基于广域信息的变电站综合负荷在线建模方法及应用

发布时间：2018-04-15 01:32

本文选题：负荷建模 + 统计综合法　；参考：《湖南大学》2016年硕士论文

【摘要】：近年来,由于电力系统工程技术的快速发展和电网规模的不断扩大,建设坚强统一的智能电网对电网智能调度提出了新的更高要求,以在线、实时仿真计算结果为主要决策依据的电网运行调度与安全稳定控制,是智能电网建设条件下电网调度的必然发展趋势。而实时、精确的电网仿真计算必须依赖于实时、准确的负荷模型,目前电网仿真计算普遍采用一套既有的负荷模型和模型参数,这不能在根本上满足精确的实时仿真计算对负荷模型的要求。本文立足于当前电力系统负荷建模工作的实际状况,针对负荷特性的随机时变性问题,在对电力系统综合负荷建模方法进行深入研究的基础上,构建了一套完整的变电站综合负荷在线建模的理论方法和实现技术。首先综合运用统计综合法和总体测辨法进行负荷建模,然后在变电站典型用电行业的构成解析的基础上,对得到的负荷模型参数进行在线修正或预测,以便得到实时的或未来的负荷模型参数。统计综合法基于统计学原理,从底层用电设备建模开始,逐步加权,先完成行业建模,再完成变电站建模。总体测辨法基于系统辨识理论,根据现场实测数据利用辨识算法来确定模型的结构和模型参数,使得模型响应能最好地拟合观测到的实测数据。变电站典型用电行业的构成解析方法是通过对海量数据平台存储的用户负控系统与变电站SCADA系统采集得到的用户和变电站日负荷功率曲线进行数据挖掘,经用户分类与综合后得到变电站各类典型用电行业的等效日负荷功率曲线,然后在典型用电行业的等效日负荷功率曲线和变电站日负荷功率曲线之间建立映射关系,通过最小二乘法辨识两者间的映射关系,最后解析得到变电站典型用电行业的构成比例。在此方法的基础上,分别利用实时的和负荷预测得到的日负荷功率曲线,解析得到变电站实时的和待预测日的典型用电行业构成比例,结合典型用电行业的负荷模型参数,加权即可得到变电站实时的和待预测日的综合负荷模型参数,最终实现模型参数的在线修正和预测。根据不同的建模方法需要,本文用于综合负荷建模的数据有：用户数据(含调查统计数据、负控数据)、SCADA/EMS数据、故障录波数据、WAMS/PMU数据等四类数据。从广域信息的来源及特点、下载及预处理方法、数据库设计以及入库流程等几个方面详细阐述了广域信息的集成方法,形成了广域信息数据库以便负荷建模系统调用。以综合负荷建模方法和广域信息数据为基础,在基于Windows的操作系统下,采用结构化设计思想,利用面向对象的可视化软件Visual Studio 2005作为编程环境,以C++作为编程语言,以关系数据库SQL Server 2000作为本地数据管理服务器,通过ADO(活动数据对象)数据接口技术实现广域信息数据库的操作,设计开发了数据源下载预处理模块、广域信息数据库管理模块、行业及变电站分类与综合模块、负荷模型辨识模块、负荷模型参数在线修正模块与负荷模型参数库管理模块,并通过WinExec函数和动态链接库(DLL)等网络软件技术实现各个功能模块的调用,共同构建了完整的基于广域信息的变电站综合负荷在线建模系统。其主要功能是对广域信息数据进行下载、预处理、查询及修改,形成历史/实时广域信息数据库；在此基础上利用历史/实时广域信息数据实现负荷模型的辨识和在线实时修正,最终形成完整的历史/实时负荷模型参数库,以便电网仿真计算应用。本文开发的负荷在线建模系统已在某省级电力调度控制中心试运行,运行稳定,与以往的负荷建模系统相比,操作更加简单灵活,功能更加完善。
[Abstract]:In recent years, due to the rapid development of engineering technology of power system and the enlargement of power grid scale, the construction of smart grid strong unified smart grid to put forward higher requirements, a new online, grid operation security and stability control real time simulation results as the main basis for decision-making, is the inevitable trend of the development of smart grid dispatching construction conditions. In real time, accurate calculation of power system simulation must rely on real-time, accurate load model, the power grid simulation generally adopts a set of existing load models and model parameters, which can not meet the real-time simulation precision in the fundamental calculation of load model. According to the actual situation in the current power system load modeling work, according to the random time-varying load characteristics, based on the in-depth study of power system composite load modeling method On the construction and implementation of a complete set of theory and method of substation load modeling. Firstly using the statistical synthesis method and the measurement method of load modeling, and then in a typical substation based on the power industry composition analysis, revision or online prediction on load model parameters, in order to get real time or the future load model parameters. The statistical synthesis method based on the principle of statistics, starting from the bottom of the electric equipment to complete the industry gradually weighted modeling, modeling, and complete substation modeling. The measurement method based on system identification theory, according to the measured data using the identification algorithm to determine the model structure and model parameters, the the model response can best fit the observed data. A typical substation analysis method of power industry is based on the massive data storage platform for users of negative control System and substation SCADA system collected by users and power substation daily load curve data mining, all kinds of typical daily load substation with the equivalent power curve power industry obtained by user classification and synthesis, and then establish the mapping relationship between power industry, daily load curve and the equivalent power substation daily load curve in typical power, by mapping the relationship between the least squares identification, finally resolved the typical substation electricity industry proportion. On the basis of the method, the daily load curve of power respectively by real-time and load forecasting, analytic and real-time forecasting of the electricity industry typical proportion, with the typical parameters of electric load model the industry, can be obtained and weighted real-time forecast load model parameters, finally realize the model parameters The online correction and prediction. According to the need of different modeling methods, this paper used comprehensive load modeling data: user data (including the survey statistics, negative control data), SCADA/EMS data, fault recorder data, WAMS/PMU data and other four kinds of data sources and characteristics. From the wide area information, download and pretreatment methods and the design of database and storage process described in detail several aspects of integrated method of wide area information, forming a wide area information database for load modeling system calls. With comprehensive load modeling method and wide area information based on the data in the operating system based on Windows, using a structured design idea, using the object-oriented visualization software Visual Studio 2005 the programming environment, using C++ as a programming language, using the relational database SQL Server 2000 as the local data management server through the ADO (active data object number) According to the implementation of wide area information database interface technology, the design and development of data source download wide information preprocessing module, database management module, and substation industry classification and synthesis module, module load model identification, load model parameters online correction module and load model parameter library management module, and through WinExec function and dynamic link library (DLL etc.) to realize the function of each module is called network software technology, to build a complete substation online load modeling system based on wide area information. Its main function is to the wide area information data download, preprocessing, query and modify, form a real-time wide area information database on the basis of history; history / real-time wide area information the data real-time identification and on-line correction of the load model, and ultimately form a complete history / real-time load model parameter library, in imitation of power grid The online load modeling system developed in this paper has been trial run in a provincial power dispatching control center and runs stably. Compared with the previous load modeling system, the load online operation system is more simple and flexible, and the function is more perfect.

【学位授予单位】：湖南大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TM714;TM743;TM63

【相似文献】