Protection-oriented supervisory interpretation of PMSM thermal virtual sensing for traction drives
Permanent-magnet synchronous motor (PMSM) traction drives require thermal supervision that supports protective action rather than accuracy reporting alone. This short communication presents a control-facing interpretation layer that converts one-step thermal virtual-sensor outputs into quantities directly usable in traction-drive supervision. Using the public Electric Motor Temperature PMSM dataset sampled at 2 Hz and a leakage-safe profile-wise split of 48 training, 10 validation, and 11 test profiles, channel-wise one-step temperature estimates were evaluated for permanent magnet, stator winding, stator tooth, and stator yoke targets. The proposed supervisory layer reports peak estimated temperature, time spent above a warning threshold, and a continuous derating factor that can be mapped to the admissible stator-current envelope within field-oriented control, including field-weakening operation. Test RMSE values were 0.0891°C for permanent magnet, 0.0693°C for stator winding, 0.0377°C for stator tooth, and 0.0312°C for stator yoke; the corresponding MAE values were 0.0381, 0.0439, 0.0236, and 0.0185°C. For a representative test profile on the magnet channel, the maximum estimated temperature remained below an illustrative warning threshold of 120°C, giving zero time-above-warning and full derating availability. The communication therefore reframes PMSM thermal virtual sensing as an auditable supervisory input for electric-vehicle protection logic.
Authors
- Remya Sree R. (ORCID: https://orcid.org/0009-0001-9597-4785)
- Rajesh G. (ORCID: https://orcid.org/0009-0003-3128-0906)
- Maruthupandi P
- Sebasthirani K
Institutions
- Abu Dhabi University (AE)
- Anna University, Chennai (IN)
- Indian Institute of Technology Palakkad (IN)
Publication Details
- Journal
- Proceedings of the Institution of Mechanical Engineers Part D Journal of Automobile Engineering
- Published
- 2026-07-22
- DOI
- https://doi.org/10.1177/09544070261470645
- Primary Topic
- Machine Fault Diagnosis Techniques
- Type
- article
- Field-Weighted Citation Impact
- 0.00