From probabilistic forecasts to statistical scenarios of short‐term wind power production

Abstract

Short-term (up to 2-3 days ahead) probabilistic forecasts of wind power provide forecast users with a highly valuable information on the uncertainty of expected wind generation. Whatever the type of these probabilistic forecasts, they are produced on a per horizon basis, and hence do not inform on the development of the forecast uncertainty through forecast series. However, this additional information may be paramount for a large class of time-dependent and multi-stage decision-making problems e.g. optimal operation of combined wind-storage systems or multiple-market trading with different gate closures. This issue is addressed here by describing a method that permits the generation of statistical scenarios of short-term wind generation that accounts for both the interdependence structure of prediction errors and the predictive distributions of wind power production. The method is based on the conversion of series of prediction errors to a multivariate Gaussian random variable, the interdependence structure of which can then be summarized by a unique covariance matrix. Such matrix is recursively estimated in order to accommodate long-term variations in the prediction error characteristics. The quality and interest of the methodology are demonstrated with an application to the test case of a multi-MW wind farm over a period of more than two years.

Use of wind power forecasting in operational decisions.

“…Since the probabilistic forecasts (i.e., set of quantiles) produced by the quantile regression method do not capture the temporal correlation between forecast errors (i.e., uncertainty) of different hours, a scenario generation method described in [78] is used to generate a number of wind power scenarios that provide information on the development of the prediction errors through the set of look-ahead times. The method takes as inputs the forecasted quantiles for each look-ahead time-step and also the observed wind power generation.…”

“…A specified number of scenarios are obtained through sampling from an inverse cumulative distribution function. More details about the method can be found in [78]. The scenarios produced by the method have the following characteristics: (i) they respect the marginal wind power forecasted distribution for the coming horizon (i.e., probabilistic forecasts); and (ii) the temporal correlation of forecast errors are embedded in the scenarios values.…”

“…To produce wind power scenarios, wind power data (forecasted and realized) were used to train the uncertainty estimation model, as well as to generate scenarios for the forecasted wind power, according to the methodology introduced by Pinson et al [78], which is equivalent to producing scenarios under a Monte Carlo process.…”

“…This large set of scenarios should be a suitable approximation of the probability density function of wind power. One of the possible approaches to generate these scenarios is presented in [78].…”

“…In this work, the scenarios sampled by using the process described in [78] are analyzed and the structure of such scenarios is shown to be unimodal, by means of using the information theoretic learning (ITL) Mean Shift algorithm [86], [87].…”