Running a high-voltage transmission grid is work in the shadows: it rarely goes wrong. Yet, a lot of challenges are faced every day. Reliable forecasting is at the heart of this challenge. A good understanding of weather will lead to better forecasts, yet the landscape is changing fast. Having accurate solar and wind forecasts is not enough; better insights in market dynamics and new...
In recent years, the Royal Meteorological Institute of Belgium (RMI) has positioned itself strongly in the field of renewable energy forecasting and its applications. This was made possible by leveraging scientific collaboration within various projects. In this presentation, we give an overview of applications such as the RMI Storm Forecast Tool developed for the Belgian TSO (Elia), and...
Over more than half a century, Numerical Weather Prediction (NWP) has achieved remarkable progress. This advancement results from the use of increasingly dense observational data types, progress in high-performance computing, and steady improvements in numerical modeling. This long-standing progress faces new challenges, with standard technologies approaching their physical limits, and models...
The on-demand Extremes Digital Twin Renewables models, developed by the Destination Earth team, provide a configurable workflow linking the detection of adverse weather, hectometric and sub-hourly NWP simulations, and impact models for wind and solar energy. Targeted forecasts of renewable energy production under extreme conditions are generated through dynamic triggering based on...
We present evaluations for the prediction of wind power ramping events in the Belgian Offshore Zone. We verify two models from Royal Meteorological Institute of Belgium: the operational ALARO-4km and its version with Wind Farm Parameterization (WFP). Power predictions are generated with power curves and machine learning (ML). As standard metrics like MAE are insufficient for evaluating ramps,...
This study investigates how climate change could affect the energy output of solar panels in Belgium, with projections extending to the year 2100. It offers a long-term perspective by starting with weather data from 1950 and building on future climate models developed by the Royal Meteorological Institute (RMI). By analysing multiple climate scenarios, the study examines potential shifts in...
The HIRENEXT project aims to co-design and implement a Pilot Service tailored to dynamic line rating (DLR) computations for electricity transmission. This service leverages innovative research from the Destination Earth On-Demand Extremes Digital Twin contract (DE_330), which provides access to unprecedented high-resolution weather predictions.
Through a user-centric co-design approach,...
EnergyProtect aims to identify present and future risk hotspots, defined as sites of renewable energy infrastructure with elevated exposure to potentially disruptive meteorological conditions. We i) use physics-informed ML to detect patterns of adverse weather, ii) dynamically downscale ensemble time slices to convection permitting resolutions, and iii) estimate uncertainties, return periods...
Elia Transmission Belgium operates a grid with a high penetration of PV-production capacity, offshore wind parks, DC cables and a particular geographical position, between the biggest consumers of the European grid – Germany and France. The exploitation happens parallel to one of the most challenging infrastructure portfolios in the whole of Europe, implying many outages happening at any given...
As the Met Office retires legacy deterministic data feeds and moves towards ensemble forecasting, a challenge arises in maintaining delivery of spot forecast products, including those supporting renewable energy applications. Rather than replicating existing data feeds, a new project is taking a fresh approach to understand requirements. Starting with plain-English questions to understand how...
Accurate forecasts of regional wind power production are crucial for power system operation and planning. Total generation in the coming hours and days depends strongly on both weather forecasts and time-varying production capacity. We propose a probabilistic forecasting approach based on Bernstein Quantile Networks (BQN) to predict aggregated power production from multiple wind farms without...
The ENTSO-E Transparency Platform provides comprehensive European electricity market data essential for renewable energy forecasting and operations. We present entsoe-apy, an open-source Python library enabling access to all ENTSO-E RESTful API endpoints with seamless data retrieval.
The package features automatic request splitting for large queries, intelligent retries, and consistent...
Accurate weather forecasting plays an increasingly important role in today's society, with implications in hydrological modelling, renewable energy production, and civil service operations. To improve the reliability of ensemble weather forecasts, post-processing of said forecasts is frequently employed. However, many variables of interest, such as precipitation or wind speed, exhibit highly...
Wind power central to the energy transition, yet its variability challenges accurate forecasting. This work introduces an adaptive nowcasting method for probabilistic forecasting method combining the generalised logit transformation with a Bayesian framework. The transformation maps double-bounded wind power data to an unbounded domain, enabling Bayesian inference, while an adaptive mechanism...
RUSH (Rapid Update Short-term High-resolution) is an AI-native, rapid-update nowcasting framework that turns heterogeneous inputs into probabilistic 0–24 h forecasts at 30-min steps on a 1-km grid over Belgium. Methodologically, RUSH couples two ingredients: (i) observation-led evolution learned from 30-min radar accumulations and 15-min SEVIRI channels, and (ii) large-scale dynamical context...
With offshore wind farms increasing in size and number, their interaction with regional wind and weather patterns plays an increasingly important role. Wind turbines extract energy from the boundary layer, resulting in downstream wakes with reduced wind speeds. When clustering many turbines together, wakes coalesce and can persist for tens of kilometers, affecting neighboring farms and...
Accurate day-ahead wind power forecasts are essential for wind farm operation strategies and grid capacity planning. This presentation demonstrates probabilistic day-ahead wind power forecasting using gradient boosting trees. We compare three probabilistic prediction methods - conformalised quantile regression, natural gradient boosting and conditional diffusion models - combined with...
AI-driven Weather Prediction Models (AIWPMs) are revolutionizing weather forecasting and surpassing traditional numerical weather models in both accuracy and computational efficiency. With weather conditions impacting numerous sectors, such as renewable energy generation, this has far-reaching implications. AIWPMs could enable improved forecasts for renewable energy generation, such as wind...
This study investigates the application of Encoder–Processor–Decoder architectures, which have been successfully employed in large machine learning–based weather forecasting models, for wind power prediction. Using the Anemoi framework from the European Centre for Medium-Range Weather Forecasts (ECMWF), we evaluate graph-based neural networks over a domain covering Belgium and the Belgian...
Solar energy forms an important pillar of climate change mitigation. Short-term forecasts of surface solar irradiance (SSI) are gaining more importance for power grid operators seeking to balance supply and demand in a secure and economical way. Regional-scale SSI forecasts are essential since most solar power is provided by decentralized PV plants. Solar nowcast models offer SSI predictions...
Post-processing methods are implemented at wind/solar farm for the power forecast. We propose two forecasting frameworks based on EMOS and neural network (NN). Boosting technique is applied for selection of the variables in the EMOS approach. The NN approach uses a multilayer perceptron model, where the training process is specifically tailored to the limited dataset size of one year by...
Increasing shares of distributed PV generation plus increased wind energy being fed into the grid challenge local grids through rising variability and peak loads, requiring accurate, high-resolution forecasts for grid flexibility. EngagePrivFlex addresses this by exploring how private households can provide flexible generation & consumption to support grid stability. The meteorological...
