Temperature prediction and cooling optimization for integrated electric drives via reinforcement learning

Accurate temperature prediction and cooling optimization are vital for integrated electric drive systems (EDS). This study develops a lumped parameter thermal network (LPTN) model to evaluate the global thermal behavior of a three-in-one EDS, comprising the motor, water-cooled controller, and oil-cooled reducer. Steady-state validation reveals severe thermal concentration at the motor winding ends, with winding end2 identified as the highest thermal load for detailed analysis. Simulations show winding end2 reaches 120.398°C at rated conditions (2.032% relative error vs measurements) and 117°C at peak speed (0.863% error). To address non-linear thermal dynamics during the NEDC driving cycle, a reinforcement learning (RL) algorithm was implemented to dynamically optimize cooling pump speeds. To further interpret the cooling-control behavior, equivalent pump-speed, pump-power, and cumulative pump-energy indices were reconstructed from the simulated thermal response and equivalent cooling demand. Compared to conventional PID control, the RL strategy mitigated overall thermal loads, reducing motor and controller temperatures by 2°C–5°C and reducer gear temperatures by 1°C–4°C. The equivalent energy-index analysis indicates that the RL strategy suppresses peak temperatures by actively allocating cooling effort during high-load operation, thereby providing a broader thermal safety margin for integrated EDS.

Authors

Institutions

Publication Details

Journal
Proceedings of the Institution of Mechanical Engineers Part D Journal of Automobile Engineering
Published
2026-11-20
DOI
https://doi.org/10.1177/09544070261506864
Primary Topic
Electric Motor Design and Analysis
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Temperature prediction and cooling optimization for integrated electric drives via reinforcement learning

Liange He, Renhua Feng, Yan Zhang, Bingqi Tong et al.
Proceedings of the Institution of Mechanical Engineers Part D Journal of Automobile Engineering
Electric Motor Design and Analysis
article

Temperature prediction and cooling optimization for integrated electric drives via reinforcement learning

Liange He, Renhua Feng, Yan Zhang, Bingqi Tong, Lantian Tan, Limin Wu, Tangzhi Liu, Xingyan Bu
article en

Abstract

Accurate temperature prediction and cooling optimization are vital for integrated electric drive systems (EDS). This study develops a lumped parameter thermal network (LPTN) model to evaluate the global thermal behavior of a three-in-one EDS, comprising the motor, water-cooled controller, and oil-cooled reducer. Steady-state validation reveals severe thermal concentration at the motor winding ends, with winding end2 identified as the highest thermal load for detailed analysis. Simulations show winding end2 reaches 120.398°C at rated conditions (2.032% relative error vs measurements) and 117°C at peak speed (0.863% error). To address non-linear thermal dynamics during the NEDC driving cycle, a reinforcement learning (RL) algorithm was implemented to dynamically optimize cooling pump speeds. To further interpret the cooling-control behavior, equivalent pump-speed, pump-power, and cumulative pump-energy indices were reconstructed from the simulated thermal response and equivalent cooling demand. Compared to conventional PID control, the RL strategy mitigated overall thermal loads, reducing motor and controller temperatures by 2°C–5°C and reducer gear temperatures by 1°C–4°C. The equivalent energy-index analysis indicates that the RL strategy suppresses peak temperatures by actively allocating cooling effort during high-load operation, thereby providing a broader thermal safety margin for integrated EDS.

Proceedings of the Institution of Mechanical Engineers Part D Journal of Automobile Engineering
Chongqing University of Technology (CN), Chongqing Jiaotong University (CN)
Electric Motor Design and Analysis
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon