Stochastic Modelling of Electricity Demand on Multiple Time Scales

This paper develops a stochastic model for half-hourly electricity demand in Great Britain in the long term. The model captures annual, weekly and daily seasonal patterns as well as temperature dependencies through semi-parametric additive models. The model allows for the incorporation of expert views and other information that may not be contained in historical data. The model is easy to implement and computationally fast, which makes it useful for long-term electricity price simulations and risk management. The model was developed in collaboration with the UK government entity, Low Carbon Contracts Company