Efficient decomposition strategy for scheduling of multistage production system and combined heat and power

• Published in: Computers & chemical engineering Vol. 133; p. 106634
• Main Authors: Zulkafli, Nur I., Hanak, Dawid P., Kopanos, Georgios M.
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 02.02.2020
• Subjects: Combined heat and power; Decomposition strategy; Large-scale optimization; Mixed-integer linear programming; Scheduling; Large-scale optimization; Scheduling; Combined heat and power; Decomposition strategy; Mixed-integer linear programming
• ISSN: 0098-1354; 1873-4375
• DOI: 10.1016/j.compchemeng.2019.106634

•A three-stage decomposition strategy is introduced as an intermediary approach.•Decomposition strategy achieves optimal solutions four times faster than the integrated approach.•Operational and cleaning schedule for CHP and multistage production system is presented.•Potential reductions in emissions for imposing minimum possible emission cap. An integrated approach, which optimizes the production and utility system simultaneously, is commonly used to arrive at an optimal scheduling solution. For the large-scale complex industrial systems, this approach requires extensive computational time. The main aim is to investigate the capability of the proposed decomposition strategy to find optimal or near-optimal solutions with relatively low computational effort. The computational results showed that the proposed three-stage decomposition strategy achieves the optimal solutions four times faster than the integrated approach. Also, a sensitivity analysis of emissions cap on the total cost was performed. It was shown that the minimum possible emissions cap (365 t/h) resulted in a 1.2% reduction in total emissions. Overall, the proposed mixed-integer programming (MIP)-based decomposition strategy could be used as an intermediary approach that combines the benefits of the sequential and integrated approach in terms of faster computational time and higher productivity of the overall system. [Display omitted]