Dynamic clustering of residential electricity consumption time series data based on Hausdorff distance

Abstract

As the analysis of electrical loads is reaching data measured from low voltage power distribution networks, there is a need for the main agents involved in the operation and management of the power grids to segment the end users as a function of their shapes of daily energy consumption or load profiles, and to obtain patterns that allow to classify the users in groups based on how they consume the energy. However, this analysis is usually limited to the analysis of single days. Since the smart metering data are time series formed by sequential measurements of energy through each hour or quarter of hour of the day, and also through each day, thanks to the implementation of Advanced Metering Infrastructure (AMI) and the Smart Grid technologies, it becomes clear that the analysis of the data needs to be extended to consider the dynamic evolution of the consumption patterns through days, weeks, months, seasons, and even years. This is the objective of the present work. A new framework is presented that addresses the dynamic clustering, visualization and identification of temporal patterns in load profiles time series, fulfilling the detected gap in this area. The present development is a generic framework that allows the clustering and visualization of load profiles time series applying different classical clustering algorithms. A novel dynamic clustering algorithm is also presented, based on an initial segmentation of the energy consumption time series data in smaller surfaces, and the computation of a similarity measure among them applying the Hausdorff distance. Following, these developments are presented and tested on two dataset of energy consumption load profiles from a sample of residential users in Spain and London.

A comparative study of clustering techniques for electrical load pattern segmentation

“…A few studies in the literature have addressed the "dynamic"/"online" clustering of load data and the problem of big data. Dynamic clustering of time series data is considered in [104] and [105] to deal with the dynamic evolution of the consumption data through time. The presented framework in [104] for dynamic clustering of load curves compares the performance of K-means and FCM algorithms with different similarity measures including the Euclidean distance, the Pearson correlation coefficient, and another measure called Hausdorff distance.…”

Smart Grids Data Analysis: A Systematic Mapping Study

“…SG failures deals with fault status detection, fault type classification, power distribution reliability. [3], [26], [40], [75], [84], [94], [112], [129], [160], [167], [173], [204], [212], [214], [232], [233], [239], [245]), power consumption pattern recognition ( [7], [44], [49], [64], [67], [131], [150], [153], [183], [265], [266]), power load forecasting ( [56], [250]), events/tasks extraction ( [50])…”

A pattern recognition methodology for analyzing residential customers load data and targeting demand response applications

“…Other algorithms such as follow-the-leader [3], density-based spatial clustering of applications with noise (DBSCAN) [12], [13], and K-shapes [14] are examined in some publications. Since the consumption values are continuously recorded over time, dynamic [15], [16] and online [17], [18] clusterings of time series data are also considered in a few studies.…”