Soft Sensor Framework Based on Semisupervised Just-in-Time Relevance Vector Regression for Multiphase Batch Processes with Unlabeled Data

Jia

Asia-Pacific J Chem Eng

2023

This paper focuses on product quality control issue of polystyrene polymerization reaction process. A novel tube‐based batch model predictive control (BMPC) strategy based on a data‐driven model is presented, which is inspired by the tube‐based robust model predictive control (MPC) strategy. First, the dynamic behavior of the polystyrene polymerization reaction process is captured with high accuracy by establishing a just‐in‐time learning (JITL) model. Then, the built JITL model is regarded as the nominal system, and a BMPC is designed on the basis of JITL model to obtain the nominal trajectory, which is an integrated control system framework composed of iterative learning control (ILC) and MPC. Meanwhile, another auxiliary MPC (AMPC) is designed to minimize the deviation between the actual trajectory and the nominal trajectory, and the actual tracking errors are limited to the tube invariant set centered on the nominal errors to restrain the performance deterioration of the control system caused by the modeling errors. Finally, the effectiveness of the proposed tube‐based BMPC method is verified by simulating an industrial batch polystyrene polymerization reaction process. The results indicate that the presented control algorithm not only enhances the tracking performance of the system but also provides a more robust system stability than existing methods. This paper provides a new solution to improve the product quality of polystyrene polymerization reaction process.

Section: Just-in-time Learning Modelmentioning

confidence: 99%

Tube‐based batch model predictive control for polystyrene polymerization reaction process

Jia

Asia-Pacific J Chem Eng

2023

“…During the past few years, the JITL method has been extensively used for the design of adaptive controllers, [36][37][38] process monitoring, [39,40] and soft sensors [41,42] for chemical processes. The JITL method uses a set of onlineobtained local models to describe a nonlinear process.…”

Section: Jitl Modelling For Batch Chemical Processesmentioning

confidence: 99%

High‐order iterative learning model predictive control for batch chemical processes

2023

In order to address two‐dimensional (2D) control issue for a class of batch chemical processes, we propose a novel high‐order iterative learning model predictive control (HILMPC) method in this paper. A set of local state‐space models are first constructed to represent the batch chemical processes by adopting the just‐in‐time learning (JITL) technique. Meanwhile, a pre‐clustered strategy is used to lessen the computational burden of the modelling process and improve the modelling efficiency. Then, a two‐stage 2D controller is designed to achieve integrated control by combining high‐order iterative learning control (HILC) on the batch domain with model predictive control (MPC) on the time domain. The resulting HILMPC controller can not only guarantee the convergence of the system on the batch domain, but also guarantee the closed‐loop stability of the system on the time domain. The convergence of the HILMPC method is ensured by rigorous analysis. Two examples are presented in the end to demonstrate that the developed method provides better control performance than its previous counterpart.

“…Their application in biotechnological and pharmaceutical processes is currently limited [ 10 ] due to the more challenging processes involved. This technology, particularly when combined with phase detection algorithms, has only been described in a few publications, for example, to determine the penicillin concentration in a simulated bioprocess [ 19 ]. The broader application of different bioprocesses in a generalist concept in the biotechnology industry, as well as the implementation of robust data-driven phase detection methods that exclude burrs, is still an open issue.…”

Section: Introductionmentioning

confidence: 99%

Generalizability of Soft Sensors for Bioprocesses through Similarity Analysis and Phase-Dependent Recalibration

Siegl

Kämpf

Geier

et al. 2023

Sensors

A soft sensor concept is typically developed and calibrated for individual bioprocesses in a time-consuming manual procedure. Following that, the prediction performance of these soft sensors degrades over time, due to changes in raw materials, biological variability, and modified process strategies. Through automatic adaptation and recalibration, adaptive soft sensor concepts have the potential to generalize soft sensor principles and make them applicable across bioprocesses. In this study, a new generalized adaptation algorithm for soft sensors is developed to provide phase-dependent recalibration of soft sensors based on multiway principal component analysis, a similarity analysis, and robust, generalist phase detection in multiphase bioprocesses. This generalist soft sensor concept was evaluated in two multiphase bioprocesses with various target values, media, and microorganisms. Consequently, the soft sensor concept was tested for biomass prediction in a Pichia pastoris process, and biomass and protein prediction in a Bacillus subtilis process, where the process characteristics (cultivation media and cultivation strategy) were varied. High prediction performance was demonstrated for P. pastoris processes (relative error = 6.9%) as well as B. subtilis processes in two different media during batch and fed-batch phases (relative errors in optimized high-performance medium: biomass prediction = 12.2%, protein prediction = 7.2%; relative errors in standard medium: biomass prediction = 12.8%, protein prediction = 8.8%).