Standard monitoring for solar assets is excellent at detecting total failures, but it consistently misses the subtle, cumulative underperformance that quietly costs asset owners 2-5% in power output, and millions in lost revenue. To solve this, Lamina developed a sophisticated analysis engine that uses a dual-model approach for high resolution accuracy and performance monitoring.

A physics-based model calculates your asset’s theoretical maximum output based on its specific hardware (panels, inverters) and high-fidelity meteorological data providing a performance benchmark. Simultaneously, a machine learning model (XGBoost) learns the asset’s unique, real-world performance signature from its own historical performance data, generating a high resolution, bespoke predictive model.

By forensically comparing the actual output against both the “perfect-world” theoretical model and the “real-world” learned benchmark, our system can precisely identify and diagnose a full spectrum of hidden losses: from long-term degradation to specific anomalies like soiling, clipping, or intermittent faults.

This transforms raw operational data into a clear, actionable report on recoverable revenue, forming the core of our “Hidden Yield-Loss Analysis” service.