基于PCA的多变量系统故障诊断的研究

摘要

应用主元分析实现多变量系统的故障检测与诊断,既消除了各变量之间的关联,也减小外界噪声的影响,能够有效检测诊断出系统的故障源,为实现复杂容错控制提供了保障.通过对多变量液位控制进行试验验证,主元分析法能够有效地对生产过程进行监测,并准确及时的分析判断系统中发生的故障.%It was realized multivariate system fault detection and diagnosis using principal element analysis.Not only eliminate the correlation between variables,but also reduce the influence of the outside noise.It was find effectively diagnose system fault source.It was security to realized fault-tolerant control.It was carried on the experimental confirmation in the variable liquid level control test verification,principal component analysis can effectively to the production process for monitoring,and accurate and timely analysis judgment fault occurred in the system.