中国电力 ›› 2013, Vol. 46 ›› Issue (10): 29-34.DOI: 10.11930/j.issn.1004-9649.2013.10.29.5

• 电网 • 上一篇    下一篇

基于模糊负荷聚类的不良负荷数据辨识与修正

刘辉舟1, 周开乐2, 3, 胡小建2, 3   

  1. 1.国网安徽省电力公司 铜陵供电公司, 安徽 铜陵 244000; 2. 合肥工业大学 管理学院, 安徽 合肥 230009; 3. 合肥工业大学 过程优化与智能决策教育部重点实验室,安徽 合肥 230009
  • 收稿日期:2013-05-20 出版日期:2013-10-23 发布日期:2015-12-10
  • 作者简介:刘辉舟(1985—),男,安徽桐城人,硕士,助理工程师,从事电力负荷预测与电力信息化研究工作。
  • 基金资助:
    国家高技术研究发展计划(863计划)资助项目(2011AA05A116); 国家自然科学基金资助项目(71071045)

Bad Data Identification and Correction Based on Load Clustering by FCM Algorithm

LIU Hui-zhou1, ZHOU Kai-le2, 3, HU Xiao-jian2, 3   

  1. 1. Tongling Power Supply Company, State Grid Anhui Electric Power Corporation, Tongling 244000, China; 2. School of Management, Hefei University of Technology, Hefei 230009, China; 3. Key Laboratory of Process Optimization and Intelligent Decision-Making, Ministry of Education, Hefei University of Technology, Hefei 230009, China
  • Received:2013-05-20 Online:2013-10-23 Published:2015-12-10

摘要: 为辨识和修正不良负荷数据,在利用模拟退火遗传算法优化的模糊C均值(Fuzzy C-Means,FCM)算法进行负荷曲线聚类的基础上,提出将待测曲线与相应特征曲线进行比较计算差量系数的方法。差量系数大于电力公司确定的阈值的负荷点即为不良负荷数据。通过算例验证表明,该方法克服了统计历史数据中不良数据的影响,提高了不良数据辨识的可操作性和实用性。同时提出了考虑不良数据测量点外所有其他测量点负荷信息的不良数据修正方法,与仅考虑不良数据测量点前后2个测量点负荷信息的修正方法相比,提高了不良数据修正的精确性和有效性。

关键词: 不良负荷数据, 辨识与修正, 负荷曲线聚类, 模糊C均值算法

Abstract: In order to identify and correct the bad load data, the load profiles are clustered by using simulated annealing genetic algorithm optimized fuzzy C-means algorithm(FCM). Based on the threshold of differential coefficient which is determined by the comparison of test load profiles with its typical load profile, the bad data whose differential coefficient is greater than the threshold value is identified. A numerical case study demonstrates that this method overcomes the impact of bad data in the statistical historical data,and as a result improves the operability and practicality of bad data identification.A new bad data correction method is presented, which takes all the measurement points load information into consideration. Compared with the correction method which considers only the load information of two points before and behind the bad data measurement point, this method improves the accuracy and effectiveness of the bad data correction.

Key words: bad data, identification and correction, load curve clustering, fuzzy C-means algorithm

中图分类号: