Jafari Sadiqa, Byun Yung-Cheol
Department of Electronic Engineering, Institute of Information Science & Technology, Jeju National University, Jeju 63243, South Korea.
Department of Computer Engineering, Major of Electronic Engineering, Jeju National University, Institute of Information Science & Technology, Jeju 63243, South Korea.
Heliyon. 2024 Feb 9;10(4):e25949. doi: 10.1016/j.heliyon.2024.e25949. eCollection 2024 Feb 29.
Global warming, a significant outcome of climate change, exerts detrimental effects on the daily lives of individuals and industries. As a result, there is an increased demand for Electric Vehicles (EVs) to reduce carbon emissions contributing to climate change. This shift underscores the critical need for accurate estimation of the State of Charge (SoC) in battery systems, which is essential for optimizing EVs' performance and ensuring effective energy utilization. This paper introduces a methodically constructed and tested SoC prediction model utilizing a comprehensive dataset derived from various driving cycles and battery records. The battery performance of EVs was assessed in our study. The essence of our innovation resides in the meticulous choice of representative driving cycles, effectively replicating real-world conditions. This methodology improves the model's capacity to apply to various driving patterns and conditions. During these cycles, a comprehensive set of battery data, encompassing voltage, current, temperature, and SoC, was systematically documented to facilitate thorough analysis. To achieve superior accuracy and robustness, our predictive model considers the strengths of the Extra Tree Regressor (ETR) and Light Gradient Boosting algorithms. Our experimental results demonstrate the remarkable performance of the ETR model in predicting SoC, surpassing the LightGBM model. The ETR model exhibited higher values of 0.9983 and lower Root Mean Square Error (RMSE) of 0.62, Mean Absolute Error (MAE) of 0.085, and Mean Squared Error (MSE) of 0.39 values, underscoring its superiority. The research emphasizes the considerable significance of battery capacity in effectively predicting the SoC of EVs. Our research highlights the significant importance of battery capacity in accurately forecasting the SoC of EVs. The proposed model facilitates accurate SoC predictions, improving energy management in EVs to optimize battery utilization and support informed decisions toward sustainable mobility.
全球变暖作为气候变化的一个重要结果，对个人生活和行业产生了不利影响。因此，对电动汽车（EV）的需求增加，以减少导致气候变化的碳排放。这种转变凸显了准确估计电池系统荷电状态（SoC）的迫切需求，这对于优化电动汽车性能和确保有效能源利用至关重要。本文介绍了一种经过系统构建和测试的SoC预测模型，该模型利用了从各种驾驶循环和电池记录中获得的综合数据集。我们的研究评估了电动汽车的电池性能。我们创新的核心在于精心选择具有代表性的驾驶循环，有效地复制了现实世界的条件。这种方法提高了模型适用于各种驾驶模式和条件的能力。在这些循环中，系统地记录了包括电压、电流、温度和SoC在内的一套全面的电池数据，以便进行深入分析。为了实现更高的准确性和鲁棒性，我们的预测模型考虑了极端随机树回归器（ETR）和轻梯度提升算法的优势。我们的实验结果表明，ETR模型在预测SoC方面表现出色，超过了LightGBM模型。ETR模型的 值更高，为0.9983，均方根误差（RMSE）更低，为0.62，平均绝对误差（MAE）为0.085，均方误差（MSE）为0.39，突出显示了其优越性。该研究强调了电池容量在有效预测电动汽车SoC方面的重要意义。我们的研究突出了电池容量在准确预测电动汽车SoC方面的重要意义。所提出的模型有助于准确预测SoC，改善电动汽车的能源管理，以优化电池利用，并支持朝着可持续交通做出明智决策。
