多元指数加权移动平均主元分析的微小故障检测

发布时间：2018-03-29 10:28

本文选题：微小故障　切入点：主元分析　出处：《控制理论与应用》2014年01期

【摘要】：主元分析(principal component analysis,PCA)是一种有效的数据分析方法,在故障诊断与状态监测方面已得到广泛应用.多元指数加权移动平均主元分析(multivariate exponentially weighted moving average principal component analysis,MEWMA PCA)方法用于解决PCA不能有效检出微小故障的问题.本文深入研究了MEWMA PCA中EWMA影响主元分析进行故障检测的机制,导出了MEWMA PCA可检出微小故障的原因.本文确定了MEWMA PCA中遗忘因子λ、单传感器故障幅值和迟延时间三者的关系,并进行了数值仿真和火电厂磨煤机组运行状态的仿真实验.实验结果验证了MEWMA PCA中EWMA提高PCA的监测性能的机制,并给出了根据系统实际要求来选取合适的遗忘因子值,从而在规定的时间内检出微小故障的实例.
[Abstract]:Principal component analysis (PCA) is an effective method for data analysis. It has been widely used in fault diagnosis and condition monitoring. Multivariate exponentially weighted moving average principal component analysis method is used to solve the problem that PCA can not effectively detect small faults. The mechanism of EWMA influencing principal component analysis (PCA) for fault detection in MEWMA PCA is studied. The reason why MEWMA PCA can detect minor faults is derived. The relationship among forgetting factor 位, single sensor fault amplitude and delay time in MEWMA PCA is determined. The numerical simulation and the simulation experiment on the operation state of coal mill unit in thermal power plant are carried out. The experimental results verify the mechanism of improving the monitoring performance of PCA by EWMA in MEWMA PCA, and the appropriate forgetting factor value is given according to the actual requirements of the system. Thus in the specified time to detect the small failure of the example.
【作者单位】：华北电力大学新能源电力系统国家重点实验室;
【基金】：国家重点基础研究发展计划“973”计划资助项目(2012CB215203) 国家自然科学基金重点资助项目(51036002) 中央高校基本科研业务费专项资金资助项目
【分类号】：TM621

【参考文献】