Evaluation of time series techniques to characterise domestic electricity demand

This paper discusses time series approaches, often used by Transmission System Operators (TSOs) to forecast system demand, and applies them at an individual dwelling level. In particular, two techniques, Fourier transforms and Gaussian processes were evaluated and used to characterise individual household electricity demand. The performance of the characterisation approaches were evaluated based on Pearson correlation coefficient, descriptive statistics and paired sample t-tests for electrical parameters: Total Electricity Consumption, Maximum Demand, Load Factor and Time of Use of maximum electricity demand. Finally, a number of time series tests were carried out to ensure certain properties remained between the original and characterised series.Both Fourier transforms and Gaussian processes were shown to be suitable techniques for characterising domestic electricity demand. Depending on customer demand load profiles, each approach has its own strengths and weaknesses. Fourier transforms are better at characterising the profiles of customers who consume electricity more evenly across the day (>1 h). In contrast, Gaussian processes are better at characterising customers whose demand is high for only short periods of time (≤1 h) throughout the day.

► We discuss time series approaches to characterise domestic electricity demand.
► Fourier transforms are better at characterising longer intervals of smaller electricity demand.
► Gaussian processes are better at characterising small intervals of larger electricity demand.