Modeling Of Dual-pipeline Crude Oil Transportation Systems In Refinery By Controlled-Colored-Timed Petri Net And Short-term Schedulability Analysis

Oil refinery industry plays an important role in economy for daily use in oil consumption or for national strategy target. However, our oil refinery industries are chaterized by high cost, high energy consumption, low efficiency and the low capability of process control, which results in lacking of competitive advantages.As one of the significant industries, oil refinery industry should untilize the integrated automation control technology to optimize its production. Generally speaking, there are three production levels in oil refining-long-term planning, short-term scheduling and process control. Up to now, as for process control and long-term planning, there are advanced unit process control system and Linear Programming-based commercial software to optimize long-term planning, resulting in great amout of profits in the plant. In terms of short-term scheduling, the complexity of procedures and large amounts of variables are the key problems to tackle. Thereforce, it is done manually by using a predict analysis tool by simulation.However, it seriously depends on workers’experience and spends too much time and lacks flexibility of addressing emergencies.In order to optimize oil scheduling, this thesis firstly models the system by a Colored Timed Petri Net (CCTPN) and then carries out the schedulability analysis. An oil refining process is characterized as a hybrid system containing continuous and discrete-event variables. Hence, its scheduling problem is combinatorial. Up to now, modeling of oil refining system is one of the most difficult problems in hybrid system modeling. For the convenience of modeling, CCTPN is applied to model crude oil operation. This paper extends the CCTPN to model the dual-pipeline crude oil transportation (high/low fusion point oil) system. Many practical production constraints, especially constraints for high fusion point oil transportation, are analyzed. In this thesis, crude oil operation is divided into four subproblems in a hierarchical way, including oil tankers arriving, oil unloading, oil transportation and distiller feeding. With a modular modeling approach, a Petri net model is developed such that the system is modeled in a hierarchical way, which makes the model compact and the system’s behavior can be analyzed effectively. Moreover, it can be used as an assistant tool for verification of the logistics information.When making the schedulability analysis, obviously, the tasks to be scheduled are unknown. Thus, heuristic and intelligence algorithms cannot be applied directly. Through a hierarchical architecture, oil refinery problem can be tackled in two phases-finding a realizable refining scheduling to optimize combinational targets and making a detailed schedule. A detailed schedule can be easily obtained, given a realizable refining schedule. By the CCTPN model, this thesis conducts the schedulability analysis of the transportation of high fusion point oil. Considering the high cost and complicated constraints, the key to schedule high fusion point oil lies in how to transport as much oil as possible during the scheduling horizon and the surplus high fusion point oil should be charged to which tank.Consequently, the scheduling problem of dual-pipeline crude oil transportation system is effectively addressed. In Chapter Four, from special cases to general ones, such as four charging tanks for two distillers, five charging tanks for two distillers, we present the safe boundary principles to operate the system. Chapter Five observes initial states by the safe boundary conditions to obtaine a detail schedule, so as to verify the above research works in this thesis.