Research On Active Disturbance Rejection Control Of VGT-EGR System In Diesel Engine

Diesel engine has the advantages of strong power,good fuel economy and reliability,and is widely used in automobiles,ships,generator sets and other fields.However,the main pollution emissions from diesel engines are nitrogen oxides and particulate matter,which can seriously harm the environment and human health.With increasingly stringent emission regulations and increasing requirements for fuel economy,variable geometry turbocharger(VGT)and exhaust gas recirculation(EGR)technologies have been applied to diesel engines to reduce emissions and improve fuel economy.However,the nonlinear,coupling,complex and uncertain characteristics of VGT and EGR system make it more difficult to control.Active disturbance rejection control is very suitable for VGT and EGR system coordination control due to its independent model and strong anti-disturbance ability.Based on the diesel engine university-enterprise cooperation project of Weichai Power Co.,Ltd.,this paper studies the active disturbance rejection control of diesel engine VGT-EGR system.The main contents of this paper include:Firstly,in order to solve the problem of mutual coupling and uncertainty between VGT and EGR systems,the double-input and double-output diesel engine VGT-EGR system was decomposed into two single-input and single-output systems.The coupling,unmodeled dynamics and the dynamic changes caused by the change of working conditions were estimated and eliminated as the total disturbance by the extended state observer.On this basis,linear active disturbance rejection controller(LADRC)and nonlinear active disturbance rejection controller(NLADRC)are designed respectively,and the two controllers are compared.The comparison results show that NLADRC has better tracking performance,anti-interference performance and robustness.Secondly,a novel nonlinear extended state observer(NLESO)is designed to solve the problem of large number of parameters in traditional NLADRC,which reduces the number of parameters to be set in traditional NLESO from the structure,thus reducing the difficulty of parameter tunning,and its convergence analysis is carried out.The novel NLADRC of diesel engine VGT-EGR system designed based on the novel NLESO was verified in terms of tracking performance,anti-interference performance and robustness,and the effectiveness of the novel NLADRC was verified.Finally,in order to solve the problem that the novel NLADRC parameters are coupled with each other and it is difficult to set manually,the NLADRC-CARLA parameter tunning algorithm is proposed based on the continuous action reinforcement learning automata(CARLA),which can automatically learn parameters satisfying the control performance according to the objective function.The novel NLADRC parameters of diesel engine VGT-EGR system are learned by using this algorithm.Compared with the control effect of the novel NLADRC with manual setting parameters,the parameters obtained by learning can optimize the control performance.Moreover,the novel NLADRC with learning parameters has good tracking effect on different types of expected signals,which verifies the effectiveness of NLADRC-CARLA parameter tunning algorithm.