Optimal control of a flywheel-based automotive kinetic energy recovery system

摘要

This thesis addresses the control issues surrounding flywheel-based Kinetic Energy Recovery Systems (KERS) for use in automotive vehicle applications. Particular emphasis is placed on optimal control of a KERS using a Continuously Variable Transmission (CVT) for volume car production, and a wholly simulation-based approach is adopted. Following consideration of the general control issues surrounding KERS operation, a simplified system model is adopted, and the scope for use of optimal control theory is explored. Both Pontryagin’s Maximum Principle, and Dynamic Programming methods are examined, and the need for numerical implementation established. With Dynamic Programming seen as the most likely route to practical implementation for realistic nonlinear models, the thesis explores several new strategies for numerical implementation of Dynamic Programming, capable of being applied to KERS control of varying degrees of complexity. The best form of numerical implementation identified (in terms of accuracy and efficiency) is then used to establish via simulation, the benefits of optimal KERS control in comparison with a more conventional non-optimal strategy, showing clear benefits of using optimal control.