中文摘要：

随着现代数据采集系统在生产中的应用,通常需要同时监控多个相关的过程变量,并且在化工等生产过程中,多元过程数据还会呈现自相关。已有的多元自相关过程控制方法大多仅能检出过程偏移却无法诊断出哪个变量或哪些变量组合会导致过程失控。本文针对多元自相关过程,提出了基于支持向量机（SVM）的过程在线监控和诊断方法。通过构建过程监控和偏移诊断两个分类器,可以对生产中的数据进行在线监控和诊断,基于matlab的仿真结果表明,提出的方法具有更好的监控性能和更高的诊断正确率。

英文摘要：

With the application of modern data collection system in manufacturing process,it is common to monitor several correlated process variables,and process data measured in some cases,such as chemical process are often highly autocorrelated.Most of existing multivariate autocorrelated process monitoring approaches can only detect process shifts,but cannot provide direct information of which variable or variables caused the out of control signal. A novel approach based on support vector machine（ SVM） learning is proposed for monitoring multivariate autocorrelated process mean shifts and identifying the sources of the shifts. Two SVM classifiers are developed,one for process monitoring and the other for fault identification. Moreover,they can be used for online process monitoring and fault identification. Simulation results show that the proposed method has better performance in process monitoring and better accuracy in identifying the sources of the shifts than compared methods.