Study On Rail Pressure Control Of Engine Based On Neural Network

The traditional PID method has many problems in rail pressure control.On the one hand,the traditional method is not robust and stable for time-varying and nonlinear systems.On the other hand,the optimal control parameters of different engines are not the same.Without adaptive adjustment capability,it is difficult to optimize the optimal parameters in a short time,which is time-consuming and labor-intensive.In addition,there is a long pressure build-up time when various operating conditions are switched,and the rail pressure adjustment has a large lag problem.In view of the above problems,the author proposes a new adaptive controller based on neural network algorithm.This paper first establishes a mathematical model of the high-pressure fuel pump,injector,common rail tube,and related components based on the composition principle of the fuel system,and analyzes the logic correctness and accuracy of the model.At the same time,the relationship between fuel injection,fuel supply and rail pressure fluctuation was further explored.Then,the rail pressure control strategy based on neural network feedforward BPNNPID control is designed,and the network parameters affecting the network convergence are studied in depth.Finally,the engine bench test is performed on a rapid prototype platform system,and the control effects of the traditional controller and the neural network controller are compared.The simulation and bench test results show that the neural network controller is superior to traditional controllers in response speed and stabilization speed.Among them,the problem of large overshooting of construction pressure during the switching of working conditions is effectively improved,and the construction pressure speed is improved.Under stable conditions,the anti-disturbance performance of neural network control is better than that of traditional controllers,especially in high-speed conditions and small cycle fuel injection.The optimization effect is obvious.The research proposed on the one hand helps to achieve the precise control of diesel fuel injection,and on the other hand provides a reference for the intelligent technology of diesel control systems in China.