Clustering-based forecasting method for individual consumers electricity load using time series representations

• DOI: 10.1515/comp-2018-0006
• Publication Date: 2018-07-25T22:16:02Z
• Authors: Peter Laurinec; Mária Lucká
• Keywords: Representation; Smart meter; Centroid

Abstract This paper presents a new method for forecasting load of individual electricity consumers using smart grid data and clustering. The from all are used clustering to create more suitable training sets methods. Before clustering, time series efficiently preprocessed by normalisation the computation various model-based representation Final centroid-based forecasts scaled saved parameters forecast every consumer. Our is compared with approach that creates consumer separately. Evaluation experiments were conducted on three meter datasets residences Ireland Australia, factories Slovakia. achieved results proved our clustering-based improves accuracy mainly residential consumers.We can also proclaim it be found such setting will perform accurately than fully disaggregated approach. scalable since necessary train model only clusters not separately