【作者】 李峰；

【导师】 汤奕；

【作者基本信息】 东南大学 ， 电气工程， 2016， 硕士

【摘要】 随着智能电网建设的不断推进,电力系统与通信系统将更紧密地结合,成为电力信息物理融合系统,即电力CPS。电力CPS环境下,先进信息量测和通信技术为电力系统实时广域分析和控制的在线应用提供支撑的同时,也引入了因通信故障导致在线应用失效的风险。这种风险在传统电力系统在线分析和控制应用中缺乏考虑,使电力系统在线分析和控制的实施效果难以保证。因此,电力CPS环境下,有必要在电力系统在线应用中计及潜在的通信风险。本文针对电力系统暂态过程中功角轨迹预测方法进行了研究,取得了如下成果:首先,本文在分析发电机暂态过程功角轨迹相似性的基础上,阐述了通过离线聚类、在线更新建立发电机功角轨迹模式库的方法;提出了一种依靠实时功角量测信息和发电机功角轨迹模式库的功角轨迹预测在线实施方案,该方案将通信故障风险纳入考虑,从而提高了电力系统暂态过程功角轨迹在线预测的可靠性。其次,本文利用影响电力系统暂态过程稳定性的关键信息区分方法,实现了对海量电力系统运行参数信息的快速筛选,并基于极限学习机学习速度快、预测准确度高的特点,将关键信息筛选与ELM数据处理相结合,建立了基于ELM的电力系统暂态过程稳定性评估模型,实现了对电力系统暂态稳定性的快速预测。最后,本文对轨迹匹配预测方法和ELM预测方法的特点进行了比较分析,提出基于稳定快速预测,轨迹匹配校验的电力系统暂态轨迹联合预测方法,实现了在预测计算速度和预测准确度上的平衡。更多还原

【Abstract】 With the development of smart grid construction,integration between power system and communication system is closer and closer.The traditional power system is becoming power cyber physical system(CPS).Power CPS can support power system real-time wide-area analysis and control with advanced measurement and communication technology.Meanwhile,it brings to power system on-line application with risks caused by communication faults.In traditional power system on-line analysis and control applications,this kind of communication risks haven’t been considered which makes implement effect of power system on-line analysis and control unreliable.Thus,it is necessary to consider potential communication risks in power system on-line applications.In this paper,angle trajectory prediction method under power system transient process is studied and achievements are listed as follows:Firstly,method of constructing generator angle trajectory pattern library by off-line clustering and on-line updating is introduced based on analyzing the similarity between generator angle trajectories under different operation condition.An implement prediction method relies on real-tine data measuring and constructed generator angle trajectory pattern library is proposed and this method takes communication risk into consideration which improves reliability of angle trajectory on-line prediction method.Secondly,in order to search key information that have important effect on power system transient angle stability form massive power system operation information,a fast data filtering method to find key information is introduced in this thesis.Considering ELM has high training speed and high predicting accuracy,power system transient angle stability assessment model is set up based on ELM by combining key information filtering and ELM data processing.In this way,power system transient angle stability fast prediction is realized.At last,characteristics of method based on trajectory fitting and method based on ELM are outlined and compared.On this basis,a new power system transient trajectory uniform prediction method carried out by fast stability predicting and trajectory fitting check is proposed and its simulation result shows that this method realizes balance between prediction speed and prediction accuracy.更多还原