基于XGBoost模型的锂电池SOC预测与影响因素分析

打开文本图片集

中图分类号：TP39；TM912 文献标识码：A 文章编号：2096-4706（2026）03-0010-04

SOC Prediction and Influence Factor Analysis of Lithium Battery Based on XGBoostModel

WANG Zhijun （Datang Sanmenxia Power Generation Co.，Ltd.， Sanmenxia 472143， China）

Abstract：Lithiumbatteriesarewidelyusedinfeldssuchasnergy，control，transportation，andelectronic industrydueto theirexcelentenergystorageanddischarge performance.Accuratelypredicting the State-of-Charge （SOC）oflithiumbaeryis a keylink tooptimzebateryseeciencyeancesystesfetyandextendaterylife.Tisstudyims toonstrtaig precision and interpretable SOC prediction method.By merging electric bus operationdata and external weather data，his paper establishesadata-driven modeltoimprove the accuracyandreliabilityofSOCestimationandprovide theoreticalsupportfor battery management and energyconsumption optimization.This paper proposes alithium baterySOC prediction method based on XGBoost model， which integrates batery status data and weather data toconstruct atraining dataset.This papercolets relevant dataoflithium-ionbatery toverifythe model.TheresultsshowthattheRootMean SquareError（RMSE）of thismodel onthetestdatasetreachesO.O928.Meanhile，feature importanceanalysisfindsthatmeantotalvoltage，averagetemperature， and average speed have the most significant influence on the model.

Keywords： lithium battery; battery SOC prediction;XGBoost; electric bus

0 引言

随着全球能源转型和可持续发展战略的深入实施，锂离子电池作为高效储能单元，其性能优化与管理技术成为关键研究领域。（剩余5117字）