Data-Based Modeling of High-Resolution Household Load Profiles
Abstract
An algorithm to generate individual household electrical load data which relies solely on a previously recorded dataset is presented. A dataset consisting of 348.766 days of load data sampled in 5 minute intervals was used to calibrate the model in terms of realistic seasonal and daily variability traces. A SARIMA model extended with random effects generates new load traces which are not part of the original dataset. Altogether, the new load traces resemble the original data characteristics. The model can be used to generate large numbers of long load profiles required for detailed simulation models of residential areas renewable energy systems.