基于非侵入式负荷检测与分解的电力数据挖掘
Power Data Mining Based on Non-Intrusive Load Detection and Decomposition

电力分项计量对优化电力系统管理、提高电力系统稳定性和可靠性、促进科学合理用电、提高电能利用效率、建设节约型社会和缓解电能压力等方面均有重要的现实意义。电力分项计量的一系列技术，是将电器识别作为物联网的一个重要研究方向，要想电器识别数据挖掘提取特征是一个不可或缺的环节，本文首先对给出的原始数据进行预处理，其次基于预处理后的数据提取出各用电设备的运行特征，然后对本文11种用电设备按照工作状态分类，在暂态环境下，分别提取用电设备的电流、电压、瞬时功率和电压噪声的特征；在稳态环境下，分别提取用电设备的有功、无功功率、电压、V-I轨迹等特征，对于建立用电设备的识别系统具有重要意义。
Power sub metering has important practical significance in optimizing power system management, improving power system stability and reliability, promoting scientific and rational use of electricity, improving power utilization efficiency, building a conservation oriented society and relieving power pressure. A series of technologies of electric power sub metering takes electrical equipment identi-fication as an important research direction of Internet of things. In order to identify electrical equipment, data mining and feature extraction is an indispensable link. Firstly, this paper prepro-cesses the original data, then extracts the operation characteristics of electrical equipment based on the preprocessed data, and then classifies 11 kinds of electrical equipment work according to their working conditions. In the transient environment, the features of current, voltage, instanta-neous power and voltage noise are extracted respectively; in the steady-state environment, the features of active power, reactive power, voltage and V-I trajectory are extracted respectively, which is of great significance for the establishment of the identification system of electrical equipment.