Hybrid Dynamic Optimization Using Artificial Chemical Process: Extended LARES−PR

Abstract

This paper presents a methodology for the dynamic optimization of hybrid processes. The proposed methodology considers the most general class of hybrid dynamic systems consisting of a finite set of dynamic subsystems, in which sequence and duration is not established a priori. The objective is to determine the optimal sequence of dynamic subsystems, the transition times from one subsystem to another, and the optimal profile of the control laws. LARES-PR algorithm for the solution of dynamic optimization problems (Irizarry, R. Chem. Eng. Sci. 2005, 60, 5663-5681) is extended to solve this type of problem. The efficiency and robustness of the method is demonstrated with a set of challenging benchmark problems. The proposed methodology is then applied to a multicomponent nanoparticle alloy synthesis. This prototype model is a very challenging one in which a mechanism switch not changes not only the model structure but also the solution methodology for each mode. For instance, in one phase, the particles grown by coalescence are modeled using discretized population-balance models. When a mechanism switch is induced, a new type of particle is introduced into the system. The subsequent growth of the binary system is modeled using the Monte Carlo method. The proposed methodology can handle this situation effectively. In general, the method is very robust, is simple to implement, and can be used to solve a wide range of hybrid optimization instances with different types of complexities.