Model Predictive Control of Advanced Hybrid Powertrain Systems

In today’s automotive industry, much of the focus has shifted to advanced vehicle propulsion systems. This is due to many factors, the main ones being climate change and the diminishing supply of gasoline. This thesis addresses some of the control objectives involved when coupling two power supplies as hybrid vehicles do. The main focus of the research is the use of Model Predictive Control to achieve both thermally efficient and fuel efficient algorithms to control the internal combustion engine in a hybrid powertrain. This is done with the use of two operational modes: Fuel Use Mode which limits the fuel consumption of the engine, and Efficiency Mode which maximizes the thermal efficiency of the engine. A mathematical overview of MPC is described to bring the reader into context. Next, this type of control is applied to a series-parallel hybrid electric vehicle. The formulation of the vehicle model in Simulink is discussed in detail. Then, a practical form of MPC is implemented, by using a torque vector to estimate the outputs of engine model, and minimizing the resulting cost function. Next, MPC parameter selection is discussed, which covers the choices for receding horizon length, cost function weights, and uncertainty correction parameters. After all the parameters have been chosen, simulations are run for the US06 and FUDS drive cycles, and the results are analyzed. The results show that Fuel Use Mode yields a higher miles-per-gallon (MPG) rating, and Efficiency Mode helps maintain the charge level of the batteries.