According to the increased penetration of wind energy into the utility\ngrids, various methods for the short-term (time horizon\n\n6h-48h) prediction of the power output of single installations as\nwell as the ensemble output of groups of turbines within\n\nspecific regions are in application or under development. For methods\nthat are based on the application of the outcome of\n\nnumerical weather prediction schemes, the overall quality of the predictions\nas measured by long term values of the mean\n\nbias and the root mean square errors has been extensively analyzed.\nFor the operational application however, the assignment\n\nof a specific confidence interval for each individual prediction would\nbe desirable. This paper aims in sketching out a way to\n\nidentify this information.\n\nWe present the results of an approach to sort the magnitude of the\nprediction error in terms of its standard deviation\n\naccording to parameters describing the meteorological situation arising\nin course of the forecast procedure (e.g. the predicted\n\npower output itself, the temporal evolution of the forecasted data).\nBased on the respective values and with respect to the non\n\ngaussian nature of the distribution of the forecast errors, a procedure\nto determine confidence intervals for the expected actual\n\npower is proposed.
|Number of pages||4|
|Journal||Proceedings of the European Wind Energy Conference, Copenhagen|
|Issue number||June 2015|
|Publication status||Published - 2001|
- forecasting methods