Study On The Forward Simulation Of A Fuel Cell Vehicle Based On Driver-Vehicle Closed-loop System

Based on the contract project of Research on Fuel Cell Mechanical-Electronic Hybrid Power Technology, the thesis studies the forward modeling of fuel cell vehicle and driver-vehicle close-loop simulation according to the development work of the fuel cell vehicle for the project.The main content is as following:1. Study of forward modeling for the powertrain of fuel cell vehicles. In this study, a model is made via applying Matlab/Simulink. Combined with the test data of subsystems, the simulation result is validated to finalize the vehicle model. Upon these works, a driver-vehicle-close-loop model is founded and applied for simulating.2. Study on driver-vehicle close-loop longitudinal modeling and simulation. Based on the traditional preview-follower longitudinal driver model, the thesis adds a pre-reactive circuit as the driver's experience reaction. This leads to the need of the research on explaining of the expectation of the driver. The model is validated according to the project fuel cell vehicle's test data.3. Study on the longitudinal control human-vehicle interface - pedals system. Some mass-product pedals and several human drivers are chosen as the experimental samples on a longitudinal driver-vehicle control simulation platform. The actions of driver's pedal-control are studied and analyzed. And some modes and parameters are extracted for describing the actions. Starting from the explanation of driver- expectation, the principal of longitudinal human-vehicle-interface design and its application are stated. 4. Study on the multi-energy control and its strategy. In the research on the vehicle power control strategy, the driver-vehicle close-loop model is applied. The offline-debugging is carried out for the strategy on the hardware controller. The controller and the strategy are applied on the fuel cell vehicle.5. Experimental research on the vehicle control. The project fuel cell vehicle is chosen as the experimental platform. The state-of-art development tools are applied in the vehicle power control strategy test and debugging.6. Foreseen the studies on the control strategy of fuel cell vehicles, driver modeling and human-vehicle-interface.Aimed at the structure of the fuel cell vehicle, the thesis describes the forward-modeling idea and its process. Starting from the driver-vehicle interface (pedals and shifter), a forward model describes the process that torque generated by the engine passes by the powertrain system components to the wheels and drives the vehicle moving forward. Principally, the dynamic models of every sub-system are mandatorily needed in forward modeling. Therefore, the real world engineering limits the approach to such appreciating models of the sub-systems. The static models and data via testing have to be substituted. The research shows that the driver model would act an important role in the forward modeling. The driver model will"drive"the forward model of vehicle to run in the virtual environments-drive cycles. Thereafter, the key and tough point of developing and using the vehicle forward model are the tire model and the driver model. A precise tire model is essential for studying the vehicle forward model to simulate under the limited conditions; such as when take-off. A driver model which could simulate the human driver well is a basic precondition for the study of fuel cell vehicle driving performances.UniTire model works well in the most limited conditions. So, this thesis mainly focuses on improving the longitudinal driver model and carries out the relevant researches.A target-speed-follow driver model is a model to describe how a driver controls a vehicle to run according to a given speed. Aiming at the error of speed following, the thesis suggests an improved model in which a feedforward is added based on the traditional preview-follower driver model. The base of physiology for the feedforward is analyzed. The thesis concludes it as the materializing of driver's expertise in a driver model. This reflects the reaction of a human driver to control the vehicle based on the driving expertise. This is the so-called driver's expertise feedforward. The thesis shows the study on the fix of the parameters describing the feedforward. Analyzing the driving speed data vs. the pedal degree records, a relative coefficient MAP is extracted which could describe the feedforward. Furthermore, the differences of driver's expertise could be described based on the idea. It is harmony with the driver's three levels actions hypothesis that the differences in the feedforward reflect the differences of driver's expertise.The human-pedal system has effects on the process of driver-vehicle longitudinal control. For studying the effects, the thesis designed a series of tests and made analysis on the test data. The chain of rings of the longitudinal human-vehicle interface is concluded. Meanwhile, the study makes the necessary preparation for the commissure of the driver model and vehicle model. Also, the thesis shows a classificatory method for the human drivers in a mathematical way. Moreover, the thesis brings forward the three principles for the design of longitudinal human-vehicle interface. Thereafter, the thesis shows the vista of the driver-vehicle closed-loop simulation.The thesis specifically built a driver-vehicle closed-loop model for the project's fuel cell vehicle based on the above study. The model could be used for the forward simulation as well as the evaluation of different design plan. The model was validated via comparing with the road test data. Series simulation results demonstrate that the driver model with an expertise feedforward could improve the speed-following effectively. Focus on the different drive cycles of the real world, the model could give direct running data so that the powertrain parameter could be designed suitably based on the support of the data. Also, the model could serve as a platform for the multi-power control strategy development of the fuel cell vehicle.The thesis makes an experimental study on the pedals system. The test method materializes the rings of the longitudinal human-vehicle interface. Also, the thesis applies the multi-power control strategy on the project's fuel cell vehicle. The vehicle was tested on its power performance. These tests help to confirm the parameters of the driver model and the vehicle model.To sum up, according to the composition of the power system on the fuel cell vehicle of the project, a forward model is build based on the detail parameters of the fuel cell vehicle. The model was made and applied in simulation validation via experiments of various working status of the fuel cell vehicle. The validated model applied in given working cycles to simulate the fuel cell vehicle running. According to the result, a suggestion was given on the power parameters'chosen as an example of the future applications.