Electric Motor Damping Torque Estimation for Early EV Calibration

The patent creates a virtual model that copies the essential dynamics of the actual electrified powertrain system. This virtual model includes a two-degree-of-freedom representation with inertia elements, spring elements, and damper elements that replicate the mechanical behavior of the real system. By working with this copy instead of the original hardware, calibration can be performed earlier in the development process without requiring physical prototypes or test vehicles, thus reducing calibration time while maintaining sufficient accuracy for torque margin allocation verification.

The virtual model enables preliminary calibration actions to be performed before the actual powertrain hardware is available or before costly on-road testing begins. The system allows calibration parameters to be determined in advance through simulation, so that when physical testing does occur, the calibration is already optimized and requires minimal adjustment. This preliminary action significantly reduces the time and resources needed for later calibration stages.

The patent extracts the essential calibration function from the complex real-world system and isolates it into a simplified virtual model. By taking out only the critical dynamic characteristics needed for AEMD calibration (inertia, stiffness, damping properties) and representing them in a two-degree-of-freedom model, the system eliminates the need for complex hardware setups, sensor arrays, and on-road data collection infrastructure. This extraction maintains sufficient optimization capability while dramatically reducing system complexity.

The virtual model allows calibration parameters to be adjusted and optimized systematically through simulation. By changing parameters such as damper coefficients, spring constants, and inertia values in the virtual model, the system can evaluate different calibration scenarios and identify optimal settings without physically modifying the actual powertrain. This parameter-based approach simplifies the calibration process compared to hardware-based trial-and-error methods.