Enhancing battery state of charge estimation through hybrid integration of barnacles mating optimizer with deep learning

The precise determination of battery state of charge (SoC) holds paramount significance and has garnered considerable attention across diverse sectors, including academia. Accurate knowledge of the SoC percentage offers numerous advantages, ranging from optimizing travel planning to enhancing the efficiency and reliability of electric vehicle operations through effective battery management systems. In response to the growing importance of SoC estimation, this study introduces a hybrid approach called the Barnacles Mating Optimizer with Deep Learning (BMO-DL) for SoC of Nissan Leaf batteries. The conventional methods for SoC estimation often suffer from limitations in accuracy and robustness, leading to suboptimal EV performance and battery management. In contrast, BMO-DL leverages the power of BMO algorithm to fine-tune the hyperparameters of DL, which is subsequently employed for the actual estimation. This synergistic combination enhances the accuracy and reliability of SoC estimation. The estimation model takes three inputs: voltage, current and conducted charge to generate a single output, the SoC percentage. he study's findings underscore the superiority of BMO-DL by revealing its capability to achieve significantly better results compared to the other benchmarking methods identified. Notably, BMO-DL exhibits significantly lower error rates when compared to competing algorithms, thereby reinforcing its potential to advance the efficiency and reliability of electric vehicle operations while addressing the critical challenge of SoC prediction.