Cycle design and surrogate-based multi-objective optimisation of magnetic induction swing adsorption for electrified post-combustion CO2 capture

Magnetic Induction Swing Adsorption (MISA) regenerates a fixed bed adsorbent by volumetric, electrically driven heating, bypassing the gas-to-solid thermal resistance of conventional temperature swing adsorption. Four MISA cycle configurations on Fe 3 O 4 @HKUST-1 were designed for post-combustion CO 2 capture, ranging from a three-step baseline to a six-step cycle with dual-stage inductive heating and two recycle loops (5-12 design variables). Each configuration was simulated with a rigorous fixed-bed model resolving coupled gas, solid, and wall energy balances together with multicomponent mass-transfer kinetics and adsorption isotherms. Multi-objective optimisation was made viable by coupling a Gaussian Process surrogate, refined through active learning, to Non-dominated Sorting Genetic Algorithm II and Multi-Objective Particle Swarm Optimisation. The six-step configuration was the only cycle architecture that simultaneously met Carbon Capture and Utilisation (purity ≥ 90%, recovery ≥ 30%) and Carbon Capture and Storage (purity ≥ 95%, recovery ≥ 90%) specifications at near-atmospheric pressure without vacuum assistance. Specific thermal energy was dominated by sensible heating of the bed and wall and governed by low CO 2 working capacity of Fe 3 O 4 @HKUST-1 rather than by the induction-heating mechanism itself.