基于LightGBM的低压配电网台区线损率估算方法  

Line loss estimation method of low-voltage distribution transformer area based on LightGBM

作　　者：于艾彤 YU Aitong(School of Electronic,Information and Electrical Engineering,Shanghai Jiao Tong University,Shanghai 200240,China)

机构地区：[1]上海交通大学电子信息与电气工程学院,上海200240

出　　处：《电气应用》2023年第11期63-69,共7页Electrotechnical Application

摘　　要：针对低压配电网台区线损数据统计难度大、线损率计算结果准确度低的现状,提出一种低压配电网台区线损率计算方法。以平均电流法理论线损计算方法为基础,确定用于训练模型的特征,依托电网状态监测系统、数据统计平台,提取模型样本数据并建立数据集,通过以LightGBM为框架的梯度提升算法训练机器学习模型,并将训练得到的模型用于低压配电网台区线损率估算。采用某地区内台区线损率数据对模型进行验证,实验结果表明,运用LightGBM算法训练的模型可实现对低压配电网台区理论线损率的拟合估算,且模型具有较强的泛化能力,测试集平均绝对误差为0.78,均方误差为1.35,决定系数为0.52。与目前常用的回归算法模型相比,可有效提升计算结果准确度及拟合效果。Aiming at the current situation that the line loss data of the low-voltage distribution transformer area is difficult to statistics and the accuracy of calculation results is low,a method for calculating the line loss rate of low-voltage distribution transformer area is proposed.Features used to train the model are selected based on the theoretical line loss calculation method named average current method,and the sample data is extracted to establish the data set by relying on the power grid monitoring system and data statistics platform,the machine learning model is trained by the gradient boosting algorithm framed by LightGBM,finally use the model to estimate the line loss rate of low-voltage distribution transformer area.The experimental results show that the model trained by LightGBM can realize the fitting and estimation of the theoretical line loss rate of the low-voltage distribution transformer area.It has strong generalization ability.The mean absolute error(MAE)of the test set is 0.78,the mean squared error(MSE)is 1.35,and the R-Square(R2)is 0.52,which can effectively improve the accuracy and fitting effect compared with the commonly used regression algorithm model.

关 键 词：低压配电网 理论线损计算 梯度提升树 平均电流法 回归算法 回归模型评价指标 

分 类 号：TM714.3[电气工程—电力系统及自动化]