Space-time trajectories of wind power generation: Parameterized precision matrices under a Gaussian copula approach

Publication: Research - peer-reviewBook chapter – Annual report year: 2015

View graph of relations

Emphasis is placed on generating space-time trajectories of wind power generation, consisting of paths sampled from high-dimensional joint predictive densities, describing wind power generation at a number of contiguous locations and successive lead times. A modelling approach taking advantage of the sparsity of precision matrices is introduced for the description of the underlying space-time dependence structure. The proposed parametrization of the dependence structure accounts for important process characteristics such as lead-time-dependent conditional precisions and direction-dependent cross-correlations. Estimation is performed in a maximum likelihood framework. Based on a test case application in Denmark, with spatial dependencies over 15 areas and temporal ones for 43 hourly lead times (hence, for a dimension of n = 645), it is shown that accounting for space-time effects is crucial for generating skilful trajectories.
Original languageEnglish
Title of host publicationModeling and Stochastic Learning for Forecasting in High Dimensions
EditorsAnestis Antoniadis, Jean-Michel Poggi, Xavier Brossat
PublisherSpringer
Publication date2015
Pages267-296
ISBN (print)978-3-319-18731-0
DOIs
StatePublished - 2015
EventInternational Workshop on Industry Practices for Forecasting - Paris, France

Workshop

WorkshopInternational Workshop on Industry Practices for Forecasting
CountryFrance
CityParis
Period05/06/201307/06/2013
SeriesLecture Notes in Statistics
Number217
ISSN0930-0325
CitationsWeb of Science® Times Cited: 2
Download as:
Download as PDF
Select render style:
APAAuthorCBE/CSEHarvardMLAStandardVancouverShortLong
PDF
Download as HTML
Select render style:
APAAuthorCBE/CSEHarvardMLAStandardVancouverShortLong
HTML
Download as Word
Select render style:
APAAuthorCBE/CSEHarvardMLAStandardVancouverShortLong
Word

Download statistics

No data available

ID: 93684427