Multi-Objective Optimization of the Energy Management Strategy for Hybrid Electric Vehicles -Examination of Sensitivities between Fuel Consumption, Electric Driving and Driving Comfort

Abstract

The combination of the electric drive and the internal combustion engine (ICE) in hybrid electric vehicles (HEV) requires the implementation of an Energy Management Strategy (EMS). The task of the EMS is to split the driving demand between the two energy converters. The design of the EMS in charge-sustaining operation is commonly targeted at the minimization of fuel consumption. For in-vehicle implementation of the EMS, supplementary objectives, such as the electric driving (ED) experience or the driving comfort, influenced by the frequency of state shifts, are considered. Therefore, this work extends the framework for EMS optimization from the fuel-optimal design to multi-objective target spaces. First, the general multi-objective optimal control problem (MOOCP) is formulated. In a next step the central target space for EMS calibration consisting of fuel consumption, ED time and number of ICE starts is considered and the resulting MOOCP is solved using Dynamic Programming (DP). The solutions are compared in WLTC, and the resulting EMS behavior is analyzed. In the field of in-vehicle calibration of EMS, it is of fundamental importance to develop a basic understanding of the sensitivities between the optimization objectives. Therefore, the pareto-optimal solutions for several real-world driving cycles (DC) are determined and their characteristics are investigated. Using a statistical evaluation, correlations between the sensitivity and DC properties are derived. To further explore the sensitivities in the multi-objective target space considered, the influence of different driving resistance (DR) is highlighted. This work contributes by examining the correlations in the multi-objective design of the EMS for HEVs as well as the determination of the influence of different boundary conditions. In future work, the multi-objective optimization framework can be used as a benchmark for the design of causal EMS utilized in production vehicles.