Model Predictive Quality Control of Batch Processes

This work addresses the problem of driving a batch process to a specified product quality using model predictive control (MPC) with data-driven models. To address the problem of unavailability of online quality measurements, an inferential quality model, which relates the process conditions over the entire batch duration to the final quality, is required for this problem. At a given sampling instant, the accuracy of this type of quality model, however, is sensitive to the prediction of the future (unknown) batch behavior. That is, errors in the predicted future data are propagated to the quality prediction, adding uncertainty to any control action based on the predicted quality. To address this "missing data" problem, we integrate a previously developed data-driven modeling methodology, which combines multiple local linear models with an appropriate weighting function to describe nonlinearities, with the inferential model in a predictive control framework. The key benefit of this approach is that the causality and nonlinear relationship between the future inputs and outputs are accounted for in predicting the final quality, resulting in more effective control action. The efficacy of the proposed predictive control design is illustrated via closed-loop simulations of an industrially relevant nylon-6,6 batch polymerization process.