Computation of eco-driving cycles for Hybrid Electric Vehicles: Comparative analysis

In this paper, the calculation of eco-driving cycles for a Hybrid Electric Vehicle (HEV), using Dynamic Programming (DP), is investigated from the complexity-solving method viewpoint. The study is based on a comparative analysis of four optimal control problems formulated using distinct levels of modeling. Starting with three state dynamics (vehicle position and speed, battery state-of-charge) and three control variables (engine and electric machine torque, gear-box ratio), the number of state variables is reduced to two in a first simplification. The other two simplifications are based on decoupling the optimization of the control variables into two steps: an eco-driving cycle is calculated assuming that the vehicle is propelled only by the engine. Then, knowing that the vehicle follows the eco-driving cycle calculated in the first step, an off-line energy management strategy (torque split) for an HEV is calculated to split the requested power at the wheels between the electric source and the engine. As is shown, the decreased complexity and the decoupling optimization lead to a sub-optimality in fuel economy while the computation time is noticeably reduced. Quantitative results are provided to assess these observations.