Evolutionary Algorithms for Spatial Control of Gas-Solid Fluidized Beds

Gas-solid fluidized beds are used in industry for fluid-solid processing, e.g., coating and mixing of particles. Fluidized beds are operated by driving a fluid flow through a collector of particles, leading to a chaotic motion of these particles. A control of the spatio-temporal variation of the fluid velocity at the inlet of the fluidized bed can achieve a prescribed pattern (the objective), thus controlling the fluidization. Our study proposes a repair algorithm to account for the suitable constraints for using hardware-in-the-loop. We evaluate the work for a single objective optimization, i.e., to maximize the fluidized bed expansion. Due to the hardware-in-the-loop feature, we focus on initialization methods to achieve a faster convergence towards the objective. We show that using evolutionary algorithms to control the inlet velocity profile in a fluidized bed using a suitable initialization approach, a control for a specific objective can be achieved.