State of the Art and Research Gaps


State of the Art and Research Gaps

6 years ago, we started IEA Wind Task 36 "Forecasting for Wind Energy" with a workshop in Barcelona on the State of the Art and Research Gaps. Now, at the beginning of Task 51 "Forecasting for the Weather Driven Energy System", we did repeat the success with a workshop on 12 and 13 September 2022 at University College Dublin, with a Task 51 meeting following the day after. Download flyer.

NEWS: The entire workshop (minus the bits which were not amenable for online presentation, like the Open Space discussion sessions) is now online in our YouTube channel: Day 1 / Day 2.


Please note that not all talks are yet on the website, the program keeps being updated.

Day 1

9:00-10:30. Session host: Gregor Giebel, DTU
Welcome and Keynotes: The power system of Ireland is standing as a prime example of the challenges encountered in the energy transition: high shares of renewables, low transmission capacities elsewhere, and growing demand for transport and heat to come from electricity.

Gregor Giebel, Operating Agent, DTU: Introduction to IEA Wind Task 51 “Forecasting for the Weather Driven Energy System”
Keynote by Conor Kavanagh, Eirgrid on Challenges with running the Irish system securely
Keynote by Eamonn Lannoye, EPRI: Integrating forecasting into power system operations and planning – current gaps and research needs

11-12:30. Session host: Jethro Browell, WP lead, University of Glasgow
Keynote from Traders: Physical availability of renewable power is one thing, but in the current setup, the power is traded on the markets. How this is done in the view of changing demands, optimisation across different time scales, the use of storage and the inability to see all data, is the topic of this session.

Edward McGarrigle, Galánta Energy: On Forecasting for Trading
Malte Rieck, Vattenfall: How are wind forecasts used operationally? A tour of marketing wind energy
Kathryn Fowler, Centrica Energy Trading: Changing forecast requirements in a trading environment

13:30-15. Session host: John Zack, UL Renewables
Weather forecasting for renewable power: For timescales from intraday to seasonal, weather prediction plays a major role for the forecast accuracy and reliability. This session will go into the details of new products and new developments in weather forecasting for renewables.

Caroline Draxl, co-Operating Agent, NREL: The WFIP projects 1-3 and Forecasting for Airborne systems.
Irene Schicker, ZAMG: Subseasonal Forecasting
Frédéric Vitart, ECMWF: Season to Season (S2S) forecasts and their relevance for energy
Remco Verzijlbergh, Whiffle: Very high resolution forecasting using LES on GPUs

15:30-17. Session host: Corinna Möhrlen, WP lead, WEPROG
Extreme event definition and forecasting: Not only the weather will get more extreme in future climate scenarios, the impacts on the power system also will get more extreme. But what is an extreme event for the power and energy system: extreme weather, or can “boring” weather with low wind and sun constitute problems too? The session starts with a short introduction, followed by an Open Space discussion and presentation of the results in plenum.

Day 2

9-10:30. Session Host: Ricardo Bessa, WP lead, INESC TEC
Data for forecasting: Many forecasting models are data driven, and data is developing into a new commodity for both meteorological and power forecasts. Which parts of the data can be open, how could data markets work, and how to generate simulated data for training the models?

Bradley Eck, IBM: AI and cloud computing developments towards integrating renewables
Ricardo Bessa, INESC TEC: Smart4RES collaborative analytics for renewable energy forecasting: federated learning and data markets
Paul Cuffe, UCD: Prediction markets as forecasting and hedging instruments within the renewable electricity sector
Discussion speakers & audience (20 min)

11-12:30. Session host: Gregor Giebel, Operating Agent, DTU
Open Space session: Your opportunity to come with input to the current state and the research gaps in three different applications: minute and hour scale forecasting, intraday and day-ahead forecasting, and week-ahead to season-ahead forecasting. We will separate into discussion groups, and the discussion leaders will report back to the plenum what the individual groups wrote down.

Bonus presentation: Kristian Horvath, WMO: WMO Study Group on Integrated Energy Services

13:30-15.00. David Lenaghan, UK NationalGrid ESO
The Journey from Deterministic to Probabilistic: Academia has shown for decades how to prepare probabilistic forecasts, and has shown that using them for decision making would yield benefits. Despite this, and despite the fact that all forecast vendors provide probabilistic forecasts, the use of them at the end users has been hesitant. In this session we discuss how the journey from deterministic forecasts to probabilistic forecasts can work in practice. We round off with playing a forecasting game motivating the benefits of probabilistic forecasts.

Introduction by David Lenaghan
George Kariniotakis, Mines Paris: How do the applications influence the forecasting tech evolution?
Juan Sopena, Solute: People just want numbers - How to fairly compare and interpret forecasts with a benchmarking framework for performance evaluation
Corinna Möhrlen, WEPROG: Introduction to the probabilistic gaming concept, the games purpose and setup to demonstrate a real-time environment and how to integrate probabilistic forecasts into decision making processes (automatically as well as with human intervention) (10min ).
Playing the IEA Wind/WEXICOM Forecast Game

15.10 -17.00. Session host: Gregor Giebel, Operating Agent, DTU
Panel discussion and Q&A based on Open Space results regarding identification of research gaps and research overview, defining the final list of research gaps. We will together identify and prioritise the research gaps and future research in the sector.


Day 3: Task 51 meeting (Task members only), 9-13.



From the general Work Stream description:

The established state-of-the art will be carried forward in the recommended practice guideline for forecasting solution selection and its dissemination to the industry at workshops, webinars, conferences, white papers and a book publications. Every WP contributes to this activity.

D 2.1: Workshop and paper on state-of-the-art and future research issues in the forecasting of weather-dependent energy system variables (M7=Summer 2022, M12=Dec 2022)

M 2.1: Version 3 of IEA Recommended Practice on Forecast Solution Selection (M36=Dec 2024)

WS: WP1 Weather WP2 Power WP3 Applications Deliverable #, Due Collaboration
State of the Art for energy system forecasting (WP2)



Workshop / Paper


RecPract on Forecast Solution Selection v3

D2.1 / M7, M12


M2.1 / M36

PVPS Task 16, Hydro TCP, Hydrogen TCP, …