Welcome to F97.BE!

Python, Photons, and Predictions.

Energy Production Forecast

Forecasting solar energy is harder than it looks. GTI forecasts, grounded in physics and solar geometry, are transparent and reliable when skies are clear, but they struggle with clouds and local effects. AI forecasts learn from years of production data and adapt to site-specific conditions, yet they can misfire under unusual or shifting patterns. Climate change makes this challenge greater: as the world warms, the atmosphere grows hazier and less predictable. AI often reads rising temperatures as more sun, when in reality haze cuts production sharply. It will take years of new training data before AI fully adjusts.

This is why I developed the Fusion Forecast, a key step in solar prediction. It unites the learning power of AI with the physics of GTI, correcting biases and dynamically balancing the most reliable source. Fusion adapts immediately, providing forecasts that are resilient and realistic under changing climate conditions. Alongside Fusion, you can explore my 24-hour energy production forecast for the House Roof System, generated with a custom AI model trained on more than 1500 days of historical weather and production data. The dashboard visualizes real-time AI predictions, GTI-based forecasts, and actual inverter output in a single chart, and includes key performance metrics (, MAE) as well as historical accuracy tables.

Each day, two models – Random Forest and Gradient Boosting – are evaluated, with the system automatically selecting the best performer for the AI forecast. In parallel, I also calculate a GTI forecast, based purely on solar geometry and irradiance predictions. These two streams are then fused into a third column: the Fusion forecast, which learns the optimal weighting between AI and GTI, adds an offset if needed, and falls back to a 70/30 blend when training data is sparse. This way the forecast stays stable while adapting to changing weather and seasonal patterns.

Download the paper: Fusion Forecasting for Residential PV: An Adaptive Regression Blend of AI and GTI with Online Calibration.

The AI model was last updated on 25.08.2025, the GTI model on 07.07.2025. The real-world dataset was last refreshed on 28.08.2025 and currently contains 1507 entries up to 22.08.2025.

Forecast Summary

7 Days Forecast Comparison AI and GTI

home | top