基于大数据算法的输电线路故障分析研究

发布时间：2018-03-16 13:22

本文选题：大数据分析　切入点：输电线路　出处：《湖北工业大学》2017年硕士论文　论文类型：学位论文

【摘要】：随着现代输电网的规模、容量和覆盖范围越来越大,输电线路在国民经济中占的地位越来越重要,电网停电故障将会给社会生产和人民生活,造成重大的经济损失。源于输电线路具有运送距离长、覆盖区域广等特点,容易受天然环境及人为成分的影响较多,致使线路的运作维护工作存在较大的困难。怎样有效提高输电线路的运行和维护的质量,以确保电网稳定安全运行,现已成为广大研究人员及电力部门努力探讨的课题。。近几年来,当代信息社会已经进入了大数据时期,大数据火速发展成长为深受学术界和产业界眷顾的热门范畴,并得到了大量的运用。大数据分析方式可以从输电系统海量的信息中,找出潜藏的模式和其中的规则,为维护职员供给相关维护策略支持。基于大数据算法的输电线路故障分析作为调度人员处理事件的协助工具,能够起到降低事故处理的耗时,防止事故扩张的作用。首先,本文分析了输电线路的常见故障,大数据分析方法的应用领域,及目前输电线路智能化诊断中运用的大数据算法和其不足。为利用大数据算法解决输电线路故障分析问题,提供了丰富的理论依据和新的思路。其次,在现有研究成果的基础上,本文设计了输电线路故障分析模型,根据k最近邻分类算法(k-Nearest Neighbor,KNN)的特点,结合模糊理论,提出了模糊KNN算法模型,解决了KNN在处理类别界限不明显数据时的问题。实验结果表明,模糊KNN算法模型提升了KNN故障类型诊断分析的准确率。同时,本文基于Spark计算平台,开发了输电线路实时故障分析模型,满足了电网实时故障分析的需求。然后,本文针对模糊KNN算法对于混合故障大数据类别分析准确率低的问题,对输电线路故障分析模型进行优化,研发了基于密度逻辑回归的多分类模型(Multi-Classification based on Density logistic regression,MCDLR),大大提升了现有输电线路故障数据分类方法的准确性。最后,本文进行了大量的输电线路故障分析实验,充分证实了文中提出的算法模型的可行性和有效性。本文提出的算法模型不仅能够满足输电线路故障分析的需求,而且实现了大数据分析算法在电力系统的应用研究,具有良好的创新性和实践可行性。
[Abstract]:With the scale, capacity and coverage of modern transmission network increasing, transmission lines play a more and more important role in the national economy. Power outages will bring social production and life to the people. Because transmission lines are characterized by long transportation distance and wide coverage, they are vulnerable to the influence of natural environment and human composition. The operation and maintenance of transmission line is difficult. How to improve the quality of operation and maintenance of transmission line effectively in order to ensure the stable and safe operation of power network, In recent years, the contemporary information society has entered the period of big data, and big data has rapidly developed into a hot category deeply favored by academia and industry. Big data can find hidden patterns and rules from the vast amount of information in the transmission system. Transmission line fault analysis based on big data algorithm, as a tool to assist dispatchers in handling incidents, can reduce the time consuming and prevent the expansion of accidents. In this paper, the common faults of transmission line, the application field of big data's analysis method, the big data algorithm used in intelligent diagnosis of transmission line and its shortcomings are analyzed. It provides abundant theoretical basis and new ideas. Secondly, based on the existing research results, this paper designs a fault analysis model for transmission lines, combining with fuzzy theory, according to the characteristics of k-nearest neighbor classification algorithm (KNN). A fuzzy KNN algorithm model is proposed, which solves the problem of KNN in dealing with unobvious data of class boundaries. The experimental results show that the fuzzy KNN algorithm model improves the accuracy of fault type diagnosis and analysis of KNN. At the same time, based on the Spark computing platform, this paper proposes a fuzzy KNN algorithm model. The real-time fault analysis model of transmission line is developed to meet the need of real-time fault analysis of power network. Secondly, the fuzzy KNN algorithm is used to analyze the big data class of hybrid faults with low accuracy. Based on the optimization of transmission line fault analysis model, a multi-classification based on Density logistic regression model based on density logic regression is developed, which greatly improves the accuracy of existing fault data classification methods for transmission lines. In this paper, a large number of transmission line fault analysis experiments have been carried out, which fully verify the feasibility and effectiveness of the proposed algorithm model, which can not only meet the needs of transmission line fault analysis. Moreover, the application of big data analysis algorithm in power system is realized, which has good innovation and practical feasibility.
【学位授予单位】：湖北工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP311.13;TM755

【参考文献】