Inventory Pinch Analysis based Bi-level Heuristic Algorithm for Scheduling of the PVC Production by Calcium Carbide Method

摘要

The scheduling of PVC production by calcium carbide method has attracted increasing concerns with the aggravation of global competition and strict environmental regulation. However, due the resulted scheduling model scale is too large to be solved in reasonable time, there is seldom successful report on plant-wide process scheduling. To accelerate the solving process, inventory pinch analysis based heuristic method is proposed in this paper. Inventory pinch analysis is introduced to divide the whole scheduling horizon into several sub-horizons, and then part of the variables of VCM production process are optimized, which makes the plant-wide scheduling problem easier. Firstly, a VCM demand curve is obtained from the PVC demand profile, and the start-up/shut-down operations of the arc furnaces for sub-horizons divided by the pinch points are determined by the local and small scale optimization. Then, the plant-wide scheduling model with reduced number of binary variables is obtained as the decision variables of start-up/shut-down operations of the arc furnaces are primarily determined. The recursive procedure continues between the two levels until the feasible schedule arrives. At the end of the paper, cases originated from a real world plant exhibit outstanding performance of the proposed algorithm, in which the significantly reduced solution time is required.