Publications

SelectionFile type iconFile nameDescriptionSizeRevisionTimeUser
ċ

View
We study the variability of the wind energy resource in France on monthly to seasonal timescales. On such long‑term timescales, the variability of the surface wind speed is strongly influenced by the large‑scale situation of the atmosphere. We investigate the relationship between the large scale circulation and the surface wind speed, summarizing the former by a principal component analysis, so that the large‑scale mass distribution is described by a small set of coefficients. We then apply a multi polynomial relationship to model the monthly and seasonal distribution of surface wind speeds given the knowledge of these few coefficients. We finally apply this model to seasonal forecasts from the European Center for Medium‑range Weather Forecasts (ECMWF) in an attempt of forecasting the monthly and seasonal distribution of the surface wind speed. For one month time‑horizon, the forecasting performance is superior to climatology.  Sep 7, 2016, 7:28 AM Peter Tankov
ċ

View
We study the optimal trading policies for a wind energy producer who aims to sell the future production in the open forward, spot, intraday and adjustment markets, and who has access to imperfect dynamically updated forecasts of the future production. We construct a stochastic model for the forecast evolution and determine the optimal trading policies which are updated dynamically as new forecast information becomes available. Our results allow to quantify the expected future gain of the wind producer and to determine the economic value of the forecasts.   Sep 7, 2016, 7:17 AM Peter Tankov
ċ

View
We focus on wind power modeling using machine learning techniques. We show on real data provided by the wind energy company Maïa Eolis that parametric models even following closely the physical equation relating wind production to wind speed are outperformed by intelligent learning algorithms. In particular, the CART-Bagging algorithm gives very stable and promising results. Besides, as a step towards forecast, we quantify the impact of using deteriorated wind measures on the performances. We show also on this application that the default methodology to select a subset of predictors provided in the standard random forest package can be refined, especially when there exists among the predictors one variable, which has a major impact.  Sep 8, 2017, 1:34 AM Lucie M
Ċ
View
Article "Tails of Weakly Dependent Random Vectors" by P. Tankov, published in Journal of Multivariate Analysis (Volume 145, March 2016, Pages 73–86). This article presents methods for quantifying the tail behavior of random vectors which can be used, for example, to evaluate risks due to spatial dependency in wind power production.    Jan 26, 2016, 8:13 AM Peter Tankov