Trusted by 100+ MWp of Solar & BESS Operations
Technical Whitepaper • 8 pages
Digital Twin Configuration Guide
Complete reference for configuring, tuning, and experimenting with NuraVolt digital twin models including CatBoost hyperparameters.
The Problem:
Default model configurations may not be optimal for all plants. Operators need guidance on tuning parameters for their specific data quality and operational requirements.
What You'll Learn:
- 11 physics-informed features: Core (irradiance, temps), Temporal (day/hour/month), Solar geometry, Derived
- CatBoost hyperparameters: iterations, learning_rate, depth, L2 regularization, early stopping
- Data filtering: MAX_YEARS, MIN_PR, MIN_IRRADIANCE, VALIDATION_SPLIT
- Quality thresholds: R² > 0.70, MAE < 50 kW for production deployment
- Performance matrix: Parameter changes vs training time, accuracy, overfitting risk
- Quick trial guide: --limit 5 --digital-twins-only for rapid experimentation
Based on: Research from NREL, Sandia Labs, IEEE studies, and operational data from 100+ MWp of GCC solar & BESS assets.
Want Custom Analysis for Your Plant?
Our UAE-based team can provide a free 15-minute data quality audit and custom ROI calculation.
Contact Us →