With the development in the understanding of the life sciences, the demand for various functional proteins is constantly growing. Extraction and purification of protein form various natural sources have continued to evolve during the past fifteen years. Liquid-solid circulating fluidized bed (LSCFB) ion exchange system, which is capable of continuous protein recovery from broth solution, has significant advantages over the other ion-exchange adsorption processes. Two separate operations (adsorption and desorption) are carried out simultaneously in the two columns (downcomer and riser) of LSCFB system and the performance of these operations are mutually dependent. Therefore, optimum design and operation of the LSCFB system for the better overall performance is very critical.

Like most real-life optimization problems, the design and operation of LSCFB system for continuous protein recovery is also associated with several important objectives such as production rate and recovery of protein, and ion exchange resin requirements, which need to be optimized simultaneously. Multi-objective optimization of the LSCFB system at both the operation and the design stages were carried out using an experimentally validated mathematical model to determine the range of optimal solutions. Elitist non-dominate sorting genetic algorithm with its jumping gene adaptation (NSGA-II-aJG) was used to solve a number of two- and three- objective function optimization problems. The optimization resulted in Pareto optimal solutions, which provides a broad range of non-dominated solutions due to conflicting behavior of the operating and design parameters on the system performance indicators. Significant improvements were achieved, for example, for the same recovery level, the production rate at optimal operation increased by 33%, using 11% less solids compared to experimental results. In the design stage optimization, the performance of the system was further improved. This multi-objective optimization study is very general and can be easily extended for the improvement of LSCFB in other applications.

Keywords: Liquid-solid circulating fluidized bed, protein recovery, multi-objective optimization, Pareto sets, genetic algorithm.