Copy the page URI to the clipboard
Vukadinovic Greetham, Danica and Giasemidis, Georgios
(2019).
DOI: https://doi.org/10.1145/3307772.3330164
Abstract
Recurrence Quantification Analysis (RQA) is becoming a popular technique to analyse time-series obtained from complex dynamical systems. In this work, a recently presented RQA-based method to analyse and manage energy demand at low sampling rate (5 min and 30 min) is tested using data-sets from three small enterprises. From recurrence plots, different RQA variables are obtained and analysed, following parameter optimisation that depends on a system observed. Based on RQA variables, energy maps of ‘normal’ behaviour are created. Here, preliminary robustness tests concern- ing the training phase length, missing data and noise are presented. Our test results show that this approach has great potential in energy management of small and medium enterprises.