Performance Related Information-Based Operating Performance Assessment of Multiphase Batch Processes

To pursue high comprehensive economic benefit, the online operating performance assessment becomes more and more critical for industrial processes. In this study, a novel performance related information-based operating performance assessment method is proposed for multiphase batch processes. To describe the effects of the process characteristics of different phases on the operating performance of the batch process, the assessment models are established based on total projection to latent structures (T-PLS) for each phase. Then the similarity between the performance related information of the online data and the assessment model of the corresponding phase is calculated and used to define the assessment index for online assessment. When the operating performance is nonoptimal, the responsible variables can be identified by the variable contribution-based nonoptimal cause identification method, which does not need any help of the expert knowledge. Finally, the effectiveness of the proposed assessment method is demonstrated through the fed-batch penicillin fermentation process.