A Comprehensive Thermal Model For System-Level Electric Drivetrain Simulation With Respect To Heat Exchange Between Components

Abstract

Monitoring critical temperatures in the electric drivetrain components is becoming more and more crucial for operational safety and achieving better efficiency. Instead of a distributed thermal model for each component, in this contribution a centralized compact lumped-parameter thermal network model for the electric drivetrain is set up, so that the thermal coupling between inverter, electric motor and gearbox can be considered. The measured and calibrated loss maps as well as empirical functions for losses distribution in the permanent magnet synchronous machine are used to calculate the losses of components. In the thermal modeling, a-priori system knowledge is taken into account in order to reduce parameter identification effort. A global linear parameter-varying identification approach is applied to find the parameters of the lumped-parameter thermal network model. The parametrized thermal model is cross-validated by independent experimental data on the chassis dynamometer. The maximum estimation error of circa 7 °C is achieved at the ambient temperature around 20 °C with the realistic coolant profiles for automotive scenario. The simulation results demonstrate how good the temperatures can be estimated by a centralized lumped-parameter thermal network regarding the thermal coupling between the components.