Minute scale forecasting
| WS: | WP1 Weather | WP2 Power | WP3 Applications | Deliverable | #, Due | Collaboration |
| Minute scale forecasting (WP2) | X | Workshop / Paper | D2.5 / M31, M36 | Wind Tasks 32 Lidar, 44 Farm Flow Control and 50 Hybrids, PVPS T16 |
On the power plant level, forecasts some minutes ahead can be used for battery control in hybrid power plants, in wind farm flow control (it takes minutes for the wind field to pass through a larger wind farm), and sometimes also in market structures like the Australian market, which operates on a 5-min schedule. Advances in minute-scale forecasting have been investigated in phase 2 and will be further developed and communicated to the industry. Since minute scale forecasting mainly uses data driven tools (statistical or machine learning), the WS is administered by WP2, but has connections to WP1 for knowing the wind flow through a farm, and to WP3 with regards to usage of the forecasts. We plan to have a workshop together with the IEA Wind Tasks on Lidar and on Hybrid Power Plants, and possibly others.
D 2.5: Workshop and paper on minute-scale forecasting for hybrid power plants or wind farm control, in conjunction with Task 32 on Lidars, Task 44 on Farm Flow Control and Task 50 on Hybrid Power Plants (M31=Summer 2024, M36)
Venue: DTU Risø Campus, Roskilde, Denmark
Attendence: 70+ in-person, 20 online.
Collaboration of multiple IEA tasks:
Forecasts of wind and clouds are important inputs for the control and value of renewable power plants. The forecasts on a time resolution of minutes or seconds are typically data driven, looking at upstream plants, all-sky images of clouds or direct measurements of wind by lidars.
To facilitate the dissemination of information about minute scale forecasting products, skill, applications, issues, and best practices to members of the electric energy community, we invited in April, 10-11 2024 to a Minute Scale Forecasting workshop with the goal of gathering information about:
- methods used to produce the forecasts
- the current state-of-the-art skill and uncertainty in forecasting for variables on high temporal resolution
- current and planned research activities intended to improve the current level of skill
- types of public and private sector operational forecasting products
- the range of minute scale applications in the energy community
- quantified or perceived value obtained from those applications
- the sensitivity of user’s application performance to variations in forecast skill
- the unmet minute-scale forecasting-related needs or desires of the energy user community
Program:
Wednesday, April 10, 2024
09:00 – 10:30 Welcome and Keynote
10:30 – 10:45 Networking Break
10:45 –12:15 Methods of minute scale forecasts and their uncertainty
12:15 –13:15 Lunch
13:15 –14:45 Applications of minute scale forecasts for control, ramps and extreme events
15:00 – 16:30 Open Space Discussion
18:30 Informal dinner gathering
Thursday, April 11, 2024
9:00 – 10:30 Data driven solar forecasting
10:30 – 10:45 Networking Break
10:45 – 12:15 Data driven wind forecasting
12:30 – 13:15 Conclusions and Lunch
Presentations:
01 Challenges and where we see the greatest benefit from accurate high-resolution forecasts
Martin Lørup Stensdal
02 Probabilistic solar forecasts as a binary event using sky camera
Methieu David
03 Minute-scale wind forecasting introduction
Elliot Simon
04 Spatio-temporal solar forecasting
Rodrigo Amaro e Silva
05 Benefit of an All Sky Imager Network for Satellite-Based Solar Irradiance Forecasts
Jorge Lezaca
06 Enhancing PV Forecasting through Sky Imaging and Irradiance Analysis
Andreas Boschert
07 High resolution solar irradiance forecasting using All Sky Imagers and machine learning
Nils Straub
08 Extreme wind and solar ramping events in All-Island of Ireland
Ndamulelo Mararakanye
09 On the benefits and challenges of high-resolution ensemble forecasts for minute-scale ramping applications
Corinna Möhrlen
10 Reducing power system imbalance forecast errors using short-term wind (and solar) power forecasts
Jon Olauson
11 Leveraging Generative Models for Enhanced Solar Irradiance Ramp Detection
Yann Fabel
12 Graph machine learning for short-term PV forecasting
Rafael Carrillo
13 Evaluation of the impact of local sensing data and feature extraction techniques on high frequency intra-hour irradiance forecasting
Erling W. Eriksen
14 Predicting surface solar irradiance from satellite imagery with deep learning radiative transfer emulation
Angela Meyer
15 Measuring the unpredictable.
Robin Conseil
16 AI based weather prediction for wind power forecasting
Jie Yan
17 Prediction of ramp events using lidar-based minute-scale power forecast
Janna Kristina Seifert