Title:
Optimization of batch crystallization processes by computational efficient frameworks
Optimization of batch crystallization processes by computational efficient frameworks
Time:
10/23 (Sat.) 7 pm PDT, 8 pm MDT, 9 pm CDT, 10 pm EDT
10/24 (Sun.) 4 am CEST, 10 am Taiwan
Time zone conversion tool
10/23 (Sat.) 7 pm PDT, 8 pm MDT, 9 pm CDT, 10 pm EDT
10/24 (Sun.) 4 am CEST, 10 am Taiwan
Time zone conversion tool
Keywords:
Chemical Engineering, Process Systems Engineering, optimal control theory, Pontryagin's minimum principle, model-based control, population balance modeling
Chemical Engineering, Process Systems Engineering, optimal control theory, Pontryagin's minimum principle, model-based control, population balance modeling
Abstract:
With modeling of evolution of crystal size distribution by population balance equations (PBEs) and specific control performance index, model-based control of batch crystallization processes has been extensively studied over the past decades. However, profiles of the optimal control input (e.g., temperature trajectory) are apparently dependent on the single objective function selected, indicating existence of strong trade-off between them. In this research, a framework consisting of a coordinate transformation of the PBEs and optimal control theory is developed to analyze the trade-off quantitatively. The coordinate transformation helps to reduce computational burden in solving the PBEs (in the form of partial differential equations) and optimal control theory provides necessary condition of optimality to relevant multi-objective optimization problems. By further extend the framework, optimization of complex batch operation such as growth-dissolution cycling process can be also conducted with low computational cost.
With modeling of evolution of crystal size distribution by population balance equations (PBEs) and specific control performance index, model-based control of batch crystallization processes has been extensively studied over the past decades. However, profiles of the optimal control input (e.g., temperature trajectory) are apparently dependent on the single objective function selected, indicating existence of strong trade-off between them. In this research, a framework consisting of a coordinate transformation of the PBEs and optimal control theory is developed to analyze the trade-off quantitatively. The coordinate transformation helps to reduce computational burden in solving the PBEs (in the form of partial differential equations) and optimal control theory provides necessary condition of optimality to relevant multi-objective optimization problems. By further extend the framework, optimization of complex batch operation such as growth-dissolution cycling process can be also conducted with low computational cost.
