%0 Journal Article %1 DeFelice2013Electricity %A De Felice, Matteo %A Alessandri, Andrea %A Ruti, Paolo M. %D 2013 %J Electric Power Systems Research %K nwp demand energy %P 71--79 %R 10.1016/j.epsr.2013.06.004 %T Electricity demand forecasting over Italy: Potential benefits using numerical weather prediction models %U http://dx.doi.org/10.1016/j.epsr.2013.06.004 %V 104 %X Although there are many works about forecasting using weather variables, the effects of information accuracy have been rarely investigated. This work will use temperature data obtained from numerical weather prediction models, considering both data reanalysis and operational forecasts. Use of weather data provided by models in all the regions leads to a clear improvement of forecasting accuracy. The observed gap between operational forecasts and reanalysis may be reduced with more accurate initial condition. Electricity demand forecasting is a critical task for energy management of power grids. Due to the wide use of refrigeration and residential air-conditioning devices, electricity demand in Italy is influenced by weather conditions, especially during summer. This paper performs daily load forecasting for Italy through statistical modeling with the aim of studying the influence of temperature. The actual capability of available weather forecasts to contribute in predicting electricity loads is evaluated by using weather data from numerical weather prediction (NWP) models. Time-series models have been used and compared with a naive predictor on working-days daily load during June and July in years 2003–2009 considering lead-times between one and five days. Results are analyzed both at the national level and at regional scale, using unprecedented historical load data provided by the Italian transmission grid manager. It is shown that the use of weather data provided by NWP models leads to performance improvements, especially for the hottest areas where the use of electricity is more heavily influenced by temperature. Furthermore, by observing the gap between load forecast models using reanalysis and operational forecast weather data we can obtain some clues about the limitations of the weather forecast models we used on specific geographic areas in Italy.

@article{DeFelice2013Electricity, abstract = { Although there are many works about forecasting using weather variables, the effects of information accuracy have been rarely investigated. This work will use temperature data obtained from numerical weather prediction models, considering both data reanalysis and operational forecasts. Use of weather data provided by models in all the regions leads to a clear improvement of forecasting accuracy. The observed gap between operational forecasts and reanalysis may be reduced with more accurate initial condition. Electricity demand forecasting is a critical task for energy management of power grids. Due to the wide use of refrigeration and residential air-conditioning devices, electricity demand in Italy is influenced by weather conditions, especially during summer. This paper performs daily load forecasting for Italy through statistical modeling with the aim of studying the influence of temperature. The actual capability of available weather forecasts to contribute in predicting electricity loads is evaluated by using weather data from numerical weather prediction (NWP) models. Time-series models have been used and compared with a naive predictor on working-days daily load during June and July in years 2003–2009 considering lead-times between one and five days. Results are analyzed both at the national level and at regional scale, using unprecedented historical load data provided by the Italian transmission grid manager. It is shown that the use of weather data provided by NWP models leads to performance improvements, especially for the hottest areas where the use of electricity is more heavily influenced by temperature. Furthermore, by observing the gap between load forecast models using reanalysis and operational forecast weather data we can obtain some clues about the limitations of the weather forecast models we used on specific geographic areas in Italy.}, added-at = {2018-06-18T21:23:34.000+0200}, author = {De Felice, Matteo and Alessandri, Andrea and Ruti, Paolo M.}, biburl = {https://www.bibsonomy.org/bibtex/2a5d3e0cad806fd8c33f88fa1e26f3001/pbett}, citeulike-article-id = {14013827}, citeulike-attachment-1 = {DeFelice_etal_2013.pdf; /pdf/user/pbett/article/14013827/1064076/DeFelice_etal_2013.pdf; c12211011ec8cd8d8e2eff0fb77e1bf912818dc2}, citeulike-linkout-0 = {http://dx.doi.org/10.1016/j.epsr.2013.06.004}, doi = {10.1016/j.epsr.2013.06.004}, file = {DeFelice_etal_2013.pdf}, interhash = {9055720f325a34f0d9f0ea569aa9e07b}, intrahash = {a5d3e0cad806fd8c33f88fa1e26f3001}, issn = {03787796}, journal = {Electric Power Systems Research}, keywords = {nwp demand energy}, month = nov, pages = {71--79}, posted-at = {2016-04-16 12:17:17}, priority = {2}, timestamp = {2018-06-22T18:36:14.000+0200}, title = {Electricity demand forecasting over Italy: Potential benefits using numerical weather prediction models}, url = {http://dx.doi.org/10.1016/j.epsr.2013.06.004}, volume = 104, year = 2013 }