Integrated molecular riser-regenerator modeling via Radau collocation for FCC operational optimization

• Integrated riser–regenerator FCC model with endogenous heat coupling. • Radau collocation converts DAEs to an NLP in CasADi/IPOPT. • Plant-calibrated kinetics with bounded, thermodynamically consistent parameters. • Profit optimization via feed preheat, air rate and heat extraction. • Industrial validation for yields, flue gas and temperature profiles. Fluid catalytic cracking (FCC) operation depends on tight heat coupling between an endothermic riser and an exothermic regenerator. Decoupled models often mispredict feasible temperatures and product trade-offs. This study develops an integrated framework that couples a 23-lump riser kinetics, derived from Molecular-Type Homologous Series (MTHS), with a four-reaction regenerator to compute flue-gas composition and heat release. Unit-level heat and mass balances are imposed endogenously so that catalyst circulation and feasible operating envelopes arise from the model solution. The steady DAE system is transcribed to a sparse nonlinear program using Radau IIA collocation and solved in CasADi/Ipopt. Kinetic parameters and reaction enthalpies are regressed to industrial data under physical bounds. The validated model reproduces major product yields, temperature profiles, and flue-gas compositions with small errors. Against plant data, absolute deviations are within 0.4 wt% and the key product yields are within ±1% relative error, while minor products (e.g., slurry/dry gas) show larger deviations. The operational optimization in the case studies shows up to 7.76% profit increase by coordinated adjustment of air rate, feed preheat, and heat extraction, demonstrating that the framework offers a mechanistically interpretable, computationally robust tool for decision support in industrial FCC units.