The gradual decline of oil reserves and the increasing demandfor energy over the past decades has resulted in automotive
manufacturers seeking alternative solutions to reduce the
dependency on fossil-based fuels for transportation. A viable
technology that enables significant improvements in the overall
tank-to-wheel vehicle energy conversion efficiencies is the hybridization
of electrical and conventional drive systems.
Sophisticated hybrid powertrain configurations require careful coordination
of the actuators and the onboard energy sources
for optimum use of the energy saving benefits. The term optimality
is often associated with fuel economy, although other measures such as
drivability and exhaust emissions are also equally important.
This dissertation focuses on the design of hybrid-electric vehicle (HEV)
control strategies that aim to minimize fuel consumption while maintaining good
In order to facilitate the design of controllers based on mathematical
models of the HEV system, a dynamic model that is capable of
predicting longitudinal vehicle responses in the low-to-mid frequency region (up to 10 Hz) is developed for a parallel HEV configuration. The model is validated using experimental data from various driving modes including electric only, engine only and hybrid. The high fidelity of the model makes it possible to accurately identify critical drivability issues such as time lags, shunt, shuffle, torque holes and hesitation.
Using the information derived from the vehicle model, an energy management
strategy is developed and implemented on a test vehicle. The resulting control strategy has a hybrid structure in the sense that the main mode of operation (the hybrid mode) is occasionally interrupted by event-based rules to enable the use of the engine start-stop function. The changes in the driveline dynamics during
this transition further contribute to the hybrid nature of the system.
To address the unique characteristics of the HEV drivetrain and to ensure smooth
vehicle operation during mode changes, a special control method is developed.
This method is generalized to a broad class of switched systems in which the switching conditions are state dependent or are supervised. The control approach involves partitioning the state-space such that the control law is modified as the state trajectory approaches a switching set and the state is steered to a location within the partition with low transitioning cost. Away from the partitions that contain switching sets, the controller is designed to achieve any suitable control objective. In the case of the HEV control problem, this objective generally involves minimizing fuel consumption.
Finally, the experimental verification of this control method
is illustrated using the application that originally motivated the
development of this approach: the control of a HEV driveline during the
transition from electric only to hybrid mode.