The 10th Asian Symposium on Computational Heat Transfer and Fluid Flow ( ASCHT2025)

This study presents an adaptive battery cooling strategy for electric vehicles using an LSTM-Encoder network to predict short-term temperature rise. By analyzing time-series data of current, voltage, and temperature, the model adjusts coolant flow in real time based on future thermal trends. Compared to fixed-flow systems, the approach improves response to dynamic heat loads, reduces energy consumption, and maintains battery temperature within a safe range, enhancing overall thermal management efficiency.
 

Lithium-Ion Battery, Liquid Cooling, Adaptive Flow Control