Improve forecasting in different timescales

When: Thursday, 23 June 2022, 11:00 - 12:30
Where: Auditorium

Session description

Although it is a standard today to use forecasts of wind production to manage power systems with high wind penetration and to optimise market participation of wind farms, there are still situations where prediction errors are high and translate into technical and financial impacts for the operators. These impacts stimulate continuous research in the field worldwide. This session aims to present selected recent advances on very short term forecasting and ramps forecasting using SCADA data, and also on seasonal forecasting for the longer time scales. The session finally emphasises on the importance of focusing not only on the accuracy of the forecasts but also on the value they bring to the applications they are used for as input. To that direction, novel developments on the optimal use of forecasts for market participation and the provision of ancillary services will be presented.

Session chair

Georges Kariniotakis

Professor, Head of Renewable Energies & SmartGrids Group, Centre PERSEE, MINES Paris, PSL

Presentations

SCADA based dynamic short-term forecasting via transfer learning

Tuhfe Göçmen

Researcher, DTU Wind Energy

The prediction of ramp events by means of an observer-based power forecast utilizing lidar and SCADA data

Frauke Theuer

research assistant, ForWind - University of Oldenburg

Seasonal forecasting: beating climatology with machine learning

Gerard Castro

R&D Data Analyst, Vortex FdC

Advanced Model Predictive Control Strategies for the Provision of Energy and Ancillary Services by Wind Farms Coupled with Storage

Simon Camal

Research engineer, MINES Paris - PSL University

Analysis of optimal use of wind power forecasting for Energy management system solution to spot markets participation

Jesús López Merino

Control Algorithms Engineer, Siemens Gamesa Renewable Energy