一种用电设备分析识别原理与应用

Principle and Application of Analysis and Identification of Electrical Equipment

非侵入式负荷监测(NILM)与识别是智能电网建设的重要方面。针对非侵入式用电负荷监测技术提出了一种基于V-I轨迹与奇次谐波分量特征的混合特征矩阵和卷积神经网络相结合的方法。具体内容为:分析用电设备的基本特征,包括输入电流稳态特征和暂态特征。建立了六种单相交流电源供电的用电设备MATLAB/Simulink仿真电路,分别为电阻加热器、电吹风、LED灯串、电源适配器、变频空调和变频冰箱,作为待分析识别的用电设备,并提取了这六种负荷模型的网侧电流波形和频谱、V-I轨迹以及奇次谐波电流分量数据作为基本特征量;提出了一种采用V-I轨迹与奇次谐波特征的混合特征矩阵以及卷积神经网络(CNN)进行用电设备分析识别,经过PLAID数据集的训练和测试,能够有效识别稳态运行下的用电设备,给出了详细的解算过程,并提供了应用算例。此外,对比分析了已有三种检测方法,验证本文提出的基于混合特征矩阵以及卷积神经网络用电设备分析识别方法的优越性。

Non-intrusive load monitoring(NILM)and identification are important aspects of smart grid construction.A method combining a mixed feature matrix and convolutional neural network based on V-I trajectory and odd harmonic component features is proposed for non-invasive electricity load monitoring technology.The specific content is to analyze the basic characteristics of electrical equipment,including steady-state and transient characteristics of input current.Six simulation circuits were established for the MATLAB/Simulink Electrical or Electronics module library of single-phase AC power supply electrical equipment,including resistance heaters,hair dryers,LED light strings,power adapters,variable frequency air conditioners,and variable frequency refrigerators,as the electrical equipment to be analyzed and identified.The grid side current waveform and spectrum,V-l trajectory,and odd harmonic current component data of these six load models were extracted as basic feature quantities;A mixed feature matrix using V-l trajectory and odd harmonic features,as well as a convolutional neural network(CNN),were proposed for the analysis and recognition of electrical equipment.After training and testing on the PLAID dataset,it was found that the system can effectively identify electrical equipment under steady-state operation.A detailed solution process was provided,and application examples were provided.In addition,three different detection methods were compared and analyzed to verify the superiority of the proposed method for analyzing and identifying electrical equipment based on mixed feature matrices and convolutional neural networks.