Regulation of the Emulsion Particle Size Distribution to an Optimal Trajectory Using Partial Least Squares Model-Based Predictive Control

Feedback Control of Crystal Size Distribution for Cooling Batch Crystallization Using Deep Learning-Based Image Analysis

The shape of the crystal size distribution directly determines the quality of crystal products. It is often assumed that distributional properties of crystal size conform to the Gaussian distribution or the log normal distribution. The mean and variance or relative crystal number are widely adopted to describe the crystal size distribution and taken as the control objectives. Therefore, the resulting control methods have difficulties in controlling the crystal size distribution with a general shape. In this article, a novel feedback control system of crystal size distribution based on image analysis is designed for the effective control of crystal size distribution with a general shape. First, a deep learning network-based image analysis method is adopted and implemented to extract the crystal size distribution. Second, the crystal size distribution is approximated by a radial basis function neural network. Consequently, a feedback controller is designed and the tracking control of the target crystal size distribution is finally realized. The results of crystallization experiments demonstrate the effectiveness of the proposed method.