evidoria

← Back to browse

Good practice Imported

Quartz Solar — Great Britain's Grid Operator Halves Solar Forecasting Errors with AI, Saving £30 Million a Year

United Kingdom · Warwick · See the United Kingdom profile

Evidence: Descriptive / self-reported Top 25% 73/100 · Ask Evidence Copilot about this practice

Great Britain's grid operator NESO deployed Quartz Solar, an AI forecasting tool built with non-profit Open Climate Fix. It halved solar-forecast errors, saving at least £30m a year in balancing costs, with up to £150m projected by 2035.

halved (~50) %
Reduction in large solar-forecasting errors (current)
~2.8 x
Accuracy improvement versus the previous forecasting tool (current)
at least 30 GBP million/year
Estimated annual savings in imbalance/reserve costs (current)
~300,000 tonnes/year
Estimated annual CO2 emissions avoided (current)
up to 150 GBP million/year
Projected annual savings by 2035 (projected to 2035)
1,000+ sites
PV sites used for historical generation data (current)
~300 supply points
Grid supply points forecast (current)
Quartz Solar — Great Britain's Grid Operator Halves Solar Forecasting Errors with AI, Saving £30 Million a Year

Details

Maturity
Established
Promoter
National Energy System Operator (NESO), with Open Climate Fix
Period
2024–2025
Keywords
energy, electricity grid, renewable energy, solar forecasting, public utilities

Context

Great Britain's electricity system is increasingly reliant on solar power, and the system operator, the National Energy System Operator (NESO), a public body since it was nationalised in October 2024, needed more accurate short-term forecasts of solar generation to avoid over-buying costly backup reserve.

Objectives

NESO worked with Open Climate Fix, a UK non-profit applying machine learning to climate problems, to build Quartz Solar, a tool to forecast solar output more accurately and more frequently than the system it replaced.

Activities

The tool combines live satellite imagery (12 spectral channels), weather data and historical generation data from over 1,000 PV sites to forecast solar output up to 36 hours ahead for around 300 grid supply points, updating every 15 minutes instead of the previous six-hourly cadence, with probabilistic (10th/50th/90th percentile) outputs.

Results

According to NESO and Open Climate Fix, Quartz Solar has halved large forecasting errors and is roughly 2.8 times more accurate than the tool it replaced. NESO estimates this avoids at least £30 million a year in imbalance/reserve costs and around 300,000 tonnes of CO2 emissions, with savings potentially rising to £150 million a year by 2035 as solar capacity grows.

Conclusions

The performance figures are self-reported by NESO and its non-profit technology partner rather than independently audited, and the tool addresses solar forecasting specifically rather than the full range of grid-balancing decisions.

Implementation
Indicative cost
Medium (€50k–€500k)
Time to results
Medium (1–3 years) — NESO was nationalised in October 2024; Quartz Solar has since become NESO's everyday forecasting tool, with projected savings rising through 2035 as solar capacity grows.
Staffing & skills
National Energy System Operator (NESO) control room team, Open Climate Fix, a non-profit machine-learning delivery partner

Conditions for success

  • Partnering with a mission-driven non-profit technology partner rather than a black-box commercial vendor
  • Combining multiple live data sources, satellite imagery, weather data and historical generation from over 1,000 PV sites
  • Probabilistic (10th/50th/90th percentile) outputs rather than single-point forecasts, to support risk-based reserve decisions

Common failure modes

  • Performance figures are self-reported by NESO and its non-profit partner rather than independently audited.
  • The tool addresses solar forecasting specifically, not the full range of grid-balancing decisions.

Where it fits

Governance type
national grid system operator (public body)
Scale
national (Great Britain transmission system)
Income level
high-income

Commonly funded by

National / regional programmes

Indicative funding routes for practices of this type — always check each programme's current calls and eligibility rules.

Do you run this practice? Claim it — verified implementers get a public contact pathway and can propose corrections.

Data sources

Where this practice's information was retrieved from, and when.

Attachments

Similar practices you may find useful