计及负荷特性的在线电压稳定性研究-电力系统及其自动化专业论文.docxVIP

  • 4
  • 0
  • 约5.41万字
  • 约 67页
  • 2019-05-03 发布于上海
  • 举报

计及负荷特性的在线电压稳定性研究-电力系统及其自动化专业论文.docx

明i摆摆出 摘要 摘要 随着电压失稳和电压崩溃事故的频发以及电力需求的不断增长,电力系统 的运行越来越接近其极限状态,电压稳定问题日益突出。同时,电压稳定问题 又是负荷稳定问题,因此研究负荷特性对电压稳定问题的影响显得尤为重要。 本文主要研究了三种不同负荷模型下的在线电压稳定性问题。 一是应用连续潮流得出各种运行状态下的电压稳定裕度,通过设计一个三 层的神经网络模型,从而把节点负荷的变化句电压稳定裕度之间建立一种映射 关系。神经网路模型的输入向量应用回归分析进行特征选择,使得输入维数减 少,既加快了训练速度,又提高了训练精度,克服了运用连续潮流在钱计算慢 的缺点。 二是提出了一种基于时间断面上的周部等值方法,在所要研究的节点处, 把电力网络周部等效成两节点系统,利用 PMU 测得的电压和电流相量,求出周 部等值模型的参数。负荷采用静态 ZIP 模型,最后推导出了节点电压稳定指标 表达式,此指标无需全系统外部等值,可提高准确度和计算速度。 三是应用连续潮流追踪系统的平衡解流形,在获得的平衡点处,建立由感 应电动机并联阻抗的动态负荷模型,推导出了一个简单电力系统电压稳定分析 的微分代数方程组模型,分析了计及负荷动态特性的小干扰电压稳定性,并推 导了电压稳定动态指标。 关键词:电力系统:电压稳定:静态负荷:动态负荷:电压指标 II II ABSTRACT ABSTRACT With the acciden臼 of voltage in阳bility have been occurred frequently and the growing demand in electricity二 power system operation is more closer ωits critical state,voltage stabi1ity problems become to be increasingly serious,也.e voltage 剖ability problem is getting attention. Meanwhile ,the voltage s阳.bi1ity problem is also ca11ed load stability,the study of load characteristics on 也e impact of voltage 剑abi1ity problem is particularly important. In创s paper,0卧line voltage 阳.bi1ity problems have been researched based on 也ree different load models. Voltage stability margin, a11 kinds of operation state,has been ca1culated by continuous power flow. Then a three-layer neural network model has been des阳led, which establishes a mapping between the changes of node load and voltage stability margin. The input vectors of neura1 network model are used the method of regression analysis for feature selection, in 由副 reduces 址le dimensions of input in purpose of speeding up 由.e training and improving the. training precision, which overcomes the disadvantage of onlinecalculation is slowing by continuous power flow. A loca1 equiva1ent method which is based on time section is presented. For a researched node,白le power grid is equiva1ent to a two-node system which is composed of a load and a power generation, works out the loca1 equiva1ent model p町ameters by using PMU

您可能关注的文档

文档评论(0)

1亿VIP精品文档

相关文档