A Hierarchical Statistical Process Monitoring Strategy for Multivariable Multi-rate Industrial Processes

A hierarchical statistical process monitoring strategy is proposed for the industrial processes with multivariable multi-rate sampled measurements. By making full use of multi-rate measurements, two-level models are adopted where the sub-PCA models are built on high-rate measurements to ensure timely abnormality detection and the super Multi-block PCA model is built on the lifted process measurements to monitor the overall operating performance. The ability of the proposed strategy is demonstrated with the benchmark TE process.