Hybrid modeling based on mechanistic and data-driven approaches for cane sugar crystallization

摘要

Cane sugar crystallization is a non-linear and non-stationary process. Some key parameters which significantly affect the process and the quality of the product such as crystal content and main crystal size distribution parameters (average size and coefficient of variation) still cannot be directly measured online by existing sensors. In this research, we developed a hybrid model which is combination of a data-driven model and a mechanistic model. Based on the previously developed data-driven modeling, a crystallization kinetic mechanism is established from the point of view of energy and mass conservations population balance. By expressing the nucleation and growth rate of the crystallization and aggregation, combining the data-driven model, not only the mother liquor supersaturation and mother liquor purity can be successfully predicted, but also the crystal content and main crystal size distribution parameters of the crystallization process can be accurately predicted. Based on this hybrid model, an intelligent integrated measurement and control system for the process of cane sugar crystallization is developed and has been successfully applied in sugar production. The experimental results show that the hybrid model based on mechanism and data-driven has high accuracy and can be used in actual industrial production process.