Integrated scheduling under uncertainty for polyvinyl chloride processes

Abstract

Polyvinyl chloride (PVC) is a major commercial polymer and is widely used as a raw material in various chemical and petrochemical products. Polyvinyl chloride (PVC) plant has large capacity and complex operating rules. Despite the large capacity and complex operating rules of the PVC plant, current production scheduling is performed by rule of thumb. Therefore, the PVC production schedule needs to be optimized more systematically. Optimizing PVC production will result in cost reductions and profit improvement. A scheduling model and a solving algorithm should be developed so that the PVC plant operates more efficiently. PVC scheduling models are proposed to optimize a production schedule for the entire process of the PVC plant. A basic PVC scheduling model is formulated as mixed integer linear programming (MILP) according to the general operating rules of the PVC plant. In addition, detailed scheduling models are proposed for two PVC processes with different characteristics. In order to solve the PVC scheduling models efficiently, a hybrid method combined with GA for integer programming (IP) and the interior point method for linear programming (LP) is developed. A heuristic technique is also proposed to reduce the number of solving LPs. In order to test the efficiency of the proposed algorithm, the computational time and convergency of the proposed algorithm are compared with those of $GAMS^{TM}$. Based on the proposed models and algorithm, two PVC scheduling S/Ws, WISS (Web-based integrated scheduling system) and PSS (PVC scheduling software), are developed. PSS is applied to two real processes, and the results are analyzed. This study also proposes a new scheduling method that uses virtual demand profiles (VDPs) to handle uncertain intermittent demand. The VDPs are generated by sampling from probabilistic distribution functions, which are built using historical data. Tens of VDPs are used for the production scheduling so that the uncertainty in demand ...