Investigations on LAM ensembles for wind power prediction

Principal Investigators

Dr Kai Sattler
Danish Meteorological Institute
Lyngbyvej 100
2100 Copenhagen
Denmark

ksa@dmi.dk

Project description

The special project Investigations on LAM ensembles for wind power prediction will investigate the potential of LAM ensembles for wind forecasting and wind power prediction. The work in this special project is part of the Danish PSO-funded project Wind power prediction by ensemble forecasting, which is supported under the reference number ORDRE-101295 (FU2001).

The aim of the PSO funded project is to look into the possibilities of improving wind power forecasts used at utilities for scheduling power plants and electricity trading. This includes a study on representation of uncertainty, which is to be addressed by using ensemble forecasts. At the same time, the forecast horizon shall be extended in order to facilityate maintenance schedules of the electricity grid and of o_shore wind farms. Therefore, several NWP models will be utilized in the project. beyond which is the ECMWF-EPS in order to cover medium range lead times up to 7 days ahead.

Besides the use of the global weather prediction models from ECMWF, DWD and NCEP, the DMI-HIRLAM shall provide short-range ensembles of high resolution LAM forecasts in order to investigate their potential for wind power prediction. In this context, DMI will perform investigations on the design of a suitable and feasible LAM ensemble, which will include different perturbation approaches in order to cover the effects of uncertainties emerging from the initial and the boundary condition, as well as uncertainties arising during the model simulation. The coupling of DMI-HIRLAM to selected members from the ECMWF-EPS is hereby one major aspect, because this guarantees consistency between initial perturbation and lateral boundary perturbation.

DMI has already some experience with ECMWF-EPS based LAM ensembles from the research done within the special project Heavy rain in Europe, which included experiments with ensembles applied on heavy rain events. In this project, we aim at designing a LAM ensemble for use in wind power prediction and investigate its potential by applying common verification methods for ensemble prediction systems. In addition, the verification will also indicate the potential for the general use of high-resolution, short-range ensemble forecasts based on the ECMWF-EPS.

The ensemble simulations with DMI-HIRLAM based on the ECMWF-EPS will be carried out at the computer facilities at ECMWF. They necessitate large amounts of computer resources, both CPU time and storage capacity in the ECFS system. In order to limit the latter, ECTMP will be utilized where appropriate.

For more details, please refer to the latest progress report.

Additional information

Project started in 2003.