摘要
针对高比例新能源电网,提出一种计及风—光—荷特征分类和曲线综合相似度指标的多源—荷联合典型场景提取方法。从电力系统电力电量平衡角度出发,选取风电日发电占比、高峰时段风电最小出力、低谷时段风电最大出力、风电出力变异系数、光伏日发电占比、光伏出力变异系数、负荷率7项指标对新能源出力和负荷功率原始数据进行转化,并构建特征评价矩阵。采用具有噪声的基于密度的聚类方法(DBSCAN)&k-medoids联合算法对风—光—荷特征评价矩阵进行聚类。提出一种基于动态时间规划(DTW)距离的曲线综合相似度指标,在聚类集群对应的原始数据中提取多源—荷联合典型运行场景。基于甘肃某地实际电网数据进行仿真,验证所提出的方法的有效性。
Aimming at the high proportion of new energy grid,a multi-source-load joint typical scenario extraction method considering the wind-photovoltaic-load feature classification and curve comprehensive similarity index is proposed.From the point of power system power balance,this paper selects wind power daily generation proportion,peak hours wind power minimum output,valley hours wind power maximum output,variation coefficient of wind power output,photovoltaic daily power generation proportion,variation coefficient of photovoltaic output and load rate,to transform the original data of new energy output and load power,and constructs a characteristic evaluation matrix.The density-based spatial clustering applications with noise(DBSCAN)&k-medoids joint algorithm is used to cluster the evaluation matrix of wind-solar-load features.A curve comprehensive similarity index based on dynamic time warping(DTW)distance is proposed,and the typical scenario of multi-source-load joint operation are extracted from the original data corresponding to the clustered sets.Based on the actual power grid data in Gansu,the simulation results verify the effectiveness of the proposed method.
作者
许青
徐铭
刘正英
王小文
尹柏祥
李天寿
XU Qing;XU Ming;LIU Zhengying;WANG Xiaowen;YIN Baixiang;LI Tianshou(State Grid Gansu Electric Power Company,Lanzhou 730030,China;Tianshui Power Supply Company of State Grid Gansu Electric Power Company,Tianshui 741000,China)
出处
《微型电脑应用》
2026年第1期104-108,共5页
Microcomputer Applications
基金
国网甘肃省电力公司咨询服务项目(SGGSJY00JYWT2400033)。
