General Motors Cuts Testing Time in Half by Simulating E-Drive System

GM is a multinational company that not only drives innovation in automotive products, but also in engineering methodologies—enabling shorter product development cycles while lowering costs associated with physical testing. To accomplish this, GM researchers utilize computer models to simulate drive systems and electronic control units (ECUs) before conducting physical testing. However, such simulations have historically been too slow to be effectively integrated directly into test and calibration workflows. Using Simulink^® and Simscape™, GM researchers have now designed a software-in-the-loop (SIL) model to simulate a high-fidelity, virtual ECU (VECU) combined with the plant models for the electric drive system.

Using only Simulink and Simscape, the GM team modeled a complete VECU that includes input from motors, sensors, and power inverters. This model enabled the virtualization of e-drive development tasks, including algorithm development and virtual calibration. The high-fidelity and near-real-time results generated by the VECU allowed it to function as a digital twin to test calibrations before they were applied to physical hardware. Using the Simulink framework also allowed the team to easily integrate and cosimulate various component models of the e-drive system with the VECU.

These virtual models and simulations are powerful enough to cut physical dynamometer testing time in half while being agile enough to operate using a standard CPU. With virtual calibration, the team also achieved 95% of their performance target before the parts and dynamometers were available for physical testing.