子,并为不同的模型参数设置不同的遗忘因子;在相关矩阵的递归分解中引入部分奇异值分解的思想,递归计算负荷矩阵和特征值对角矩阵;提出一种控制限递归更新方法,实现控制限的自适应更新。对某型雷达发射机工作过程的监测结果表明,改进的RPCA方法能自适应地跟踪过程的时变,有效地减少了对正常工况调整的误警和对微弱故障的漏报。
Aiming at the problems that the existing monitoring methods are prone to false alarms in the time-varying process and the detection capability of weak faults is insufficient, an improved recursive principal component analysis (RPCA) method based on variable forgetting factor is proposed for self- Adapt to fault monitoring. A variable forgetting factor is introduced into the online updating of the principal component model, and different forgetting factors are set for different model parameters. The idea of partial singular value decomposition is introduced in the recursive decomposition of the correlation matrix, and the load matrix and eigenvalue are recursively calculated Diagonal matrix. A new control limit recursive updating method is proposed to realize adaptive updating of control limits. The monitoring results of the working process of a certain type of radar transmitter show that the improved RPCA method can adaptively track the time-varying process, effectively reducing false positives and negligent reports of weak faults.