Industrial batch process monitoring with limited data. (May 2019)

• Record Type: Journal Article
• Authors: Tulsyan, Aditya; Garvin, Christopher; Undey, Cenk
• Is Part Of: Journal of process control. Volume 77(2019)
• Page Start: 114
• Page End: 133
• Publication Date: 2019-05
• Subjects: Biopharmaceutical manufacturing -- Real-time multivariate process monitoring -- Machine-learning -- Bayesian methods; Process control -- Periodicals; Fabrication -- Contrôle -- Périodiques; Process control; Periodicals; Electronic journals; 660.281
• DOI: 10.1016/j.jprocont.2019.03.002
• Languages: English
• ISSNs: 0959-1524

Abstract

Highlights: Considers a problem of batch process monitoring with small data. Addresses the small data problem by generating in silico batch data. Proposes a machine-learning method to generate in silico batch data. The efficacy is demonstrated on industrial biopharmaceutical systems. Abstract: This article addresses the problem of real-time statistical batch process monitoring (BPM) for processes with limited production history; herein, referred to as the 'Low-N' problem. The Low-N problem is a longstanding, industry-wide problem in biopharmaceutical manufacturing that challenges the theoretical foundations and practical applicability of the existing BPM platform. In this article, we propose an approach to transition from a Low-N scenario to a Large-N scenario by generating an arbitrarily large number of insilico batch data sets. The proposed method is a combination of hardware exploitation and algorithm development. To this effect, we propose a block-learning method for a Bayesian non-parametric model of a batch process, and then use probabilistic programming to generate an arbitrarily large number of dynamic insilico campaign data sets. The proposed solution not only alleviates the monitoring issues associated with a Low-N scenario, it is also compatible with the industrial BPM framework. To the best of authors' knowledge, this is the first article that describes a systematic approach to address the small data problem using the tools for large data sets. The efficacy of the proposed solution is elucidated on an industrial biopharmaceutical process.