基于非侵入式技术的负荷检测及辨识方法研究

CORC > 沈阳自动化研究所 > 中国科学院沈阳自动化研究所 > 工业控制网络与系统研究室

题名	基于非侵入式技术的负荷检测及辨识方法研究
作者	李择豪
学位类别	硕士
答辩日期	2016-05-31
授予单位	中国科学院沈阳自动化研究所
导师	臧传治
关键词	非侵入式负荷检测 BP神经网络粒子群算法负荷分解遗传算法
其他题名	Research on methods about load detection and identification based intrusive technology
学位专业	控制工程
中文摘要	本文提出了一种基于非侵入式技术的负荷检测方法，并开展了相关辨识方法的研究。(1) 通过对课题背景和国内外研究现状的分析，明白非侵入式负荷监测的原理，确定非侵入式监测系统所需的模块组成。(2) 采集居民家庭中典型电器负荷的相关波形，对负荷的常见负荷特性进行了研究分析，经过权衡选取了其中部分负荷特性作为负荷辨识的依据，并介绍了各负荷特性特征值的计算方法，为后文非侵入式负荷检测与分解打下基础。(3) 提出了一种负荷监测算法，该方法针对传统单一BP神经网络的缺陷，引入改进粒子群算法进行优化，采用设备电流有效值，电流峰值，波峰系数，有功功率、无功功率组成特征向量作为输入，选取电加热器、微波炉、台灯三种设备作为输出，采集设备启动后的有关波形数据，由MPSO-BP网络进行识别。通过对实例的仿真计算分析，验证了这种方法的有效性和准确性。(4) 提出了一种功率分解模型，该方法用不同类型电器的稳态电流的线性组合去近似拟合负荷总电流，使总电流与组合电流在一个周期内数值的残差平方和最小，目标是求解该线性组合中的系数，之后计算得到不同类型电器的功率消耗情况。通过对实例的仿真计算分析，验证了这种方法的有效性和准确性。
英文摘要	In this paper, a new method of power load monitoring, which named non-intrusive power load monitoring technology, is developed to be researched. Based the analysis of the research background and research status at home and abroad, we know the principle of non-intrusive load monitoring clearly, so as to determining the necessary functional modules of non-intrusive monitoring system. Acquired typical electrical waveform of households load, the common load characteristics of the load were analyzed, and some of the load characteristics were chosen as the basis of the load identification after consideration, the computing methods of these load signatures eigenvalues were proposed, to build the foundation for the non-intrusive load detection and decomposition. Presents a load monitoring algorithm, and introduces the improved particle swarm optimization algorithm to improve the defects of the traditional single BP neural network. Then chooses the device current effective value, peak current, crest factor, active power, reactive power to constitute feature vectors as the input, selects electric heater, microwave oven, lamp as the output. Firstly acquires the waveform data of devices, after that uses MPSO-BP network to achieve load identification. The validity and accuracy of the method are verified by the simulation of an example. A type of non-intrusive power load power decomposition optimization model is proposed. The model uses different types of the electric equipment of the steady-state current by the linear combination to approximately fitting the total load current, within a cycle numerical residual sum of squares and minimum, the goal is to solve the linear combination coefficients to determine power load in different types of electrical equipment of power consumption ratio of the total current and the current combination. The validity and accuracy of the method are verified by the simulation of an example.
语种	中文
产权排序	1
页码	75页
内容类型	学位论文
源URL	[http://ir.sia.cn/handle/173321/19614]
专题	沈阳自动化研究所_工业控制网络与系统研究室
推荐引用方式 GB/T 7714	李择豪. 基于非侵入式技术的负荷检测及辨识方法研究[D]. 中国科学院沈阳自动化研究所. 2016.