Refinery operations planning under uncertainty

Abstract

Nowadays, the petroleum refinery business is very competitive resulted in a low economic margin. In this situation, refinery planning becomes a very important tool as it can bring all potential opportunities to push the economic margin to the maximum limit. However, most formulations presented are based on nominal parameter values without considering the uncertainty. In reality, the deterministic planning obtained may become unfeasible. Consequently, this study proposed a model for refinery operations under uncertainty based on two-stage, stochastic. First, a deterministic model was developed for decision of crude oil purchased in order to attain the product specifications and demands. Then, the uncertainty in demand and price of products was introduced. The proposed objective function was based on optimizing the profit by maximizing the product sale and minimizing the crude oil cost, inventory cost, storage cost, and lost demand volume. A stochastic formulation was then developed to perform the financial risk management. Sampling algorithm was used to find the optimal solution and alternative plan that reduced risk. The model was tested on the simplified process of Bangchak Petrolum Public Company Limted. The optimization results from the deterministic and stochastic models were compared. The results show that the stochastic model can predict higher expected profit and lower risk compared to the deterministic model.