Statistical analyses

This group of methods comprise: basic statistics, regression analysis, continuous probability distributions. A statistical analysis approach to load modelling requires large amounts of measurements of energy use. The probability sample must have a high level of statistical significance in order to provide a relevant interpretation of the evolution of load an energy demand.

Linear or multivariate regression analyses or probability distributions are usually used for load and energy predictions. A regression analysis expresses the mathematical correlation between different variables (e.g. climatic or behavioural factors).

Relevant load modelling tools based on statistical analysis methods are: ARX model, Conditional demand analysis (CDA), Energy signature, EModel, The Finnish load model and USELOAD.

Published August 3, 2012

Contact: Gerd Kjølle, SINTEF Energy Research