基于改进模糊聚类分析的电力系统不良数据辨识

发布时间：2018-04-16 10:22

本文选题：电力系统 + 不良数据检测辨识　；参考：《东北石油大学》2017年硕士论文

【摘要】：当前电力行业发展迅速,电力系统的规模也在不断扩大,随着电力网络的结构和运行模式更加复杂化,电网数据传输中产生错误数据的概率越来越高,由于客观原因的存在而产生的的错误数据又被称为不良数据,由于不良数据无法正确显示系统的运行状况,其存在将严重影响状态估计的准确性与可靠性,进而对电力系统的安全稳定运行产生不利的影响。电力系统不良数据检测辨识的目的在于去除量测数据中的不良数据,为电力系统状态估计提供准确的数据。目前使用的不良数据检测辨识的方法主要是基于量测数据残差的方法,随着电网结构的复杂化,该方法检测结果存在漏检与误检的弊端日益凸显。本文采用基于模糊聚类的方法实现将原始量测中的良性数据与不良数据分离,算例仿真结果显示了该方法相比较传统方法的优越性。本文首先介绍了电力系统不良数据检测辨识的知识背景和研究现状,对比分析了当前不良数据检测辨识方法的优缺点。针对不良数据间存在相关性而容易出现残差污染和残差淹没的情况,传统的残差检测方法表现不佳,于是本文提出使用EGSA-FCM算法实现不良数据的检测辨识,该方法以模糊聚类算法中的模糊C均值算法为基础,通过引入本文提出的增强型万有引力搜索算法实现对SCADA系统上传的量测数据进行前期搜索,该方法提高了计算效率和准确性。最后将用于聚类有效性判断的COS指标应用于对最佳聚类数目的判定,获得最佳聚类结果,通过已知良性数据所在聚类,最终得到量测数据中良性数据与不良数据的分类。针对基于EGSA-FCM算法的不良数据检测辨识方法,本文编制了检测辨识程序。将该方法应用IEEE14节点电力系统和某地区电网变区中,检测辨识结果表明本文所提出的方法与传统的检测辨识方法相比有效避免了误检和漏检的发生,检测结果更加准确。
[Abstract]:With the rapid development of power industry and the expansion of power system scale, with the complexity of power network structure and operation mode, the probability of generating wrong data in power network data transmission is higher and higher.The wrong data caused by the existence of objective reasons is also called bad data. Because the bad data can not correctly display the operating status of the system, its existence will seriously affect the accuracy and reliability of the state estimation.Furthermore, it has an adverse effect on the safe and stable operation of power system.The purpose of power system bad data detection and identification is to remove the bad data from the measurement data and to provide accurate data for power system state estimation.At present, the method of detection and identification of bad data is mainly based on the residual of measured data. With the complexity of power network structure, the defects of the detection results of this method are increasingly prominent.In this paper, the method based on fuzzy clustering is used to separate the benign data from the bad data in the original measurement. The simulation results show that this method is superior to the traditional method.This paper first introduces the knowledge background and research status of power system bad data detection and identification, and compares and analyzes the advantages and disadvantages of current bad data detection and identification methods.Because of the correlation between bad data and the possibility of residual pollution and residual inundation, the traditional residual detection method is not good, so this paper proposes to use EGSA-FCM algorithm to detect and identify bad data.This method is based on the fuzzy C-means algorithm of fuzzy clustering algorithm. By introducing the enhanced universal gravity search algorithm proposed in this paper, the pre-search of the measurement data uploaded by SCADA system is realized.This method improves the calculation efficiency and accuracy.Finally, the COS index, which is used to judge the clustering validity, is applied to the determination of the best clustering number, and the best clustering result is obtained. Finally, the classification of the benign data and the bad data in the measured data is obtained through the known clustering of the benign data.Aiming at the bad data detection and identification method based on EGSA-FCM algorithm, a detection and identification program is developed in this paper.The method is applied to the IEEE14 node power system and a region power system. The detection and identification results show that the method proposed in this paper is more effective than the traditional detection and identification methods to avoid the occurrence of false detection and miss detection, and the detection results are more accurate.
【学位授予单位】：东北石油大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TM732

【参考文献】