Solar
AI forecasts solar output hours ahead, finds panel faults from drone imagery before they cut yield, and flags soiling and inverter problems as they start. Fewer surprises for the grid, more energy off the same array.
- Output-forecasting models built on your plant and weather data, so dispatch and curtailment are planned, not reacted to.
- Computer-vision inspection on drone thermal and RGB imagery to catch cracks, hot spots, and soiling early.
- Performance monitoring wired into your SCADA to flag inverter faults and degradation in real time.
- Any of it on your own hardware when site data must stay on your network.
The UK grid operator (NESO), with the Alan Turing Institute, uses machine learning to sharpen solar forecasting and balance the grid.
National Grid ESO & The Alan Turing Institute — ML to balance the GB grid, 2024 (opens in new tab)