风电机组监测数据可视分析研究

发布时间：2018-02-14 19:33

本文关键词： 多维数据风电机组监测数据多视图协同可视分析可视分析　出处：《东北电力大学》2016年硕士论文　论文类型：学位论文

【摘要】：近年来,风力发电迅猛发展,风电场现场监测装置和通信装置越来越先进,风电机组监测数据量呈指数级增长,庞大的风机监测数据量使得风电场运行管理人员“一头雾水”,产生了“数据海量,知识匮乏”的问题。研究能将风电机组运行状态信息以更直观更迅速的方式显示出来并及时地将异常数据醒目地展现出来的可视化方法,有利于风电场运行管理人员从整体上把握数据,及时发现风电机组各个装置的潜在隐患信息,对保障风电机组安全稳定地运行具有举足轻重的意义。本文的目的在于结合风电机组监测数据时序、多维、快速的特点,提出直观有效的可视化方法,从而进行有效的数据分析。为此,针对目前风电机组监测数据可视化存在的直观性不强和交互性差的问题,分别从数据预处理,增强可视化的直观性与增强可视化的交互性三个方面进行了研究。首先,本文提出了基于随机森林与主成分分析的预处理方法,对风电机组监测数据进行了可视化前的预处理,以获得良好的数据分布结构,达到辅助可视化的目的。其次,在增强可视化的直观性方面,本文提出了散点图矩阵可视化和平行坐标图可视化,首先通过散点图矩阵获得各个属性的大致相关性,然后由平行坐标图展现整个数据集的全貌及属性相关的具体情况,提高了可视化的直观性。最后,在增强可视化的交互性方面,本文提出了多视图协同可视化方法,在平行坐标图中融合散点图、散点图矩阵,通过Overview+Detail交互方式及基于平行坐标的交互技术来实现多视图之间的协同可视分析,克服了单一可视化方法的缺陷,综合了多个可视化方法的优点,增强了可视化的交互性。基于以上可视化方法,本文给出了风电机组监测数据的可视化应用结果,结合风电机组监测数据时序、多维、快速的特点,进一步展开了可视分析,通过图形图像的方式将监测数据的属性相关性、数据集中特性及分布规律进行了直观地展现及分析。
[Abstract]:In recent years, with the rapid development of wind power generation, the field monitoring devices and communication devices of wind farms are more and more advanced, and the monitoring data of wind turbines are increasing exponentially. The huge amount of fan monitoring data makes the managers of wind farm operation "confused", resulting in "huge amount of data." The problem of lack of knowledge. This paper studies the visualization method that can display wind turbine operating state information in a more intuitive and rapid manner and display the abnormal data in a timely manner. It is beneficial to the wind farm operation management personnel to grasp the data as a whole and to find out the potential hidden trouble information of each wind turbine unit in time. The purpose of this paper is to combine the monitoring data sequence of wind turbine with the characteristics of multidimensional and fast, and to put forward an intuitive and effective visualization method to ensure the safe and stable operation of wind turbine. In order to carry out effective data analysis, aiming at the problems of visualization of monitoring data of wind turbine units, which are not intuitive and poor in interaction, the data preprocessing is carried out separately from the point of view of data preprocessing. In this paper, three aspects of visual visualization and visualization interaction are studied. Firstly, a preprocessing method based on stochastic forest and principal component analysis (PCA) is proposed to preprocess the monitoring data of wind turbine before visualization. In order to obtain a good data distribution structure and achieve the purpose of auxiliary visualization. Secondly, in the aspect of enhancing visualization visualization, this paper proposes scatter plot matrix visualization and parallel coordinate graph visualization. Firstly, the correlation of each attribute is obtained by scatter plot matrix, and then the whole picture of the whole data set and the specific condition of attribute correlation are presented by parallel coordinate graph, which improves the visualization intuition.Finally, In order to enhance the interaction of visualization, this paper proposes a method of multi-view collaborative visualization, which combines scattered plot, scatter plot matrix in parallel coordinate graph. The cooperative visual analysis among multiple views is realized through Overview Detail interaction and interaction technology based on parallel coordinates, which overcomes the defects of single visualization method and integrates the advantages of multiple visualization methods. Based on the above visualization methods, this paper gives the visual application results of wind turbine monitoring data, combined with the time series, multidimensional and fast characteristics of wind turbine monitoring data, the visual analysis is further carried out. The attribute correlation, the characteristics of data set and the distribution law of the monitored data are displayed and analyzed intuitively by the way of graph and image.
【学位授予单位】：东北电力大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TM315

【相似文献】