MPC-Based Energy Management of a Parallel Hybrid Electric Vehicle Using Terrain Information

A parallel hybrid electric vehicle (HEV) combines the power produced by electric machines and a combustion engine to enable improved fuel economy. Optimization of the power-split algorithm managing both torque sources can be readily achieved offline, but online implementation results often show great deviation from expected fuel economy due to traffic, hills, and similar effects that are not easily modeled. Of these external influences, the road grade for a travel route is potentially known a priori given a set destination choice from the driver. To examine whether grade information can improve the performance of a hybrid powertrain controller, we first formulate the vehicle model as a low-order dynamic model, recognizing that the primary dynamics of the energy system are slow. A model predictive control (MPC) strategy utilizing the terrain data is then developed to obtain a time-varying power split between the combustion engine and the electrical machine. Simulation results of the HEV model over multiple standard drive cycles, with different terrain profiles and different cost functions, are presented. Testing of the MPC performance compared to Argonne National Lab’s powertrain simulation software Autonomie shows that the MPC strategy utilizing terrain data gives an improvement of up to 2.2% in fuel economy with respect to the same controller without terrain information, on the same route.