Novel flexible model predictive control strategy to improve the yield of light products in fluid catalytic cracking unit

Due to the dynamic characteristics of the fluid catalytic cracking (FCC) reactor-regenerator system, the yield of light products would decrease with the conventional reaction temperature control strategy when the regenerator temperature increases. To mitigate this issue, a novel flexible model predictive control (MPC) strategy is proposed. Precisely, this MPC strategy integrates innovative reaction heat control, reaction temperature control, and a decision-making switch to ensure safe operation. In addition, the Conv_Transformer algorithm is employed to capture the dynamic behavior of the reactor-regenerator system in a noisy environment, facilitating the construction of a predictive model for MPC. The effectiveness of proposed MPC strategy is validated through simulations based on a dynamic model of the FCC reactor-regenerator system. Simulation results indicate that this approach achieves a higher yield of light products compared to conventional temperature control strategy when perturbations occur, demonstrating its potential to enhance the unit’s profit.