基于动态概率主元分析的统计过程监测

Process monitoring based on dynamic probabiUstic principal component analysis

概率主元分析(PPCA)已广泛应用于工业过程监测。然而,PPCA法仅构造了生产过程的静态线性关系,处理具有较强动态特性的实际工业生产过程效果较差。为此提出动态概率主元分析(DPPCA)法,对经过时谱扩展后的变量数据阵.通过期望最大化(EM)算法建立生成模型,从而将静态PPCA推广到动态多变量过程。最后将此法应用于TE过程的仿真研究.结果表明该法有效。

Probabilistic Principal Component Analysis（PPCA） has been widely used for monitoring industrial process. However,PPCA only constructs linear static relations among the process variables, it can＇t effectively deal with the real industrial process which possesses strong dynamic characteristic. Using the expectation and maximization （EM） algorithm, the dynamic PPCA model is built to cope with the data matrix extended by time series. According to the technique, static PPCA can be extended to monitor dynamic multivariate process. At last, the simulation results of TE process reveal this method is very effective.