Spatial and temporal enhancement method for predicting heat generation characteristics of pouch lithium-ion battery

Detecting the time-evolving non-uniform distribution of temperature fields in pouch lithium-ion batteries remains a significant challenge in battery research. This study introduces two methods for real-time temperature reconstruction of pouch lithium-ion batteries. A numerical simulation model of the pouch-type lithium-ion cell was developed and validated using experimental data. The heat generation characteristics were analyzed at different discharge rates, and spatial and temporal data enhancement strategies were used to reconstruct the real-time temperature field under varying discharge rates and depths of discharge (DOD). Results showed the non-uniform temperature distribution of the battery was influenced by discharge rate and DOD. The proposed method allowed for reconstruction of the non-uniform temperature distribution using temperature measurements from a limited number of sensors on the battery's external surface. K-means clustering was employed to identify three primary regions of heat generation and optimal sensor placement within these clusters, enabling global field reconstruction with just three sensors. However, the reconstruction accuracy was dependent on the proportion of DOD selected for the training dataset; when the training dataset exceeds 30 % DOD, the reconstruction mean squared error (MSE) remains below 0.3 %. On the other hand, Temporal resolution enhancement was achieved using extended proper orthogonal decomposition (EPOD), which accurately captures high-resolution temperature features, reducing the need for frequent temperature monitoring. Under a discharge rate of 1.0C, the MSE remaining below 1.2 % for most of the discharge process from 13 % to 80 % DOD. This study may provide a guideline for the spatial and temporal reconstruction of the time-evolving non-uniform temperature distribution in pouch lithium-ion batteries.