Estimating Electric Motor Temperatures With Deep Residual Machine Learning

Abstract

Most traction drive applications lack accurate temperature monitoring capabilities, ensuring safe operation through expensive oversized motor designs. Classic thermal modeling requires expertise in model parameter choice, which is affected by motor geometry, cooling dynamics, and hot spot definition. Moreover, their major advantage over data-driven approaches, which is physical interpretability, tends to deteriorate as soon as their degrees of freedom are curtailed in order to meet the real-time requirement. In this work, deep recurrent and convolutional neural networks (NNs) with residual connections are empirically evaluated for their feasibility on predicting latent high-dynamic temperatures continuously inside permanent magnet synchronous motors (PMSMs). Here, the temperature profile in the stator teeth, winding, and yoke as well as the rotor's permanent magnets are estimated while their ground truth is available as test bench data. With an automated hyperparameter search through Bayesian optimization and a manual merge of target estimators into a multi-head architecture, lean models are presented that exhibit strong estimation performance at minimal model sizes. It has been found that the mean squared error and maximum absolute deviation performances of both, deep recurrent and convolutional NNs with residual connections, meet those of classic thermodynamics-based approaches, without requiring domain expertise nor specific drive train specifications for their topological design. Finally, learning curves for varying training set sizes and interpretations of model estimates through expected gradients are presented.

Machine Learning for Design Optimization of Electromagnetic Devices: Recent Developments and Future Directions

“…Consequently, surrogate models based on parametric or semi-parametric approaches might not follow accurate results in general. Through several attempts on machine learning methods, it is found that deep learning algorithms, like CNN and RNN, are good at the distribution estimation of magnetic field and temperature [129][130][131], and the prediction of torque and efficiency for motors [132][133][134]. These works established a solid foundation for the generalizable data-driven model for the analysis, design, and optimization of electromagnetic devices [129].…”

Thermal Monitoring of Electric Motors: State-of-the-Art Review and Future Challenges

“…This testing step analyzes how well the found model generalizes to unseen data, which were not part of the training set. [106]): X denotes model inputs, i.e., measured quantities which are basically available in the motor control unit while Y represents the temperature targets to be estimated. For the data acquisition process, special motors with additional temperature sensors are required.…”

“…Typical examples are recurrent networks with internal memory cells or convolutional neural networks with large time spans of past data at their input (cf. [106], [115]). Recently, state-space neural networks with a special recurrent structure mimicking state-space model dynamics have been also introduced [118].…”

A Review of Thermal Monitoring Techniques for Radial Permanent Magnet Machines

“…[89] relies on the high-frequency inductance for temperature estimation. References [90][91][92] belong to the AI-based methods. [93][94][95][96][97][98][99] are literatures about the characteristics of PM materials.…”