Multi-Objective Optimization of Fluidized Bed Hydrodynamics via Spatial Modulation of Inlet Gas Velocities

• Distributed inlet actuation enables practical, real-time structuring of fluidized beds • Evolutionary algorithms steer inlet gas velocities toward predefined hydrodynamic targets • Actuator-aware constraint handling guarantees feasible flows under hardware limits • Multi-objective optimization reveals trade-offs between bed expansion and bubble size • R-NSGA-II achieves up to 200% bed expansion with only marginal bubble growth We present a method for spatially steering the inlet gas velocity in bubbling gas-solid fluidized beds using evolutionary algorithms (EAs) to achieve predefined hydrodynamic targets related to bubble dynamics and bed expansion. A hardware-in-the-loop framework couples real-time bubble detection with a synchronized hardware-software loop, enabling the closed-loop actuation of spatially and temporally modulated inlet velocity profiles. The bubble-detection algorithm extracts bubble properties from each frame in real time, while the optimization loop manipulates inlet velocities and evaluates hydrodynamically meaningful fitness functions for the optimization process. Candidate solutions are represented by sinusoidal inlet velocity distributions around a fixed operating point, and a repair step is introduced to enforce smoothness, mass-flow conservation, and hardware constraints. A systematic hyperparameter study is performed to identify robust EA settings for the targeted reduction of bubble size. Furthermore, multi-objective optimization techniques (NSGA-II and R-NSGA-II) are employed to jointly enhance bed expansion and decrease the mean bubble diameter. Experiments demonstrate that R-NSGA-II achieves up to a 200% increase in bed expansion relative to uniform flow, with only a marginal increase in bubble diameter.