The model for generation of load profiles has been described and validated in [5] and [6]. It constructs household electricity load profiles for a number of end uses by adding appliance loads together, according to use patterns described by time-use data in a representative sample of Swedish households. In this respect it can be classified as a bottom-up model, similar to that in [7].

The model is dependent on (a) time-use data, (b) data on mean appliance powers and (in some cases) maximum appliance runtimes and (c), for modeling of lighting, data on daylight availability. The time-use data set used, collected by Statistics Sweden, has undergone thorough examination and has been used previously in activity visualization studies (see for example [8]). Representative appliance characteristics were collected from publicly available product tests from the Swedish Consumer Agency [9] and daylight availability data were generated for a standard building in Stockholm with Satel-Light [10].

The resulting load data consists of annual series of aggregated hourly electricity demand for 11 different end-uses in the almost 200 households in the time-use data set, corrected for seasonal variations and validated for consistency against preliminary results from ongoing large-scale measurements performed by the Swedish Energy Agency [11].