Reachability Analysis Of Starting State Of Short-Term Scheduling For Crude Oil Operations In Refinery With Two Transportation Pipelines

Crude oil refinery belongs to a typical process industry. It is not only the pillar industry of national economy, but also the main driving force in the development of other industries. As the refinery is a hybrid dynamic system with both continuous and discrete events, it causes high energy consumption, high coupling degree and high cost. With the increasing market globalization, the development of crude oil processing technology and control methods provides an effective way for enterprises to stay competitive.Production plan of refinery contains long-term scheduling and short-term schedu ling. Short-term scheduling is generally a production plan for a day or a week and long-term scheduling is a production plan for a month and a quarter. There are a lot of achievements of the research for long-term scheduling. The linear-programming-based software tools for making a long-term scheduling have been successfully applied in the practice. Due to the hybrid property of the oil refinery, the research for short-term scheduling of the refinery is still at the starting stage. Because the hybrid system contains both continuous and discrete variables, the short-term scheduling problem for refinery is extremely complicated. Also the oil refinery process contains some resource constraints and process constraints. However, ignoring any constraints can result in an infeasible solution. There are no effective methods to solve this problem and the short-term scheduling is still done manually. However, it is time-consuming. Meanwhile, it is difficult to guarantee the accuracy. Therefore, there is still much room for improvement.Because of the complexity of this problem and different kinds of constrains, it is very difficult to solve this problem by the mathematical programming method. Petri net is a kind of modeling and analysis tool, and Petri net has unique advantages over the traditional modeling methods. By comparing with the traditional modeling methods, Petri nets can accurately describe the process of oil refinery and avoid the complexity of the calculation. Therefore, Petri net is an effective way to model oil refinery process.Based on developed colored and timed Petri net, these problems can be decomposed into some sub-problem and the feasible initial state of short-term scheduling of crude oil operations in refinery can be obtained. This paper focuses on how the system can switch from a steady state to the required initial state. With the control theory, effective methods are proposed to switch different states such that the required initial state can be reached. The proposed techniques are applicable to practice. Finally, a real industrial case is given to illustrate the application of the proposed method.