多采样率主元分析的过程故障检测

针对多采样率过程监测问题,提出了一种基于多采样率主元分析的故障检测方法.该方法构建了一种重新采样机制,直接利用多采样率数据计算模型中的协方差矩阵,充分利用了样本中的大量不完整数据信息,减小了多采样率数据带来的偏差,给出了离线建模和在线故障检测算法.分别在数值平台和Tennessee Eastman(TE)工业平台进行了仿真分析.仿真结果表明,所提出的方法更适合多采样率过程的故障检测,效果良好.

Multi-Rate Principle Component Analysis for Process Monitoring

Abstract： To monitor multi-rate processes, a multi-rate principle components analysis algorithm was proposed in which the covariance matrix was calculated using the incomplete multi-rate data samples. To avoid the bias of the covariance matrix, the resampling method was adopted. Besides, the offline modeling strategy and online monitoring strategy were proposed. Then two case studies on both numerical and Tennessee Eastman process (TEP) simulation was given to prove the effectiveness of the proposed algorithm compared to other methods. The result shows that the proposed method has a better performance in multi rate process monitoring.