摘要
针对电力系统中出现的电力量测数据缺失的问题,本文采用统计方法、插值方法和机器学习方法进行了研究和实践。首先,本文分析了电力量测数据缺失的原因,重点探讨了量测数据在采集、传输、存储以及其他环节对数据缺失的影响。接着,本文详细介绍和分析了三种量测数据缺失补齐方法,并对不同方法进行了实验评估,包括相关系数评价、拟合优度评价和平均绝对误差占比评价等多种评价方法。实验结果表明,机器学习方法在量测数据缺失补齐精度和效果方面优于其他两种方法,表现出更好的效果。最后,本文对研究结果进行了总结和展望,指出机器学习方法在电力量测数据缺失补齐中的应用前景,本文的研究成果可为电力系统中量测数据缺失处理提供一定的参考价值。
In response to the problem of missing power measurement data in the power system,this paper conducts research and practical applications using statistical methods,interpolation methods,and machine learning techniques.Firstly,the paper analyzes the reasons for missing power measurement data,with a focus on exploring the impact of data collection,transmission,storage,and other processes on data loss.Next,three methods for completing missing measurement data are detailed and analyzed.Different methods are experimentally evaluated using various assessment metrics,including correlation coefficient evaluation,goodness of fit evaluation,and percentage of mean absolute error evaluation,among others.The experimental results indicate that machine learning methods outperform the other two methods in terms of accuracy and effectiveness in completing missing measurement data,demonstrating superior performance.Finally,the paper summarizes and looks ahead at the research findings,pointing out the application prospects of machine learning methods in completing missing power measurement data.The research outcomes of this paper can provide a certain reference value for handling missing measurement data in power systems.
作者
陆嘉铭
奚增辉
瞿海妮
许唐云
姚嵘
屈志坚
LU Jiaming;XI Zenghui;QU Haini;XU Tangyun;YAO Rong;QU Zhijian(State Grid Shanghai Municipal Electric Power Company,Shanghai 200122,China)
出处
《电力大数据》
2023年第7期40-49,共10页
Power Systems and Big Data
关键词
电力量测数据
数据缺失
线性插值
随机森林
神经网络
electric power measurement data
data missing
linear interpolation
random forest
neural network