基于LightGBM的蓄电池容量预测方法研究

Research on Battery Capacity Prediction Method Based on LightGBM

通过对通信机房中浮充状态的电池的放电电压曲线进行分析,确定了放电曲线中对容量影响显著的陡降复升段的特征。利用LightGBM算法进行训练,建立了预测容量的数学模型。通过实验对模型进行了验证,实验结果表明所提出的预测方法能对10h放电条件下的容量做出准确预测,其平均绝对百分比误差最小可达8%。同传统的检测方法相比,该方法所需的测试时间短,具备广阔的应用前景。

By analyzing the discharge voltage curve of the battery in the floating charge state in the communication room,the characteristics of the“coup de fouet”that have a significant impact on capacity in the discharge curve have been determined.The LightGBM algorithm is used for training,and a mathematical model for predicting capacity is established.The method is validated through experiments,and the experimental results show that the proposed prediction method can accurately predict the capacity under 10 hour discharge conditions,with an average absolute percentage error of at least 8%.Compared with the traditional detection methods,this method requires less testing time and has broad application prospects.