Thermal heat flux distribution prediction in an electrical vehicle battery cell using finite element analysis and neural network

In terms of battery design and evaluation, Electric Vehicles (EVs) are receiving a great deal of attention as a modern, eco-friendly, sustainable transportation method. In this paper, a novel battery pack is designed to maintain a uniform temperature distribution, allowing the battery to operate within its optimal temperature range. The proposed battery design is part of a main channel where a portion of cool air will pass from an inlet then exit from an outlet where a uniform temperature distribution is maintained. First, a 3-D model of a battery cell was created, followed by thermal simulation for 15C, 25C, and 35C ambient temperatures. The simulation results reveal that the temperature distribution is nearly uniform, with slightly higher values in the middle portion of the cell height. Second, using finite element analysis (FEA), it was determined that the heat flux per unit area is nearly uniform with a slight increase at the edges. Third, a machine learning model is proposed by utilizing a neural network (NN). Lastly, the heat flux values were predicted using the NN model that was proposed. The model was assessed based on statistical measures where a root mean square error (RMSE) value of 0.87% was achieved. The NN outperformed FEA in terms of time consumption with a high prediction accuracy, leveraging the potential of adopting machine learning over FEA in related operational assessments. [Display omitted] •Optimizing battery temperature: the study focuses on designing a battery pack to maintain a uniform temperature distribution in electric vehicle (EV) battery cells, ensuring they operate efficiently.•Thermal analysis: through thermal simulations and finite element analysis (FEA), the research shows nearly uniform temperature distributions in different ambient conditions, with slight variations in the middle of the cell height.•Machine learning for heat flux: a neural network (NN) model is introduced to predict heat flux within the battery cell. It achieves high prediction accuracy with an RMSE of 0.87%, outperforming FEA and reducing analysis time.•Efficiency gains: the study highlights the potential of machine learning for faster and accurate assessments in battery design and evaluation, crucial for electric vehicle development.