Unsupervised-Multiscale-Sequential-Partitioning and Multiple-SVDD-Model-Based Process-Monitoring Method for Multiphase Batch Processes

• DOI: 10.1021/acs.iecr.8b02486
• Publication Date: 2018-12-03T06:20:11Z
• Authors: Jianlin Wang; Kepeng Qiu; Weimin Liu; Tao Yu; Liqiang Zhao
• Keywords: Batch processing

For the effective monitoring of batch processes with uneven multiphases, phase partitioning and discriminant analysis are two critical problems. To fully solve these problems, a systematic strategy including fuzzy hybrid is proposed. First, using new unsupervised, multiscale, sequential partition (UMSP), each divided into phases transitions different clustering scales. On this basis, multiple-support-vector-data-description (SVDD) models built for online monitoring, hybrid-discriminant-analysis method then developed fault detection. The effectiveness advantages proposed illustrated 2D, handwritten example fed-batch penicillin-fermentation process.