作者: José Ramón Cancelo , Antoni Espasa , Rosmarie Grafe
DOI: 10.1016/J.IJFORECAST.2008.07.005
关键词: Names of the days of the week 、 Building process 、 Load forecasting 、 Probabilistic forecasting 、 Electricity 、 Real time forecasting 、 Statistics 、 Operator (computer programming) 、 Consensus forecast 、 Computer science
摘要: Abstract This paper discusses the building process and models used by Red Electrica de Espana (REE), Spanish system operator, in short-term electricity load forecasting. REE's forecasting consists of one daily model 24 hourly with a common structure. There are two types forecasts special interest to REE, several days ahead predictions for data, day hourly forecasts. Accordingly, forecast accuracy is assessed terms their errors. To do this, we analyse historical, real time errors data year 2006, report performance week, type day. Other aspects prediction problem, like influence predicting temperature on ahead, or need an adequate treatment days, also investigated.