In this study, a practical numerical method is proposed to estimate losses of high specific power density electric motors, using few simulated temperature data. In such electric motors, these losses generate high heat fluxes inside the motor components that can be critically sensitive to temperature. Electromagnetic and mechanical friction phenomena are behind the occurring of these thermal dissipations. For both phenomena, losses could be difficult to compute with electrical or mechanical approaches. However, thermal management of electric motors requires a precise knowledge of those losses, in particular for high-performance motors such as those considered in future hybrid planes. To determine electric motor losses in a permanent magnet synchronous motor (PMSM) in real time, an inverse method using a lumped parameter thermal model (LPTM) is elaborated. In the first step, the dynamic profile of losses is determined through the inverse method, based on temperature data at easy-access points of the motor. In a second step, the identified losses are used to find temperatures at critical non-accessible hot spot points of the motor through forward LPTM. The method is applied for three useful cases, from the simplest case scenario, where only one type of losses has to be identified, to the most complicated case where all losses are simultaneously estimated. A global strategy for the choice of the number of future time steps used for regularization of the ill-posed problem is also proposed. Results show that this method enables adequate real-time supervision of the critical motor temperatures, mainly rotor and winding core.

中文翻译：

---

通过逆向方法确定电动机损耗和临界温度

在这项研究中，提出了一种实用的数值方法来估计高比功率密度电动机的损耗，使用很少的模拟温度数据。在此类电动机中，这些损耗会在对温度极为敏感的电动机组件内部产生高热通量。这些热耗散的发生背后是电磁和机械摩擦现象。对于这两种现象，使用电气或机械方法可能难以计算损耗。然而，电动机的热管理需要对这些损失有准确的了解，特别是对于高性能电动机，例如未来混合动力飞机中考虑的那些。为了实时确定永磁同步电机 (PMSM) 中的电机损耗，详细阐述了使用集总参数热模型 (LPTM) 的逆方法。在第一步中，基于电机易于访问点的温度数据，通过逆方法确定损耗的动态曲线。在第二步中，识别出的损耗用于通过正向 LPTM 查找电机不可接近的关键热点处的温度。该方法适用于三种有用的案例，从最简单的案例场景，其中只需要确定一种类型的损失，到最复杂的案例，所有损失都同时估计。还提出了用于选择用于正则化不适定问题的未来时间步数的全局策略。结果表明，该方法能够充分实时监控关键电机温度，主要是转子和绕组铁芯。