Scheduling And Operation Optimization Of Multiproduct Pipeline Network With Multi-source And Multi-terminal

Refined product is an indispensable fuel to maintain the normal development of the national economy,society and armed forces,and most of them are transported in long-distance sequence through pipelines,that is,a variety of oil products are transported alternately in batches.To ensure the efficient,safe and economical completion of the oil product pipeline transportation task,it is necessary to quickly formulate a scheduling plan to meet the upstream and downstream demands of the multiproduct pipeline while considering the planning period,the initial state of the pipeline,refinery production capacity,oil injection and delivery demand along the pipeline,and other constraints.On the one hand,with the establishment of the national pipeline network and the interconnection of multiple pipelines,the research object of the scheduling plan has gradually changed from a single pipeline with single-source to a complex pipeline network with multi-source,which further increases the difficulty of establishing and solving the scheduling optimization model of multiproduct pipeline network.The existing research on the scheduling of multi-source and multi-terminal pipeline network is less,and a general and effective model and algorithm have not been formed.On the other hand,the refined oil pipeline network will gradually undergo market-oriented transformation,and the dynamic market environment increases the frequency of planning adjustments.Traditional scheduling optimization methods mostly use branch-and-bound algorithms,which cannot meet the timeliness when solving large-scale pipeline network scheduling plans and led to the poor timeliness of plan adjustment,resulting in accidents such as emergency shutdown of pipelines,an increase in the amount of mixed oil,and the failure of subsequent oil products to be delivered on time.Aiming at the above bottlenecks,this paper takes four types of typical refined oil network topologies as the research objects,and develops the scheduling optimization mathematical model and efficient algorithm suitable for single-source and multi-source pipeline network with multiple terminals,so as to provide suggestions and guidance for the efficient operation and management of multiproduct pipeline.Firstly,for the basic scheduling optimization model of single-source and multi-terminal pipeline,the model is improved from the perspective of the number of time windows affecting the solution efficiency,allowing multiple batches of injection or delivery during one slot,so as to reduce the number of time windows without affecting the solution quality.Secondly,for the structure of single-source and multi-terminal pipeline network,the coupling relationship and key influence mechanism between pipeline flow,node flow,batch migration and node inventory are explored.Based on the above improved scheduling model,an efficient solution algorithm is developed by decoupling the model,combining the two-stage model and backtracking mechanism.Thirdly,aiming at the structure of multi-source and multi-terminal pipeline,the injection constraint of the improved basic scheduling optimization model is improved.By inserting virtual batch and introducing the linear expression method of 0-1variable,the mathematical representation problems of the same oil injection when the upstream pipeline of the node is not shut down and different oil injection when the upstream pipeline of the node is shut down are skillfully solved,and the mixed integer linear programming(MILP)model for batch planning optimization of multi-source and multi-terminal pipeline is completely established.Finally,aiming at the model structure of multi-source and multiterminal pipeline network,it is difficult to realize batch tracking across the transfer node with fixed number due to batch reorganization and injection into the downstream pipeline in the transfer oil depot.Therefore,this paper proposes a dynamic index batch numbering method to realize accurate batch tracking under the transition between intermediate injection source and transfer oil depot.Taking the intermediate oil depots as the boundary,the scheduling problem of pipeline network is decoupled into the batch planning optimization subproblem and the resource allocation optimization subproblem of intermediate oil depots,thus reducing the space complexity of the problem effectively.The single-source and multi-terminal pipeline network based on different topologies are tested respectively.The results show that compared with the traditional basic scheduling optimization model,the improved scheduling optimization model and decomposition algorithm proposed in this paper can shorten the preparation time of scheduling plan by 1-3 orders of magnitude.For multi-source and multi-terminal pipeline network,the model and algorithm proposed in this paper can quickly formulate the optimal scheduling plan,improve the automatic operation and management level of pipe network,and also provide a reference for the automatic preparation of scheduling plan of domestic multi-source and multi sink product oil pipeline network.