基于聚类和关联分析的居民用户非侵入式负荷分解

现有的非侵入式负荷监测方法主要采用监督学习模型,该类模型需要具有针对性的大量训练数据,而且无法有效识别在训练数据中未出现的负荷。在分析多种家用电器负荷特征的基础上,选取负荷投切过程中暂态功率波形和功率变量作为负荷特征,并提出一种基于聚类和关联分析的无监督学习居民用户非侵入式负荷分解方法。首先根据功率变化情况提取电流和电压数据,并计算得到暂态功率波形;然后通过动态时间规整算法计算当前暂态功率波形与历史暂态功率波形的匹配度,并利用动态聚类算法和其他暂态负荷特征判别该功率波形对应的负荷操作;最后以周为单位对负荷操作进行关联分析,确定每种电器对应的多个暂态特征。仿真结果表明,所提方法易于实现,在准确率和可靠性方面有明显提高。

Non-intrusive load decomposition of residential users based on cluster and association analysis

The current non-intrusive load monitoring methods mainly use supervised learning models, which need a large amount of targeted training data and cannot effectively identify the load that does not appear in the training data. On the basis of analyzing the load characteristics of various household appliances, the transient power waveform and power variables in the load switching process are selected as the load characteristics, and an unsupervised learning non-intrusive load decomposition method of residential users based on cluster and association analysis is proposed. Firstly, the current and voltage data are extracted according to the power variation, and the transient power waveform is calculated. Secondly, the dynamic time warping algorithm is used to calculate the matching degree between the current transient power waveform and the historical transient power waveform, and the dynamic cluster algorithm and other transient load characteristics are used to determine the load operation corresponding to the power waveform. Finally, the association analysis of load operation is carried out in weeks to determine the multiple transient characteristics corresponding to each appliance. The simulative results show that the proposed method is easy to implement, and it has obvious improvement in both accuracy and reliability.