基于运行数据的风电场动态等值建模及故障预警研究

发布时间：2018-02-02 23:06

本文关键词： 运行数据风电场相关性分析故障预警动态等值　出处：《华侨大学》2017年硕士论文　论文类型：学位论文

【摘要】：近年来,随着风电场的规模急剧扩大,确保风电场的安全运行显得越来越重要。本文基于SCAD A系统监测运行数据,从时间和空间的角度出发对风电场的动态等值建模及故障预警进行研究,为减少风电场因事故性机组停机次数和进行事故后的故障原因分析提供理论指导。首先,基于风电场安装的SCAD A系统,对风电场内各机组的监测项目进行相关性分析,直观地挖掘出风电机组子系统和元件间的交互耦合性,进而提出风电机组运行状态间存在着动态灰色关联性的观点,并对数据样本的跨度进行确定。相关性分析所得的结论为风电机组的故障预警和风电场动态等值建模提供依据。其次,基于机组子系统和元件间存在的交互耦合关系以及机组发生事故前的临界状态特性,将动态网络标记引入风电机组故障预警研究中。结合回归支持向量机和动态网络标记,对风电机组的监测运行数据进行降噪处理后,通过机组运行过程中动态网络标记值的变化对机组故障发生前的临界转换点与故障可能发生的部位进行追踪。该方法能够全面把握机组运行状态的发展变化趋势,为制定合理的检修计划、保障机组安全运行提供可靠的技术支持。最后,基于风电机组运行状态间存在的动态灰色关联性,将灰色关联分析理论引入风电场动态等值建模研究中。结合动态灰色关联分析方法和聚类算法,构造一个动态灰色关联度矩阵G并将其作为机群聚类指标,将机组运行状态间的动态灰色关联性融入到动态等值模型中。仿真结果证明,所建立的风电场动态等值模型与详细模型较接近,能够准确地反映风电场并网点的动态特性,适用于风电并网故障的暂态稳定分析。本文提出的基于运行数据的风电场动态等值建模及故障预警研究方法,完全基于风电场的实测运行数据,可靠真实,且直观简单,具有一定的推广价值。
[Abstract]:In recent years, with the rapid expansion of the scale of wind farm, to ensure the safe operation of the wind farm becomes more and more important. In this paper, SCAD A monitoring system based on operation data, time and space from the perspective of dynamic equivalent modeling and fault prediction of wind farm are studied. The analysis provides theoretical guidance to reduce the failure reasons for accidents of wind farm the shutdown times and after the accident. First of all, the SCAD A system installed in wind farm based on the correlation analysis of monitoring items of each generator in wind farm, directly mining interaction coupling of wind turbine system and the element, and then put forward the operation state of wind turbine dynamic grey correlation exists between point of view, and the data sample span were determined. Correlation analysis results provide the basis for modeling the equivalent fault warning and wind farm dynamic wind turbine. The time base In the unit subsystem and component interaction and coupling relationship between unit critical state characteristics before the accident, the dynamic network markers in wind turbine fault warning research. Combined with support vector machine and dynamic network marker regression, on the wind turbine operation monitoring data for noise reduction, critical transition point of the dynamic changes of network fault markers during normal operation and fault value before the position of possible tracking. This method can fully grasp the development trend of the change of the running status of the unit, is the formulation reasonable overhaul plan, provide reliable technical support to protect the safe operation of the unit. Finally, a dynamic grey correlation operation state of wind the power unit based on grey relational analysis theory into the research on equivalent modeling of wind farm dynamic. Combined with dynamic grey correlation analysis method and poly Algorithm, constructed a dynamic grey correlation matrix G as cluster clustering index into dynamic grey correlation between the running status of the unit to the dynamic equivalent model. The simulation results show that the dynamic equivalent model of wind farms and the detailed model is close to that can accurately reflect the dynamic characteristics of wind farm network, transient stability analysis for wind power grid fault. Wind farm dynamic research on equivalent modeling and fault prediction method based on the operation data in this paper, fully measured operation data of wind farm based on real, reliable, simple and intuitive, and has a certain popularization value.

【学位授予单位】：华侨大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TM614;TM743

【参考文献】