基于独立源分析的过程监测及故障诊断方法

多元统计过程控制(MSPC)要求观测数据服从正态分布,而实际的工业过程数据大都不满足正态分布条件。独立源分析(ICA)是近几年才发展起来的一种新的统计方法,可以克服对数据分布的依赖。为此以ICA算法为核心,引入一种新型的过程监测及故障诊断方法。应用ICA提取独立源,利用I2图,Ie2图和SPE图进行故障检测,将变量重构图用于诊断故障。以三水箱系统为背景进行的实验研究,验证了该方法的有效性。

Process Monitoring and Fault Diagnosis Based on Independent Component Analysis Method

Multivariate statistical process control (MSPC) is based upon the assumption that the observed data must be subject to normal probability distribution, which sometimes can not be satisfied. Independent component analysis (ICA) is a recently developed method, which can overcome the need of the data distribution. A new method was introduced for process monitoring and fault diagnosis based on ICA. ICA was used to extract the independent components, and I2, Ie2and SPE charts were used for fault detection, and Variables contribution plots were considered for fault diagnosis. At last, the simulation results of three-tank system reveal this method is very effective.