A methodological proposal using simulation-based digital twins for closed-loop operating-point transitions in complex chemical processes

The chemical industry must adapt production to market fluctuations, demanding operating-point changes that preserve stability, product quality, and operational feasibility despite strong inter-unit interactions and process time delays. This work presents a sequential methodological proposal that integrates steady-state and dynamic simulation-based digital twins to link economic optimization with dynamic evaluation and support feasible closed-loop operating-point transitions in continuous chemical processes with interconnected units and recycle streams. The methodology articulates a cosimulation environment in which the rigorous dynamic process model, operating under an active regulatory control layer, is coupled with a trajectory controller designed on a surrogate model that captures the closed-loop plant dynamics. The methodological proposal was demonstrated on the purification of lactic acid by catalytic distillation, an intensified process characterized by strong interactions among columns and recycle streams. Transitions across purities of 80, 88, and 99 mol% were carried out, achieving offset-free responses and settling times consistent with the column dynamics. Complementary robustness tests under feed disturbances and measurement noise confirmed that the control architecture maintains its performance under operationally representative conditions. These results support the effectiveness of the proposed sequence for executing operating-point changes in the evaluated system and suggest its potential applicability to other complex process configurations where similar hierarchical control requirements arise.