Multiphysics Simulation Approach for Evaluating Thermo-Structural Performance of Electric Motors Using Spatial Electromagnetic Loss Mapping

<div class="section abstract"><div class="htmlview paragraph">High power and torque density electric motor is finding increasing demands in modern-day electric and hybrid vehicles because of compact and light-weight designs. These high-performance requirements are achieved by increasing the current flow, strengthening the magnetic field as well as downsizing the motor dimensions and hence can lead to multiple failure modes if not designed properly. Higher current flow results in increased magnitude of losses within the motor components such as ohmic loss, iron loss, hysteresis loss and mechanical losses. All these localized losses contribute to higher operating temperature and temperature gradient that can act as a catalyst to several modes of failure. Hence, accurate prediction of temperature distribution across the motor components is very crucial to come up with a robust and durable motor design. A common approach of predicting component temperature is by assuming bulk losses for lamination stack, hairpin and magnets. This approach might be beneficial for comparison between different design suggestions but from lifetime durability point of view, appropriate spatial distribution of losses and its transient history must be analyzed. This study focuses on a coupled electromagnetic and thermo-structural simulation approach to predict the overall temperature distribution in motor components by considering spatially distributed losses. The electro-magnetic (eMag) analysis highlights the impact of magnetic saturation and the non-linear behavior of core materials on loss distribution while the thermo-structural analysis highlights the impact of orthotropic thermal and structural behavior of the core materials during motor operation. A special mapping technique using K-Nearest neighbor algorithm is also highlighted in this paper to seamlessly propagate the loss distribution from electromagnetic solver to the thermo-structural solver leveraging dissimilar finite element (FE) mesh. The difference in temperature distribution from this approach is also compared with that using traditional bulk-loss approach. The predicted temperature distribution is also utilized to understand the motor durability against different failure modes and hence this overall multi-physics analysis approach can be used as a decision-making tool in the initial design phases of high-end electric motors.</div></div>