Nova-sight | ESA Space Solutions

Objectives of the service

Solar energy production is inherently unpredictable — clouds cause output to fluctuate within minutes, creating costly imbalances for energy traders, grid operators, and industrial sites trying to match consumption with local generation.

Nova-sight addresses this by delivering site-specific solar power forecasts with unprecedented accuracy and speed. The service ingests near-real-time imagery from the Flexible Combined Imager (FCI) aboard the Meteosat Third Generation satellite MTG-I1, which scans Europe at 0.5 km resolution every 2.5 minutes. Combined with on-site measurements and AI-based models, this enables cloud movement to be tracked and translated into actionable power forecasts — updated every few minutes and delivered via a REST Application Programming Interface (API).

Users benefit in three concrete ways: industrial companies and energy communities maximize self-consumption of locally generated solar power; Balance Responsible Parties reduce imbalance penalties and improve trading decisions; and grid operators gain sharper visibility into short-term renewable variability.

This project builds and validates the full technical pipeline — from satellite data ingestion to forecast delivery — and demonstrates its value at a real pilot site. It establishes pricing, integration pathways, and commercial readiness, positioning Nova-sight as a standalone API product and as a core enhancement to the existing Fore-sight energy optimization platform.

Users and their needs

Targeted User Communities

Nova-sight targets three user groups:

Energy communities and industrial consumers (Transfo Leiedal, AZ Groeninge)

Users: automation engineers and energy managers

Need to match solar production with local consumption and flexible assets (cooling, electric vehicle charging, storage)

Challenge: forecast latency and accuracy currently too low to act on sub-hourly solar variability

Balance Responsible Parties (BNewable, Ecopower)

Users: trading engineers and portfolio managers

Need site-level and aggregated solar forecasts updated within minutes to optimize intraday trading and reduce imbalance penalties

Challenge: no existing commercial service exploits MTG-I1 satellite data for this purpose; current solutions lack the required temporal resolution

System integrators and energy consultants (Ikologik, Oktow)

Users: automation engineers deploying energy management solutions for industrial small and medium enterprises

Need a reliable, easy-to-integrate forecasting data stream via a standardised programming interface

Challenge: building in-house nowcasting is prohibitively expensive; off-the-shelf solutions do not yet meet accuracy requirements

Cross-cutting challenge: all user groups require forecasts with update cycles below five minutes and latency below one minute — a performance threshold that drives the core technical ambition of this project.

Service/ system concept

What Nova-sight delivers to users

Nova-sight provides site-specific solar power forecasts, updated every few minutes and accessible via a web-based data interface. Users receive:

Short-term solar production forecasts (0–6 hours ahead) at their specific location, refreshed as new satellite images arrive

Cloud movement maps showing incoming cloud cover in near real time

Uncertainty estimates alongside each forecast, supporting risk-aware decisions

A training interface allowing users to upload their own historical data and personalise the model to their site

With these capabilities deployed, an energy manager can automatically pre-charge a battery before a cloud arrives, a logistics operator can schedule truck charging in sync with solar peaks, and a trading engineer can adjust energy market positions minutes before an imbalance occurs — all without manual intervention.

How it works (in simple terms)

A geostationary satellite photographs Europe every 2.5 minutes at high resolution. Nova-sight ingests these images, detects where clouds are and where they are heading, and combines this with on-site measurements and weather data to predict solar output at each user's location. The result is delivered as a data stream that plugs directly into existing energy management software.

Space Added Value

Space Asset Used

Nova-sight relies on the Flexible Combined Imager (FCI) aboard MTG-I1, the first satellite of the Meteosat Third Generation programme operated by the European Organisation for the Exploitation of Meteorological Satellites (EUMETSAT). Launched in December 2022 and fully operational since December 2024, MTG-I1 is a geostationary satellite that continuously observes Europe across 16 spectral channels at a spatial resolution of up to 0.5 km and a refresh rate of every 2.5 minutes.

Added Value Over Existing Solutions

Current solar forecasting competitors either rely on ground-based sky cameras (e.g. Reuniwatt), numerical weather prediction models, or lower-resolution legacy satellite imagery. These approaches share a fundamental limitation: they cannot track cloud formation and movement fast enough or at fine enough spatial scales to support sub-hourly, site-specific forecasting.

MTG-I1 changes this entirely. Its combination of high spatial resolution and rapid refresh rate enables cloud patterns to be detected and tracked in near real time — something no existing commercial nowcasting service currently exploits. The result is:

Forecast updates every 2–3 minutes
Site-specific accuracy at scales below 1 km

Current Status

During this period, the Nova-sight project moved from infrastructure buildout into active model development and satellite data acquisition. The satellite dish and EUMETCast Europe reception equipment were procured, with installation now pending final insurance approval; in parallel, model training was launched on VSC (Flemish Supercomputing Center) GPU infrastructure. Model validation is ongoing using PV measurement data from our pilot sites. On the commercial side, the project reached new milestones with contracts signed with Kortrijk Business Park and AZ Groeninge. Meetings are scheduled with balance responsible/service providers to validate the monetisable value of improved near-real-time forecasting for imbalance cost reduction.