融合物理机理的图神经网络电机热状态感知方法及系统
By integrating low-order thermal networks and graph neural networks, a motor internal temperature estimation system is constructed, which solves the problems of difficult sensor integration and poor physical interpretability in traditional methods. It achieves high-precision, real-time motor temperature monitoring and is suitable for embedded motor control systems.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- XIAN UNIV OF TECH
- Filing Date
- 2026-04-22
- Publication Date
- 2026-07-17
AI Technical Summary
Existing methods for estimating motor temperature are difficult to achieve high-precision and real-time internal temperature monitoring in high-performance electromechanical systems. Traditional methods suffer from problems such as difficulty in sensor integration, electromagnetic interference, high computational complexity, and poor physical interpretability.
A graph neural network method integrating physical mechanisms is adopted. By identifying key component nodes and their heat transfer paths through a low-order thermal network model, the topology of the graph neural network is constructed. Combined with graph convolutional layers and bidirectional gated recurrent networks, a residual learning strategy is introduced to establish a physical constraint loss function, thereby achieving accurate estimation of the internal temperature of the motor.
It achieves high-precision, real-time, and robust monitoring of the internal temperature of the motor, avoiding the problems of difficult sensor integration and poor physical interpretability of traditional methods. It is suitable for embedded motor control systems with limited computing power and improves the reliability of motor temperature monitoring across the entire range.
Smart Images
Figure CN122087355B_ABST