Simulation Platform for Powertrain and Thermal Systems Development and Testing

Developing vehicle powertrain and thermal control algorithms requires a lengthy proof-of-concept process, and simulation tools can significantly accelerate this. At Hyundai, our team’s job is to develop and prototype advanced control strategies using simulation software. This presentation shows how the team established a high-fidelity, high-performance simulation and co-simulation platform for virtual testing and development of powertrain and thermal control logic. The project used the Hyundai Ioniq 5 as the test platform and focused on developing a thermal model predictive control algorithm.

The existing vehicle model was built in GT-SUITE and the control algorithm was prototyped in Simulink^®. The team chose to use the existing models to build a standalone Ioniq 5 digital twin entirely in MATLAB^®. The powertrain model was constructed using lookup tables and mathematical representations, while the thermal system was built with Simscape^™ using the GT-SUITE model as a reference. On-road data collected from an instrumented vehicle was used to validate the model.

The simulation takes imposed torque and speed as inputs and incorporates replicated powertrain and thermal control logic to simulate temperatures of key components, including the battery, coolant, motors, inverters, and refrigerant. The high-fidelity thermal refrigerant loop significantly slowed simulation performance, so the team developed a data-driven reduced-order model (DDROM) to replace the refrigerant loop, greatly improving computational efficiency. The DDROM was trained using on-road data collected under various ambient temperature conditions. Throughout the model development process, reduced-order modeling techniques and Simscape were used to achieve both accuracy and performance.