Process Control

Abstract

This chapter discusses the control aspects of continuous crystallization processes. Common control objectives for continuous crystallization are related to crystal product quality, process stabilization, economic performance, and environmental impact. Supersaturation is often used as controlled variable to obtain desirable crystal quality attributes, although direct approaches with a crystal quality attribute as controlled variable have also been developed. Sensors to measure crystal quality attributes or supersaturation in situ are readily available, which makes the application of automated feedback control loops attractive. A mixed-suspension mixed-product-removal crystallizer has limited options for process actuation unless fines dissolution is employed. Novel plug-flow crystallizers allow for the adoption of different control strategies (e.g., controlled cooling profiles with seeding). Model-based controllers in combination with state observers can handle time-varying model uncertainty, input constraints, sensor and actuator faults and asynchronous measurements, whereas model-predictive control has the unique capability to enforce multiple process constraints and is most effective when dealing with complicated interactions between multiple inputs and outputs. State observers can also be used to design effective filters for actuator fault detection. The rapid development of dynamic process models, advanced analytical techniques and improved numerical methods are main drivers of the current trend towards model-based control strategies for continuous crystallization.

Data‐driven identification of crystallization kinetics

“…The performance of the SINDy method is first characterized using simulation data for crystallization in an MSMPR crystallizer. The process dynamics is described by the following population balance equation and solute material balance when assuming the crystallizer is well‐mixed, and operates at a constant temperature and constant volume: 35,36 where is the number density distribution of the crystals in the crystallizer, is a characteristic length of the crystals, denotes time, is the growth rate of crystals and …”

Optimal trajectory tracking control of batch crystallization process based on reinforcement learning

“…• The control of the particle size distribution, crystal habit and crystal purity are crucial in most crystallization processes to meet the targeted critical quality attributes of the final product [1]. • This work tries to address the increasing demand for more advanced, versatile, robust and cost-effective control technologies for crystallization processes [2]. • Reinforcement Learning (RL), has gained a lot of interest for process control and optimization, while positively impacting both research and industries [3].…”