(485b) Adjustable Robust Optimization for the Planning Operations of Integrated Refinery-Petrochemical Site Under Demand Uncertainty

Facing the challenges of tight competition and environmental restriction from global market, petrochemical industry has been forced to seek all potential possible opportunities to improve the overall profit. Therefore, instead of considering planning optimization in refinery and petrochemical plant individually, a comprehensive standpoint of global enterprise optimization has received more attention than ever before [Li et al., 2016; Zhao et al., 2017; Uribe-Rodriguez et al., 2020]. The integrated refinery-petrochemical planning optimization provides a new route to achieve a higher profit margin by producing more valuable products, saving operating cost and raw material purchasing. Furthermore, ignoring the uncertainty in practical manufacturing, such as demand and sale price of final products, can reduce the total profit significantly. To improve the profitability of enterprise-wide refinery-petrochemical planning operations and avoid excess conservativeness derived from the worst-case based classical robust optimization, the framework of adjustable robust optimization under dynamic uncertainty set is proposed for a multi-period problem of integrated refinery-petrochemical site under uncertain demands.

Focusing on the operation planning of a large industrial refinery-petrochemical site which is consist of 24 and 15 processing units in each part, respectively, the nonconvex mixed-integer nonlinear programming (MINLP) is formulated. Complex process operation, such as the scheduling of ethylene cracking process and polymerization process are taken into account simultaneously. Binary variables are defined to denote the type of processed crude oil, purchase decision of raw crudes and selection of unit operation modes. Uncertainty lies in demand of final products and corresponding constraints are reformulated using robust counterpart, affinely adjustable robust counterpart, affinely adjustable robust counterpart with dynamic uncertainty set, respectively. The fluctuation and connection in demand of different periods are drawn from historic statistics. The problems are considered in three scenarios with time length varying in 5, 10, 15 periods respectively and solved by commercial solver BARON. Obtaining from the different models, the respective operational strategies are examined in diverse scenarios with random demand of final products. The computational results show that the method of affinely adjustable robust optimization coupled with dynamic set provides a best average performance on these sampling scenarios, following by the affinely adjustable robust optimization alone. The average profit of static robust optimization is the lowest but fluctuation between scenarios is smallest.

References

1. Li, J., Xiao, X., Boukouvala, F., Floudas, C. A., Zhao, B., Du, G., Su, X., & Liu, H., Data-driven mathematical modeling and global optimization framework for entire petrochemical planning operations, AIChE Journal, 62, 3020-3040 (2016)
2. Uribe-Rodriguez, A., Castro, P. M., Gonzalo, G.-G., & Chachuat, B., Global optimization of large-scale MIQCQPs via cluster decomposition: Application to short-term planning of an integrated refinery-petrochemical complex, Computers & Chemical Engineering, 140 (2020)
3. Zhao, H., Ierapetritou, M. G., Shah, N. K., & Rong, G., Integrated model of refining and petrochemical plant for enterprise-wide optimization, Computers & Chemical Engineering, 97, 194-207 (2017)