Metamodellbasierte Optimierung von Betriebsstrategien im Thermomanagement elektrischer Fahrzeuge

  • Metamodel-based optimization of operational strategies in thermal management of electric vehicles

Wulff, Carsten Wilhelm Ingo; Pischinger, Stefan (Thesis advisor); Kneer, Reinhold (Thesis advisor)

Aachen (2020)
Dissertation / PhD Thesis

Dissertation, Rheinisch-Westfälische Technische Hochschule Aachen, 2020


In 2019, consumers’ interest in battery electric vehicles has risen significantly. More than half of consumers prefer alternative propulsion technologies for their next vehicle. At the same time, expectations regarding the driving range of these vehicles remain very high. Thermal management effects the achievable range of BEVs considerably, especially at extreme ambient temperatures. Therefore, this work investigates how a globally optimal control strategy in the thermal management can contribute to an increased driving range. To this end, the heat flows required for cabin conditioning are considered as a given boundary condition while a system with many degrees of freedom for the distribution of heat flows within the vehicle is analyzed. A metamodel-based method for global optimization is developed, which is able to identify optimum control strategies for the thermal system with regards to energy consumption of the vehicle using a thermal full vehicle model. The high complexity of this model enables the interdependencies in the system to be considered. Therefore, possible effects of the thermal system on the energy consumption of the drivetrain can be considered. By using the optimization, optimum target values for the control of the cooling system are set. Alternatively, the target values for the actuators within the system can be set directly, comparable to a model-predictive control. When using the optimization algorithm for identifiying optimum control target values, an energy consumption reduction of 2,89 % can be realized in the WLTC Class 3 for 35 °C ambient temperature by optimization of the control of the drivetrain cooling circuit alone. When using the same algorithm for the direct optimization of the actuator target values in the thermal system, energy savings weighted for the annual european temperature distribution of 11,97 % can be realized in the same driving cycle. A stringent requirement for the realization of these potentials is a fully known driving profile of the vehicle for the optimization. Hereby, efficiency potentials can be uncovered which remain unused in rule based control strategies due to component protection against thermal stress. A suitably accurate prediction function is therefore a prerequisite for an implementation of the metamodels developed during the optimization in a model-predictive controller on the vehicle control unit. In this way, most of the potential shown in this work could be transferred to real driving conditions.