The Application of Event-Tree Based Approach in Long-Term Crude Oil Scheduling

LaiXi Zou,ChuFu Li,Ming Zhang,XiaoRong He
DOI: https://doi.org/10.1007/s11426-010-3183-0
2010-01-01
Science China Chemistry
Abstract:This paper addresses the problem of optimal operation in long-term crude oil scheduling, which involves unloading crude oil from vessels, transferring it to charging tanks and feeding it to the distillation units. The application of a new approach for modeling and optimization of long-term crude oil scheduling is presented and the event-tree based modeling method that is very different from mathematical programming is employed. This approach is developed on the basis of natural language modeling and continuous time representation. Event triggered rules, decomposition strategy, depth-first search algorithm and pruning strategy are adopted to improve the efficiency of searching the optimum solution. This approach is successfully applied to an industrial-size problem over a horizon of 4 weeks, involving 7 vessels, 6 storage tanks, 6 charging tanks, 2 crude oil distillation units, and 6 crude oil types. The CPU (AMD 3000+, 2.0GHz) solving time is less than 70 seconds.
What problem does this paper attempt to address?