Receding horizon control of LPGBE gas mass flow based on neural network

In order to improve fuel economy and emission of liquefied petroleum gas bus in driving condition, this paper researches a receding horizon control method of gas mass flow for liquefied petroleum gas bus engine based on neural network. Firstly, the relation between the valve opening angle, gas pressure and gas mass flow of liquefied petroleum gas bus engine is according to gas mass flow dynamics. On this basis, gas mass flow of liquefied petroleum gas bus engine is estimated by the neural network identification. Then the minimum cost function of receding horizon control is established according to gas mass flow and engine speed, which can control gas valve opening angle and minimize the difference between estimated gas mass flow and expected one. Finally, taking YC6112LQZ of liquefied petroleum gas bus engine as an object of study, receding horizon control system of engine gas mass flow is designed. Simulation of receding horizon control system for engine gas mass flow is performed in which the throttle step angle is 25° to 40°, engine load is 10% and engine speed of 1000rpm and 1400rpm. Simulation result shows that the system response time decreases by 2/3 than common control system of engine gas mass flow, the control precision of engine excessive air coefficient increases by 25% and engine NOx emission decreases by 20%.