Robust Control Research On Idle Speed Of Gasoline Engine

Automobile is a necessary tool in nowadays transportation, and the dependence on vehicle is becoming more and more great. As the heart of vehicle, engine's performance would directly affect vehicle's performance (power and comfortable). When driving in city, facing the different roads and often in heavily traffic, idle speed condition is one of the most important running operations. So, effective control at idle condition would greatly improve engine's performance and is a very important part in Engine Management System. There were many problems at idle condition when load torque changed, such as hunting, skip, surge and so on. Thus it is an ineluctable problem to improve idle speed stability of gasoline engine.In this dissertation, the engine real-time simulation software enDYNA, which is the product of Tesis corporation in Germany, is used to make engine parameters and operation data. A mean value model of gasoline engine is carefully developed, and some coefficients are determined by using the data from the virtual engine model—enDYNA software. Finally, a robust controller based LMI is formulated. The validation by Matlab/Simulink indicated that the controller would greatly reduce the idle speed's change and fluctuation. And the throttle hard constraints would also be controlled.The main tasks as following:1. The introduction of the LMI theorySome basic concepts about LMI were introduced. Some formulation processes and proofs are presented in detail. Focusing on the control requirement in this dissertation, some useful applications of LMI were also given and discussed.2. The modeling of gasoline engine at idle speed condition Especially for the four-stroke four-cylinders gasoline engine, the general methods of modeling was discussed, Using the mean value model theory and the basic theory of thermodynamics, an mean-value based gasoline engine model was developed, including dynamic model for intake manifold, static model for throttle body, static model for cylinder air induction, fuel path dynamics, torque production and engine speed dynamics. Using the engine real-time simulation software enDYNA, some important coefficients in gasoline engine model was determined. The total nonlinear engine model was developed by Matlab/Simulink toolbox. With impulse and step input signals, model validation was given and indicated the availability of this nonlinear engine model.3. The LMI-based robust H∞controller designBasing the developed nonlinear model, a linear engine model was also developed at the engine model's operation point. It was necessary for robust H∞controller design. For the purpose of controlling idle speed, throttle angle is the only control input that influences engine torque generation, and thus engine speed. The model has two state variables: manifold pressure and engine speed, the disturb input is load torque. It is also necessary to unitary the system's parameters for designing accurate controller. The parameter's uncertainty and hard constraints of the engine model were given after analyzing the system's inputs variables, state variables, output variables. And then an LMI-based robust H∞controller was designed with the limits of the hard constraints such as throttle angle and manifold pressure are actually limited in a certain range. Finally the designed controller was used into both the linear model and nonlinear model, the simulation verified the engine speed could be controlled effectively with disturb load torque, at the same time, the requirements of input constraint and output constraint could be meet. Simulation also shows that model can be controlled well under parameters uncertainty。4. Rapid control prototyping simulationBrief introduced the equipments and tools of rapid control prototyping simulation and especially the dSPACE products. The integrative experiments using dSPACE were made to validate the proposed control approaches in"actual"condition. Controller designed were downloaded to a hardware and run in a real-time environment, respectively. Experiments showed again that the designed controller based-LMI robust H∞control was availability. By the research of modeling and controller design, following conclusions were made:1. The engine model using the mean-value based theory could also shows the physical characteristic. The disturbance of load torque would greatly affect engine speed and a well controller would be useful to avoid engine speed fluctuation and improve the engine's power, NVH and so on.2. Robust controller based LMI would effectively reduce the engine speed fluctuation at the same time satisfied the constraints of systems.3. Robust controller based LMI would also effectively reduce the engine speed fluctuation under the model parameter uncertainty.