Design And Development Of The Natural Regulator Skid Pressure Intelligent Control System

The natural regulator skid control system is one of the important components of natural gas distribution station.Its outlet pressure has a direct impact on the stable operation of distribution station and safety of downstream users.Due to the long-term work of the pressure regulator skid under unsteady conditions,PID controller has the problems of low control precision and poor parameter setting.At the same time,there are some non-linear characteristic such as continuous dead zone,which cause frequent movement of valve and unstable of pressure.Therefore,this thesis deeply studies pressure control methods under different conditions,and completes the following work.Firstly,due to the wide pressure fluctuation and untimely control,the regulator skid pressure control system has non-linear feature and unsuitable PID parameters.An improved RBF-PID pressure controller is proposed.In order to improve the transient response performance of the controller,the PID parameters are adjusted online through using information of system.In order to improve the response speed,using L-M algorithm to replace the traditional gradient descent method.An adaptive variable step strategy is proposed to improve the adjustment efficiency.Meanwhile,an online tuning discriminant condition is proposed to improve the operation efficiency of the program.Simulation results showed the validity of the method.At the same time,in order to accelerate the convergence speed of online tuning,an PID parameter initialization scheme based on multi-objective optimization and multi-attribute decision is proposed.This scheme uses NSGA ? algorithm to get Pareto optimal solution set,then the appropriate PID parameters as online control initial parameters are selected through the multiple-attribute decision method.Finally,RBF_PID controller initial parameters software based on Matlab software is developed.Secondly,aiming at the work conditions that gas fluctuations is serious and valve has serious dead zone,a Human-simulated pressure intelligent system based on Human-simulated intelligent control ideology is designed in this thesis which includes interference controller,PI controller and Human-simulated intelligent controller.In this system,the interference controller uses the feedforward control algorithm to compensate the traffic interference,PI controller adjusts the pressure,and for the Human-simulated intelligent controller,it can realize dynamic switching control based on the operating state of the system and it also designs an unequal amplitude switch strategies to overcome the influence of dead zones and improve the real-time performance of the control.Simulation result and field application show that the control system can not only improve control precision but also effectively reduce the switching frequency of the valve.Finally,based on the Visual Studio 2010 software platform,the regulator skid pressure controller software was developed.The software is designed based on modular principles and contains the communication module,human-computer interaction module and control module.The functions of system setting can be realized by the human-computer interaction module;the communication between the software and PLC can be realized by the communication module,and the function of pressure control and adaptive switching control can be realized by the control module.The results of system test shows that the controller software owns the merits of comprehensive functions,stable performance.