EV Battery Adaptive Test Unit

Lithium-ion batteries are an important component in electric vehicles (EVs), and ensuring their optimal performance is vital for the longevity and efficiency of the vehicle. Efficient testing of battery cells plays a major role in verifying their performance. Traditionally, a single-cell testing unit is used for batch processing during battery pack assembly. However, this method is time-consuming and lacks scalability, which limits productivity. To address this, the research proposes a multicell testing approach that concurrently estimates the state of charge (SOC) and state of health (SOH) of multiple battery cells in parallel. By doing so, the approach significantly reduces testing time and improves efficiency. The dual filter concept is incorporated to categorize cells based on their performance, ensuring only high- quality cells are selected for inclusion in the battery pack. Furthermore, a custom Temporal Convolutional Network (TCN) model, achieving an accuracy of 89%, is employed to accurately estimate SOC and SOH. In addition, a predictive battery temperature forecasting model is introduced to forecast the temperature of the battery cells over the next three days, which aids in proactive temperature management and prevents potential degradation. Overall, the proposed approach enhances battery testing productivity and ensures higher accuracy in SOC and SOH estimation, contributing to the development of more reliable and efficient EV batteries.