【作者】 王振宇；

【导师】 郑贵林；

【作者基本信息】 武汉大学 ， 机械设计及理论， 2011， 博士

【摘要】 目前,随着社会的发展能源需求总量也在不断提高,能源的使用效率问题便日益突出,特别是各种资源逐渐匮乏的今天,于是作为一种针对能源使用效率的思路,电力需求侧管理(DSM)越来越被人们重视。这是一种在政府法规和政策支持下,采用有效的激励和引导措施以及适宜的运作方式,让各方共同协力提高终端用电效率和改变用电方式,能够在满足相近用电功能的同时减少总电量消耗或降低电力需求增长,使各方能够受益的最小成本电力服务的管理活动,从源头上达到节约资源和保护环境的目的。但由于DSM面向的用电对象情况复杂多变、难以监测、牵涉范围广、负荷管理措施标准难以制定实施等问题,使得目前DSM发展取得的成效有限,本文针对这一难点以低压侧负荷管理为背景对其中一些关键问题进行研究。本文的主要研究内容可以概括如下：(1)通过对低压侧用电环境进行调查分析,以其中最为复杂的居民区用电情况为例,对用电负荷设备进行详细的研究与分类,并对基于负荷特性辨识的用电管理系统的组成结构及其各部分主要功能进行了设定。(2)通过对原始采样数据进行模型重建,以新数据描述模型对数据进行简化,同时让数据保真度满足系统需求,设计相应的算法完成数据描述模型的转换,最后对算法进行优化以降低其复杂度,使其能够内嵌至电能表中的主控芯片中,以达到节约管理系统硬件成本的目标。(3)对负荷特性辨识的关键问题进行了研究,负荷辨识过程可总结为独立负荷的特征提取、分类以及辨识三个部分,这些问题将通过聚类算法与线性判别函数分别完成,算法的理论基础及具体设计在文中将逐一讨论。(4)对DSM中负荷管理模式进行分析与研究,以是否直接对负荷进行操作为标准把其分为软硬两种模式,并以新型负荷监测系统为基础,对传统DSM负荷管理手段进行改进与创新,提出了以用电设备使用必要性为新负荷分类标准,并以此为基础建立软管理模式中负荷必要性电价模型,以及硬管理模式中分级中断负荷模型。基于以上研究内容,改造设计了能完成文中要求功能的新型电表,并以MATLAB为平台建立模拟的数据处理中心与信息交互系统,完成相应的现地实验,并给出相应的数据结果与分析。更多还原

【Abstract】 Now, with the increasing development of energy demand, the energy efficiency is becoming prominent, especially the gradual lack of resources today, so as an scheme for energy efficiency, demand side management (DSM) has been growing attention. DSM is an electric power service management which makes all quarters can profit at smallest cost, it runs under the supports of laws and government policy, let all quarters unite efforts together enhances the electricity efficiency and change the way of energy consumption. The goal of DSM is reduce power consumption or electricity demand growth at the precondition of keeping the same function of power system, which save resources and protect the environment from the source.The main contents and contributions are as follow.(1) Make the investigation to the electrical environments of low-voltage side, and set residential area as the research target which is the most complicated in the low-voltage side. A detailed investigation of residential loads is implemented for proper load classification of identification, the structure and functions of main parts in power management system are the discussed and settled.(2) Rebuild the raw sampling data flow by a new desription model which is simpler at the storage space without cost the fidelity of the raw data. The algorithm is designed simple enough for embedded into the energy meter, which ensure the cost of manage system can be accepted.(3) Research on the main problems of load identification, which can be summaried into three parts, feature extraction, objects(loads) classification and identification function, these can be soluted by clustering algorithm and linear discriminant function, the algorithm theory and specific design will be present in this paper in details.(4) Analysis and research on the load management modes of DSM, classifies these modes into hard and soft way by whether do operation to loads directly, then builds manage modes based on the new load identification algorithm. The new manage mode sets necessity of the loads as the classification criteria and build rules for each load class, such as necessity-traiff modes(soft management) and interruptible load class management(hard management).Based on the contents above, new multi-function energy meter which meet the requirments of load identification system, and an platform based on MATLAB for data and information exchanging and processing, experiments and analysis results are presented in the papar.更多还原