Synthesis Optimization Of Pipeline Network Based On Petri-Net

Pipelines represent the most reliable and cost-effective way of transporting large quantities of refined petroleum products from major supply sources such as refineries to local distribution centers. It can bring greater economic benefits that optimizing design and scheduling of petrochemical supply chain. Meanwhile, due to the frequent chemical safety accidents in recent years, it urgently needs to develop a set of systematic methods to solve the problems of economic and security simultaneously. Petri-Net is not only a promising modeling tool, but also contains a mathematical operation itself. This paper proposed a Petri-Net superstructure to solve the design, scheduling and safety problems of pipeline network.(1) For the design and scheduling problem involved in the multiproduct pipeline network, it is more complex and time-consuming to synthesize the minimization of pumping costs and pipe fixed charges based on the cost of the piping system. This work presents a new Petri-Net-based discrete-time MINLP formulation for the design&scheduling of multiproduct pipeline systems. The contributions are the following:(ⅰ) Arc-Variable Petri-Net was first used in modeling pipeline network;(ⅱ) the proposed State-Time-Space Matrix Equation derived from Petri-Net contains all mass balance and sequence constraints, and reduces redundant infeasible solutions;(ⅲ) the problem of design and scheduling of the pipeline network was optimized simultaneously.(2) For synthesizing production economy and safety operating procedure, according to the concepts of Petri-Net, the pipeline, equipment and fault topology modules are established, and the superstructure of operating procedure is built up. Then, constrains are determined on the basis of the tokens transportation in Petri-net. Finally, a bi-objective function is set up with the consideration of the balance between the shortest action path and the lowest fault expectation. Furthermore, the satisfying optimization solution in the Pareto front is attained using CPLEX solver in GAMS. An illustrative example is presented to demonstrate the validity and advantages of the proposed approach.