Grid-Tied Photovoltaic Performance Analysis Under Mediterranean Operating Conditions: A Residual Performance Diagnostic Framework

Conventional photovoltaic performance indicators provide useful benchmarking but do not determine whether production deficits arise from expected weather and module effects or from unresolved plant-side behaviour. This study develops a residual performance diagnostic framework for data-limited grid-connected photovoltaic plants and applies it to two adjacent fixed-tilt crystalline-silicon systems, each rated at 99.36 kWp, in South Corfu, Greece, over 2016–2023. An expected-energy baseline was constructed using hourly irradiance and meteorological data from the Photovoltaic Geographical Information System after explicit correction for incidence-angle effects, irradiance-level response, module temperature, and first-order air-mass and spectral effects. Measured alternating-current energy delivered to the grid, obtained from the Hellenic Electricity Distribution Network Operator, was then compared with the corrected expected energy through a weather-adjusted system-efficiency indicator and its complementary residual-loss coefficient. The explicit module and weather loss envelope remained close to 9.01%, with temperature as the largest modelled component. Residual losses increased from 10.76% to 14.53% in Plant 1 and from 10.51% to 16.26% in Plant 2. The fitted apparent annual declines were 0.453 percentage points per year for Plant 1 and 0.609 percentage points per year for Plant 2. A generation-hour uncertainty analysis based on independent observations from the Hellenic National Meteorological Service gave expanded uncertainties of ±5.80% for expected energy and ±5.89% for the system-efficiency indicator. The framework is therefore presented as an uncertainty-bounded screening method for prioritizing inspection and maintenance, not as a root-cause diagnostic or a formal performance-loss-rate assessment.