•The Aiolos 2.0 and Cilia 2.0 models work with day type classification, i.e. it searches for historical material to establish hourly day curves for different day types. The program constructs hourly means for the forecasted series as well as for all input data parameters that the regression uses as predictors. The period used to construct these hourly day type patterns is called the ‘reference period’, since the result will be a load and a weather that are used as a reference in the regression model.
•The models distinguish between day light saving time and normal/winter time so that the model calculates separate day type patterns for days during dst-time and normal time. This means the model can allow for changed social patterns in the changeover between summer and winter time that depend on the astronomical time, e.g. reduced need for lighting in the morning when winter time starts.