Multi-objective Optimization Of Short-term Crude Oil Scheduling In A Refinery With Dual Pipelines

As the global ecological environment pollution becomes increasingly serious,it is crucial to improve the economic and social benefits of enterprises by reducing the cost of refinery operation and energy consumption and optimizing production plans.Based on the principle of control theory,the short-term crude oil scheduling problem in a refinery with dual pipelines can be divided into two layers.The upper level refining planning problem has been well solved by linear programming method.However,due to the complex resource and operation constraints in the lower level,it is difficult to guarantee the feasibility of the scheduling plan.In this thesis,the detailed scheduling optimization problem in the lower level is studied under the premise of a given upper refining plan.The details are as follows:First of all,considering the influence of crude oil flow rate on energy consumption,a short-term crude oil scheduling problem for refineries with single pipeline is studied.In order to solve it by using heuristic algorithm and intelligent search optimization algorithm,firstly,the problem is transformed into a charging tank assignment one,and a backtracking search algorithm is proposed to ensure the feasibility of the generated detailed scheduling plan.Then,an improved version of NSGA-III algorithm is proposed.Combined with the characteristics of proposed short-term crude oil scheduling problem,a ternary chromosome structure based on real number coding is put forward,and the traditional crossover and mutation operators are improved.In order to verify the effectiveness of the improved algorithm,through an industrial example,it is compared with a variety of representative multi-objective optimization algorithms.Then,the short-term crude oil scheduling problem for refineries with dual pipeline is studied,and some key issues are analyzed,including how to calculate the volume of crude oil in one transportation process and how to ensure that the high melting point crude oil pipeline can always obtain empty charging tanks during the schedule horizon.At the same time,based on the theoretical analysis of the above key issues,two different heuristic strategies,including high melting point crude oil pipeline first strategy(HS)and parallel competition strategy(PS),are proposed to guide the solution generation process and ensure that the generated scheduling plans meet the operation constraint that once the high melting point crude oil pipeline starts,it cannot stop halfway.In addition,the periodic maintenance of charging tanks is considered.In the actual refining production,in order to improve the production efficiency,distillation tower should keep on refining during the scheduling horizon.Eventually,the problem is modeled as a constrained multi-objective optimization one.By using the search advantages of these kind of algorithms,feasible scheduling plans can be found.Then,a variable length chromosome coding structure and a decoding strategy to determine the selection scheme of a charging tank and distillation tower through a real coding bit are proposed,which effectively reduces the length of chromosome and improves the efficiency of the algorithm.Aiming at the problem of insufficient selection pressure of traditional NSGA-II-CDP algorithm during solving high-dimensional multi-objective optimization problems,an adaptive enhanced pressure algorithm NSGA-II-APE is proposed.Finally,combined with an industrial example,the performance of the two heuristic strategies is analyzed and compared,and the performance of NSGA-II-APE algorithm is verified by comparing with other 11 excellent constrained multi-objective optimization algorithms.