基于动态自适应粒子群算法的非侵入式家居负荷分解方法

A Non-Intrusive Household Load Monitoring Method Based on Dynamic Adaptive Particle Swarm Optimization Algorithm

非侵入式负荷监测可以在保证用户隐私的前提下深入分析用户独立负荷的用电信息,是智能用电技术体系的关键内容。为提高负荷辨识的准确性,提出一种基于动态自适应粒子群算法（dynamic adaptive particle swarm optimization,DAPSO）的非侵入式负荷分解方法。在传统功率特征的基础上,将总谐波失真系数（total harmonic distortion,kTHD）作为负荷新特征引入目标函数,采用DAPSO算法对实测用电数据进行负荷分解。仿真结果表明,在不同噪声背景下,DAPSO算法的负荷辨识率和收敛速度均得到一定提高,从而验证了DAPSO算法对家居负荷分解具有更优的可靠性和鲁棒性。

Non-intrusive load monitoring（NILM） is a key technology in intelligent power utilization system, which can acquire internal load compositions in the premise of protecting users＇ privacy. In order to improve accuracy of load identification, a non-intrusive household load decomposition method based on dynamic adaptive particle swarm optimization algorithm（DAPSO） is presented. Based on traditional power signature, total harmonic distortion（kTHD） is introduced to objective function as a new load signature. According to real sampling load data, DAPSO algorithm is used to realize load monitoring. Simulation results show that in different noise environments, recognition accuracy and convergence rate of DAPSO algorithm are improved, verifying that DAPSO algorithm has better accuracy and stability for household load monitoring.