Long-term forecasting of hourly electricity load: Identification of consumption profiles and segmentation of customers

Data for aggregated hourly electricity demand shows systematic variations over the day, week, and seasons, and forecasting of aggregated hourly electricity load has been the subject of many studies. With hourly metering of individual customers, data for individual consumption profiles is available. Using this data and analysing the case of Denmark, we show that consumption profiles for categories of customers are equally systematic but very different for distinct categories, that is, distinct categories of customers contribute differently to the aggregated electricity load profile. Therefore, to model and forecast long-term changes in the aggregated electricity load profile, we identify profiles for different categories of customers and link these to projections of the aggregated annual consumption by categories of customers. Long-term projection of the aggregated load is important for future energy system planning, and the hourly load profile is an important input to energy system models that serves this purpose. In particular, these models often assume an unchanged hourly load profile (although the level may change). In contrast, our model suggests that the hourly load profile also changes as the shares of consumption by categories of customers change and new consumption technologies such as electrical vehicles and (for Denmark in particular) individual heat pumps are introduced.

[1]  Eva Gonzalez-Romera,et al.  Monthly electric energy demand forecasting with neural networks and Fourier series , 2008 .

[2]  R. Ramanathan,et al.  Short-run forecasts of electricity loads and peaks , 1997 .

[3]  C. K. Hansen A prognostic model for managing consumer electricity demand and smart grid reliability , 2012, 2012 IEEE Conference on Prognostics and Health Management.

[4]  Gloria Martin-Rodriguez,et al.  Modelling the hourly Spanish electricity demand , 2005 .

[5]  Juan R. Trapero,et al.  Mid-term hourly electricity forecasting based on a multi-rate approach , 2010 .

[6]  P. Meibom,et al.  Optimal investment paths for future renewable based energy systems—Using the optimisation model Balmorel , 2008 .

[7]  Mikko Kolehmainen,et al.  Data-based method for creating electricity use load profiles using large amount of customer-specific hourly measured electricity use data , 2010 .

[8]  Simon Sansregret,et al.  Load duration curve: A tool for technico-economic analysis of energy solutions , 2008 .

[9]  Heike Brand,et al.  WILMAR: A Stochastic Programming Tool to Analyze the Large-Scale Integration of Wind Energy , 2009 .

[10]  Farshid Keynia,et al.  Mid-term load forecasting of power systems by a new prediction method , 2008 .

[11]  Ömer Nezih Gerek,et al.  A novel modeling approach for hourly forecasting of long-term electric energy demand , 2011 .

[12]  Matteo Manera,et al.  Modelling the Load Curve of Aggregate Electricity Consumption Using Principal Components , 2003, Environ. Model. Softw..

[13]  Shu Fan,et al.  Machine learning based switching model for electricity load forecasting , 2008 .

[14]  J. Jardini,et al.  Daily load profiles for residential, commercial and industrial low voltage consumers , 2000 .

[15]  James D. Hamilton Time Series Analysis , 1994 .

[16]  Manuel A. Matos,et al.  A new clustering algorithm for load profiling based on billing data , 2012 .

[17]  Cheng-Ting Lin,et al.  A novel economy reflecting short-term load forecasting approach , 2013 .

[18]  S. Koopman,et al.  An Hourly Periodic State Space Model for Modelling French National Electricity Load , 2007 .

[19]  Donald L. Harnett,et al.  Introduction to statistical methods , 1970 .