Neural Dynamics Model for Temperature Estimation of Permanent Magnet Synchronous Motor

Accurate temperature estimation of permanent magnet synchronous motors is the basis for ensuring safe operation and designing effective thermal management strategies. Model-based estimation methods exhibit superior real-time performance, but the intricate modeling process requires substantial expert knowledge and lacks versatility. To the contrary, data-driven estimation methods, while offering flexibility, often lack physical implications in terms of system dynamics. This paper proposed a structured linear neural dynamics model for motor temperature estimation. The stability of the data-driven pipeline is facilitated by the Perron-Frobenius theorem, and the smooth evolution of the state between time steps is achieved by utilizing a physics-informed loss term. The effectiveness of the proposed method is verified by temperature estimation tasks at the edge end, and its practicality is demonstrated by measuring the memory and power usage of a deployed STM32 microcontroller.