Delegates are invited to meet and discuss with the poster presenters during the poster presentation sessions between 10:30-11:30 and 16:00-17:00 on Thursday, 19 November 2015.
Lead Session Chair:
Stephan Barth, ForWind - Center for Wind Energy Research, Germany
Hans-Peter (Igor) Waldl (1) F
(1) Overspeed GmbH & Co. KG, Oldenburg, Germany
Printer friendly version: Print
Presenter's biographyBiographies are supplied directly by presenters at EWEA 2015 and are published here unedited
Physicist and holding a PhD ("Modelling and Optimisation of Wind Farms". Since 25 years inolved in wind energy. Managing director of Overspeed.
PosterDownload poster (6.33 MB)
The Anemos prediction system as generic platform for forecast handling, optimisation and benchmarking
Since many years, the Anemos prediction platform has been developed in several research and commercial contexts. The most challenging environment where this prediction system is used are the in-house forecasts for the Australian Energy Market Operator AEMO producing 10 day forecasts with an update frequency and time resolution of 5 minutes, a high-availability implementation which operates with 100 % availability since 6 years.
Many of these developments are generic, and may be used for the implementation of individual, but standardized wind and solar power prediction systems.
In practical all prediction applications, there are tasks which can be tackled on a generic level. For these tasks, a mature and proven solution with a high level of quality management and resilience may be an alternative to a dedicated in-house development. In our presentation, we present how the Anemos prediction platform may contribute to the handling of general prediction challenges and set free person power and resources to concentrate on the prediction quality instead of solving IT issues
Main body of abstract
In practical all prediction applications, there is a number of tasks which may be solved generically and don’t need to implemented individually:
* Data handling, acknowledgement, fail-over and storage, as well for time series data as for static information, including data archiving.
* Optimal combination of several prediction feeds in order to gain an improved combines forecast. For this task, the Anemos system uses a generic adaptive combination model from ENFOR, DK, which learns from historical predictions and selects the weights between the prediction feeds automatically.
* Data presentation via a graphical user interface, of time series data as plots, tables and as derived information like deviations between predictions and realized power outputs.
* Reporting of forecast accuracy as NMAEs, RMSEs, scatter plots and correlation analysis
* Reporting and surveillance of data feeds, prediction data availability, …
* Guaranteeing a high system availability by mirror sites, internal robust data handling, automatic data export failover, and data feed redundancy
All these tasks could and are solved in the Anemos prediction platform.
By using a prediction platform which is developed under a high level of quality management and which is tested intensively by a energy market operator, the Anemos prediction platform may contribute to a higher accuracy of wind power forecasts with lower costs and efforts, or may set free more resources for prediction services and saving effort for implementing the prediction handling.
End-users and prediction service providers will develop an idea where they may benefit from existing developments and systems and how to introduce these developments in their own prediction implementation.