| Hybrid vehicle technology has been widely studied in recent years because of its potential to significantly improve fuel economy and reduce emissions of ground vehicles. The supervisory control problem of hybrid vehicles is commonly referred to as the power management problem, i.e., the design of the control algorithm that determines the proper power split between different power sources in order to minimize fuel consumption and emissions. However, due to complex configuration and multiple objectives, control strategies based on engineering intuition frequently fail to fully explore the potential of these advanced vehicles. In this dissertation, a model-based control design procedure based on system-level modeling and dynamic programming is proposed to solve the power management problem of hybrid vehicles.; A systematic control design procedure is first illustrated by using a parallel hybrid electric truck. An integrated hybrid electric truck model suitable for driving-cycle simulation is developed. The control design starts with the formulation of a finite-horizon dynamic optimization problem for minimizing fuel consumption and selected emission species over a driving cycle. Deterministic dynamic programming (DDP) is then utilized to find the optimal control action. Through the analysis of the behavior of the DDP control actions, near-optimal control rules can be extracted. The dynamometer-testing results of a prototype hybrid electric truck confirm that the proposed control strategy achieves significant fuel economy improvement compared with a conventional rule-based control method. In order to enhance control performance over diverse driving scenarios, an on-line control scheme using driving pattern recognition is also presented.; We next employ an alternative method which uses a stochastic approach to tackle the power management problem. The power demand from the driver is modeled as a random Markov process. The optimal control strategy is then solved by using stochastic dynamic programming (SDP). The obtained control law is in the form of a stationary state feedback and can be directly implemented.; Finally, the power management problem of a hybrid fuel cell vehicle is investigated. An integrated dynamic model with PEM fuel cells, a DC/DC converter, a battery, and an electric drive is developed. The SDP methodology is applied to obtain an optimal policy of the fuel cell current command. Simulation analysis shows that significant fuel (hydrogen) savings can be achieved for this hybrid fuel cell vehicle. |
