家用电器负荷在线参数辨识方法的研究

发布时间：2019-04-02 08:21

【摘要】：负荷监测是获取用户详细的用电情况、用电行为和开展开节能工作的基础,供电企业通过分析用户行为制定合理的需求响应策略,正确引导用户用电,保证电力供应的稳定性和经济性,用户通过了解供电情况、电网政策以及自身精确的用电信息,合理安排自身用电行为降低能源的消耗,减少开支。非侵入式负荷监测方法只需要在被监测系统电力入口处安装信息采集装置,并运用合适算法对用电数据进行处理和分析,就能获取系统内各个负荷的用电信息,有效的解决了监测设备安装难、维护难及管理难等问题,是未来负荷监测发展的方向。本文主要研究了一种非侵入环境下的家用电器负荷在线参数辨识方法,从负荷特性入手,就负荷投切辨识、特征提取和负荷分类辨识等方面进行了研究,并完成了家用电器负荷在线参数辨识演示系统搭建,具体工作如下：(1)阐述了家用电器负荷在线参数辨识方法的研究背景与意义,研究国内外负荷监测系统和家电负荷特征提取与分类的相关现状,通过智能电网方法论的方案从用户需求出发,研究非侵入式负荷监测系统的优势、物理架构和工作原理；(2)建立负荷采集硬件系统,完成常见家用电器单一运行数据、混合运行数据的采集,分析负荷用电特征,以负荷电流、电压、有功、无功、高次谐波含量、相角等多维特征为先验训练样本,结合负荷硬件结构,挖掘不同负荷的独有特征；(3)研究基于负荷投切状态的投切识别技术,利用主成分分析法对8种典型电器的负荷特征样本进行降维处理,得到最优辨识特征,根据负荷评估值建立二层分类器,通过Fisher有监督线性判别将家用电器的主特征值矩阵在一维空间解聚,建立一层分类器,实验环境下使用Matlab下完成分类器设计。(4)使用虚拟仪器完成家用电器在线参数辨识系统搭建,实现登陆、电能监测、开断状态监测、单负荷运行状态显示、历史数据查询以及通信功能,更加方便、直观的演示本文所提出的负荷在线参数辨识方法。
[Abstract]:Load monitoring is the basis for obtaining users' detailed electricity consumption situation, power consumption behavior and energy saving work. Power supply enterprises make reasonable demand response strategies by analyzing user behavior, and guide users to use electricity correctly. In order to ensure the stability and economy of power supply, users can reasonably arrange their own power consumption and reduce energy consumption and expenditure by understanding the power supply situation, power grid policies and their own accurate power consumption information. The non-intrusive load monitoring method only needs to install the information collection device at the power entrance of the monitored system, and use the appropriate algorithm to process and analyze the power consumption data, so that the power consumption information of each load in the system can be obtained. It effectively solves the problems of difficult installation, maintenance and management of monitoring equipment, which is the development direction of load monitoring in the future. In this paper, a method of on-line parameter identification for household appliances under non-intrusive environment is studied. Starting with the load characteristics, the load switching identification, feature extraction and load classification identification are studied. The main work is as follows: (1) the research background and significance of on-line parameter identification method for household appliance load are described. This paper studies the current situation of load monitoring system and home appliance load feature extraction and classification at home and abroad, and studies the advantages, physical structure and working principle of non-intrusive load monitoring system based on smart grid methodology. (2) establish the hardware system of load collection, complete the collection of common household appliances single operation data and mixed operation data, analyze the characteristics of load electricity, take load current, voltage, active power, reactive power, high-order harmonic content, and make use of load current, voltage, active power, reactive power and high-order harmonic content. The multi-dimensional feature such as phase angle is a priori training sample, combined with load hardware structure, the unique characteristics of different loads are excavated. (3) the switching identification technology based on load switching state is studied. The load feature samples of eight typical electrical appliances are reduced by principal component analysis, the optimal identification characteristics are obtained, and a two-layer classifier is established according to the load evaluation value. The principal eigenvalue matrix of household appliances is deaggregated in one-dimensional space by Fisher supervised linear discriminant, and a classifier is established. In the experimental environment, the classifier is designed by using Matlab. (4) the on-line parameter identification system of household appliances is built with virtual instrument, which realizes the landing, electric energy monitoring, interrupt state monitoring, single load operation status display, and so on. Historical data query and communication functions make it more convenient and intuitive to demonstrate the load online parameter identification method proposed in this paper.
【学位授予单位】：华北电力大学(北京)
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TM925.06

【相似文献】