基于数据挖掘的电力负荷特性与用户分类研究

发布时间：2018-04-15 09:26

本文选题：负荷特性分析 + 负荷波动指标　；参考：《北京邮电大学》2016年硕士论文

【摘要】：近年来,随着经济的高速发展,电力作为国家经济发展的基础保障,社会各方面对电力的需求日益增大,已经与当前电力的总供应严重失衡。而且现阶段电力系统整体的电能量利用率不断下滑,电力用电集中在尖峰时间段,电量供应不足,区域性、季节性缺电现象时有发生,而平谷时间段电力需求较小,发电量剩余现象严重,日负荷峰谷差持续增大,电力供需矛盾尖锐。因此,考虑当前形势下的电力系统特点,对电力负荷特性进行深入分析,并在此基础上开展负荷管理工作,实行峰谷分时电价政策,对提高电能量利用率,优化电网用电结构,具有十分重要的意义。目前全国电力用户已达到相当大的规模,如果对这些用户都进行一对一的负荷特性分析不具备可行性。因此,电力用户分类是电力系统负荷特性分析的基础和关键环节。本文提出一种多角度聚类的电力用户分类方法。首先根据深圳市行业性质以及行业用电量百分比,选取9个典型行业。然后从负荷波动指标和实时负荷曲线特征两个角度,分别对深圳市9个典型行业的电力用户进行聚类分析,使具有相似或相同用电指标或特征模型的电力用户归为一类。根据其负荷波动指标聚类结果得出每个行业的电力用户的负荷波动指标特征;根据实时负荷曲线聚类结果得出每个行业的电力用户的用电特征模型,建立深圳市典型行业电力用户负荷曲线特征模型数据库,并对已得到的电力用户负荷曲线特征模型进行校验验证,实现长期跟踪校验、修正,最终得出适用性更强的电力用户负荷曲线特征模型。经实践验证,本文所得到的深圳市典型行业电力用户用电特征模型与用户实际的用电特征相似度较高,有较高的准确性。电力企业可以利用电力用户的用电特征模型对电力用户进行分类,并且进一步研究其各类别用户的用电趋势,进行相关负荷预测、管理规划等工作。
[Abstract]:In recent years, with the rapid development of economy, electric power as the basic guarantee of national economic development, the demand for electricity in all aspects of society is increasing day by day, which has been seriously out of balance with the current total supply of electricity.And at the present stage, the overall utilization rate of electricity in the power system is declining, the electricity consumption is concentrated in the peak period, the electricity supply is insufficient, the regional and seasonal electricity shortage occurs from time to time, while the power demand in the Pinggu period is relatively small.The phenomenon of electricity generating surplus is serious, the daily load peak and valley difference increases continuously, and the contradiction between power supply and demand is sharp.Therefore, considering the characteristics of power system under the current situation, this paper makes a thorough analysis of the power load characteristics, and on this basis, carries out the load management work, implements the peak-valley time-sharing price policy, improves the utilization ratio of electricity energy, and optimizes the power consumption structure of the power grid.It is of great significance.At present, the national electric power users have reached a considerable scale, if the one-to-one load characteristic analysis of these users is not feasible.Therefore, power user classification is the basis and key link of power system load characteristic analysis.This paper presents a multi-angle clustering method for power user classification.Firstly, according to the nature of Shenzhen industry and the percentage of electricity consumption, 9 typical industries are selected.Then, from the point of view of load fluctuation index and real-time load curve feature, the electric power users in 9 typical industries in Shenzhen are clustered and analyzed, which makes the electric power users with similar or the same power consumption index or characteristic model as a class.According to the clustering results of load fluctuation index, the load fluctuation index characteristics of power users in each industry are obtained, and the power consumption characteristic model of each industry is obtained according to the clustering results of real-time load curve.The database of power user load curve characteristic model for typical industry in Shenzhen is established, and the obtained power user load curve characteristic model is verified and verified to realize long-term tracking calibration and correction.Finally, a more applicable power user load curve characteristic model is obtained.It is proved by practice that the electricity consumption feature model of typical industry in Shenzhen has high similarity and high accuracy with the actual power consumption characteristics of the users.The electric power enterprise can classify the power users by using the power user's characteristic model, and further study the power consumption trend of each kind of users, carry on the related load forecast, manage the plan and so on.
【学位授予单位】：北京邮电大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TP311.13;TM714

【相似文献】