基于改进随机森林算法区域短期售电量预测方法  

Regional short⁃term electricity sales forecasting method based on improved random forest algorithm

作　　者：魏梦飒 李强 许成娣 黄瑶 陈译耘 WEI Mengsa;LI Qiang;XU Chengdi;HUANG Yao;CHEN Yiyun(State Grid Shanghai Songjiang Electric Power Company,Shanghai 201600,China)

机构地区：[1]国网上海市松江供电公司,上海201600

出　　处：《电子设计工程》2023年第12期151-154,159,共5页Electronic Design Engineering

摘　　要：为避免因短期售电量过大对区域电网造成的损伤性影响,实现对电网运营环境的合理维护,提出基于改进随机森林算法区域短期售电量预测方法。根据随机森林的决策树构建条件,分别计算电压精度数值与电流精度数值,完成基于改进随机森林算法的售电量估算。在此基础上,规划区域电网的布局形式,通过合理安排售电行为的方式,预测具体的电网扩容量计算结果,完成区域短期售电量预测方法的设计与应用。实例分析结果表明,在改进随机森林算法作用下,区域短期售电量预测数值始终不会出现超过其额定预测承载上限的情况,能够对电网运营环境起到较强的保护作用。In order to avoid the damaging impact on the regional power grid caused by excessive short⁃term electricity sales,and to achieve reasonable maintenance of the power grid operating environment,a regional short⁃term electricity sales forecasting method based on the improved random forest algorithm is proposed.According to the decision tree construction conditions of the random forest,the voltage accuracy value and the current accuracy value are respectively calculated,and the electricity sales estimation based on the improved random forest algorithm is completed.On this basis,the layout form of the regional power grid is planned,and the specific grid capacity expansion calculation results are predicted by reasonable arrangement of power sales behavior,and the design and application of the regional short⁃term electricity sales forecasting method are completed.The case analysis results show that,under the effect of the improved random forest algorithm,the regional short⁃term electricity sales forecast value will never exceed its rated forecast load upper limit,which can have a strong impact on the grid operation environment.

关 键 词：随机森林 短期售电量 决策树 电压精度 电流精度 电网扩容 

分 类 号：TN-9[电子电信]