Recurrent Neural Network-Based Model Predictive Control for Continuous Pharmaceutical Manufacturing

Abstract

The pharmaceutical industry has witnessed exponential growth in transforming operations towards continuous manufacturing to effectively achieve increased profitability, reduced waste, and extended product range. Model Predictive Control (MPC) can be applied for enabling this vision, in providing superior regulation of critical quality attributes. For MPC, obtaining a workable model is of fundamental importance, especially in the presence of complex reaction kinetics and process dynamics. Whilst physics-based models are desirable, it is not always practical to obtain one effective and fit-for-purpose model. Instead, within industry, datadriven system-identification approaches have been found to be useful and widely deployed in MPC solutions. In this work, we demonstrated the applicability of Recurrent Neural Networks (RNNs) for MPC applications in continuous pharmaceutical manufacturing. We have shown that RNNs are especially well-suited for modeling dynamical systems due to their mathematical structure and satisfactory closedloop control performance can be yielded for MPC in continuous pharmaceutical manufacturing.

Smart Manufacturing for Smart Cities—Overview, Insights, and Future Directions

“…Datadriven ML approaches, on the other hand, which primarily rely on data rather than the physics and thermodynamics of these processes, offer a more generic and computationally faster approach for sustainable design, process monitoring, quality control, and effective system integration in a manufacturing environment. [37,57] Lv et al [40] sourced data from the ERP system and devised an artificial neural network (ANN) and neighborhood component feature selection-coupled ANN (NCFS-ANN) using a set of 40 input features to predict the scrap rate and material feeding (target labels) during the production of printed circuit boards and observed that the performances of the NCFS-ANN were better than the standalone ANN model. Li et al [41] presented an online monitoring approach to predict the surface roughness of 3D-printed parts during the fused filament fabrication process using the process data sourced from in situ sensors.…”

Effects of Preservation Technology Investment on Waste Generation in a Two-Echelon Supply Chain Model

“…Lee [42] proposes a cost/benefit model to asses and predict the impact of quality investments on returned profit in a multi-level production assembly. A non-linear mathematical model was proposed by Iqbal and Sarkar [43] to give an optimum amount of preservatives to increase the lifetime of products, while Wong et al [44] studied waste reduction in pharmaceutical manufacturing. The study suggests that the preservation-based improved life length of products is associated with the price of the products, as the coordination of certain aspects like costs and quality of products in a supply chain infrastructure is a fundamental parameter in keeping product cost in-control and making a supply chain profitable.…”

Recursive Algorithms for Multivariable Output-Error-Like ARMA Systems

“…System identification is an important branch in the field of modern control and is an important method to establish systematic mathematical models from the combination of observation data and prior knowledge [1][2][3][4][5][6][7][8], and has been applied in many fields for decades, such as controller design [9][10][11][12][13][14][15] and system analysis [16][17][18][19][20]. Parameter identification is an important part of system identification and is to estimate the parameters by using the measurable data [21][22][23][24][25][26][27].…”