Optimal cyclic scheduling of ethylene plants in backup furnace mode with separation capacity constraints


Ethylene plants (EPs) are a critical output source of ethylene. The cyclic scheduling optimization is crucial to increase the profitability of an EP. The EP comprises the ethylene cracking furnace system (ECFS), quenching, compression, separation sections, etc., which culminates in a large-scale optimization problem. In addition, the performance-decaying and semi-continuous characteristics of the ECFS inevitably lead to fluctuations in product throughput, resulting in an unstable production process and product quality decline. To address this issue, we propose a mixed-integer quadratic programming (MIQP) model for the cyclic scheduling of the EP in backup furnace mode with separation capacity constraints. This novel model can reduce fluctuations in product throughput by running a backup furnace when one furnace is shut down for decoking. Besides, the separation capacity constraints ensure that high-quality products can be separated from the cracked gas downstream. Moreover, the reformulated normalized multiparametric disaggregation technique (RNMDT), which relies on convex envelopes, is employed to relax the proposed MIQP model into a mixed-integer linear programming (MILP) model by partitioning the domain of the variables in the quadratic terms. Subsequently, an optimization framework based on the RNMDT is established to enhance the solution efficiency of the original problem. Finally, a case study from the real world reveals the effectiveness of the proposed method, which can provide instructive optimal cyclic scheduling for the EP.

Computers and Chemical Engineering