基于Hadoop的电力设备状态监测数据处理与分析的研究

发布时间：2018-03-03 23:28

本文选题：智能电网　切入点：状态监测　出处：《华北电力大学》2014年硕士论文　论文类型：学位论文

【摘要】：随着智能电网的逐渐发展和完善，对其环境下的电力设备运行的安全性和稳定性要求越来越高，设备监测力度越来越大，电力设备种类多、数量大，在极端情况下（大雾天、阴雨天），短时间内就会累计大量的监测数据，早已超出现有监测方法的处理能力；如何利用这些海量的监测数据及时准确的预测电力设备的状态，已成为新的研究课题。Google提出云计算理念后，云计算平台Hadoop异军突起，具有海量存储与运算、高扩展性和高可靠性等优势，为解决上述状态监测问题提供了可能。本文首先详细归纳和总结了电力设备状态监测方法的研究现状，重点分析了现有方法在处理效率和处理准确率方面所存在的问题；随后对云计算Hadoop平台的应用情况和发展现状进行汇总，并对Hadoop平台各大软件进行了详细的介绍。接下来又说明了本文要解决的问题是如何及时地存储短时间内激增的设备监测数据，并利用这些海量监测数据对电力设备的状态进行高效、及时、准确的评估；通过对Hadoop平台的特性进行分析，确定使用Hadoop/HBase平台存储短时间内激增监测数据，并利用MapReduce实现自行设计的状态评估算法进行电力设备的状态评估。分析现有状态评估方法的问题和当前智能电网的应用需求，以及Hadoop平台的相关技术优势，针对智能电网对电力设备状态监测的高要求，提出了一种基于Hadoop平台的状态评估系统，并对系统总体架构、各子模块结构、所使用的算法、算法实现进行了详细的介绍，解决了现有评估方法效率低、实时性差的问题。文中详细介绍了Hadoop集群搭建方法、集群的调优、MapReduce程序设计的技巧、状态评估算法、及状态评估算法的MapReduce实现方法，并模拟了现实环境，对MapReduce算法进行了实时性和准确率进行了评估；尤其是算法实时性的测试，，证明了在电力设备状态监测各类应用环境中相关设计的有效性和正确性。本课题为利用海量监测数据进行电力设备状态评估和预测提供了新的思路；是云计算Hadoop平台在智能电网领域的一次有益尝试。
[Abstract]:With the development and perfection of smart grid, the security and stability of power equipment in its environment are higher and higher. The monitoring of power equipment is more and more powerful, and the variety and quantity of power equipment are large. On rainy and cloudy days, a large amount of monitoring data will accumulate in a short period of time, which has already exceeded the processing capacity of existing monitoring methods. How to use these massive monitoring data to predict the state of power equipment in a timely and accurate manner, The cloud computing platform Hadoop has the advantages of mass storage and operation, high scalability and high reliability, which provides the possibility to solve the above problems. Firstly, this paper summarizes the research status of power equipment condition monitoring methods in detail, and analyzes the existing problems in processing efficiency and processing accuracy. Then the application and development of cloud computing Hadoop platform are summarized. The software of Hadoop platform is introduced in detail. Then the problem is how to store the equipment monitoring data in a short period of time. And using these massive monitoring data to evaluate the state of power equipment efficiently, timely and accurately, through the analysis of the characteristics of Hadoop platform, it is determined to use Hadoop/HBase platform to store the surge monitoring data in a short time. The state evaluation algorithm designed by MapReduce is used to evaluate the state of power equipment. This paper analyzes the problems of current state assessment methods and the current application requirements of smart grid, as well as the relative technical advantages of Hadoop platform, aiming at the high requirements of smart grid for power equipment condition monitoring. This paper presents a state assessment system based on Hadoop platform, and introduces in detail the system architecture, each sub-module structure, the algorithms used and the algorithm implementation, which solves the problems of low efficiency and poor real-time performance of the existing evaluation methods. This paper introduces the method of Hadoop cluster construction, the skills of optimizing MapReduce program, the algorithm of state evaluation, and the method of MapReduce implementation of state evaluation algorithm, and simulates the real environment. The real-time and accuracy of the MapReduce algorithm are evaluated, especially the real-time test of the algorithm, which proves the validity and correctness of the related design in various application environments of power equipment condition monitoring. This paper provides a new idea for power equipment state evaluation and prediction using massive monitoring data, and is a useful attempt of cloud computing Hadoop platform in the field of smart grid.
【学位授予单位】：华北电力大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TM507

【相似文献】