Research On Washing Tower Temperature Control System Of Shale Oil Refining Factory

With the traditional resources are exhausted increasingly, shale oil refining industry entered a golden period of development. In shale oil refining process, one of the most important equipment of oil recovery system is washing tower. The main role of washing tower is to cool and wash hydrocarbon vapors, whose effect will directly affects the product quality and yield of the whole chemical process.Therefore, the reserach on washing tower temperature control system has practical significance.Based on ATP project of the shale oil refinery, this paper expatiates on the inner structure and the industrial art of the hydrocarbon vapors washing tower. By analyzing the process of the hydrocarbon vapors washing tower control system, the paper selects the outlet temperature of hydrocarbon vapors and the outlet temperature of bottom oil as the controlled variable, heavy oil flow and the circular bottom oil flow as control variables.Meanwhile, the paper simplifies the washing tower model, and combined with the assumed conditions, the paper builds an mathematical model based on the energy conservation equation and mass conservation equation. By analyzing the coupling properties of washing tower temperature control system, the paper come to a conclusion that the hydrocarbon vapors temperature control process and the bottom oil temperature control process have strong coupling degree. Thus, it’s necessary to design decoupler. For the strong coupling characteristic of washing tower temperature control process, the paper designs respective decoupler according to the design principle of feed-forward compensation decoupling control and unit matrix decoupling control. In view of the controlled object model’s time-variable and uncertainty, the single neural PID decoupling controller based on genetic algorithm is used to replace the traditional decoupling controller, so as to maintain the neural network adaptive and learning ability, and to avoid the neural network’sensibility to the initial value by using genetic algorithm to optimize the parameters of the single neural PID.Finally, compared to the performance of the traditonal decoupling control,the single neural PID decoupling control not only has smaller overshoot,shorter response time and stronger anti-disturbance ability but also can adapt to parameter changes of the controlled object better, that is, it is can achieve better control effect.