Identification and Update of Thermal Models in Digital Twins of ePowertrain Motors

Authors

J. Garcia Urbieta, M. Marijuán, P. Díaz, A. Rodríguez, I. García, S. Armentia, F. González - GKN Automotive

Summary

Thermal phenomena are of critical importance in automotive applications of per- manent-magnet synchronous motors (PMSMs). Heat accumulation in sensitive locations, such as the magnets or the endwindings, can decrease the perfor-mance of the device and damage it in extreme cases. Preventive actions can be defined based on the accurate knowledge of the temperatures of the device, but monitoring thermal effects through direct measurements is often unfeasible in production units. An alternative consists in building real-time capable digital twins to represent the thermal effects in the motor, using to this end a limited set of sensor data. Here, an approach to build these digital twins that relies on a Lumped-Parameter Thermal Network (LPTN) representation of the heat flows and temperatures in the PMSM is presented, whose predictions are fused with sensor information by means of state, input, and parameter observers. For the LPTN to provide a meaningful description of the motor behaviour, it is necessary to identify the relevant thermal effects that take place inside the device and translate them into appropriate parameters and input functions.
...