Why Installer Yield Quotes Diverge by 25% for the Same Roof
Two quotes for the identical 6.6 kWp rooftop system can easily promise 7,400 kWh/yr on Quote A and 9,800 kWh/yr on Quote B. Almost never does hardware efficiency explain that 2,400 kWh gap. What separates them is whether the estimator ran a physics-bounded irradiance model (such as the European Commission's PVGIS 5.3 or NREL PVWatts v8) or multiplied optimistic peak-sun hours by STC nameplate wattage.
Because both PVGIS and NREL PVWatts publish open, non-commercial hourly datasets, you can audit any installer claim in five minutes without trusting a proprietary sales simulator.
1. What PVGIS and PVWatts Actually Measure
Every photovoltaic simulation splits yield into two distinct layers:
- Solar Resource on the Tilted Plane (H_POA): Satellite-derived Global Horizontal Irradiance (GHI) and Direct Normal Irradiance (DNI) transposed onto your roof azimuth and tilt angle.
- Conversion Chain Losses (Performance Ratio, PR): Real electrical and optical losses separating Standard Test Conditions (1,000 W/m², 25 °C cell temperature) from AC kilowatt-hours delivered to your meter.
| Simulation Step | Honest Residential Input | Sales-Inflated Trap | Impact on Annual kWh |
|---|---|---|---|
| Radiation Database | PVGIS-SARAH3 / NREL NSRDB | Clear-sky theoretical | +8% to +15% inflation |
| System Losses | 12% – 14% combined | 4% – 6% optimistic | +7% to +10% inflation |
| Horizon / Shading | Local terrain + chimney mask | Unshaded flat horizon | +5% to +25% inflation |
| Module Thermal Model | Close-mounted NOCT | Free-standing open rack | +2% to +5% inflation |
2. Step-by-Step Auditable PVGIS Workflow
- Pin Exact Coordinates: Enter latitude and longitude to four decimal places rather than city center (coastal fog or valley shadows shift irradiance steeply within 10 km).
- Select Mounting Position: Choose Building integrated / close-mounted for flush rail-mounted residential arrays (<150 mm roof standoff). Poor backside airflow elevates cell temperature by 8–12 °C over open ground racks.
- Enter Installed Peak Power in kWp: Use total STC DC wattage (N × Pmax), not the inverter's AC continuous rating.
- Preserve Realistic Default Losses: Leave PVGIS system losses at 14% (or adjust between 11% and 14% only when using microinverters with zero shade mismatch and short runs). Zeroing this field produces physically impossible AC expectations.
- Upload Horizon Profile: Whenever trees, neighboring gables, or parapet walls rise above 12° elevation south, east, or west of the array, upload a horizon table.
3. Accounting for Temperature Coefficient and Clipping
Standard PVGIS defaults assume a generic crystalline silicon temperature coefficient around −0.35%/°C. If you compare an N-Type TOPCon (−0.29%/°C) or HJT / Back Contact (−0.24 to −0.26%/°C) module in hot desert or tropical climates (Rajasthan, Arizona, Spain, Outback Australia), lower thermal sensitivity preserves roughly 2% to 4% more summer midday yield.
Conversely, never ignore inverter DC/AC clipping. Pairing 6.8 kWp of south-facing modules onto a 5.0 kW continuous AC inverter (1.36 DC/AC ratio) boosts shoulder-hour harvesting while clipping midday clear-sky peaks above 5.0 kW AC.
4. What PVGIS and PVWatts Do Not Include
- Utility Export Curtailment / Zero-Export Limits: Grids enforcing export caps or low midday export tariffs penalize unconsumed midday energy unless stored in LiFePO4 batteries.
- Dust, Pollen & Monsoon Soiling: Dry agricultural or urban corridors lose 5–12% between cleanings; flat 5° commercial roofs lose self-cleaning rain runoff.
- Snow Accumulation: High-latitude roofs (>45° N) can lose 100% of output for days or weeks when snow rests on low-pitch arrays below 25°.
5. Bottom Line for Reviewing Quotes
Always divide quoted annual kWh by system DC kWp to inspect Specific Yield (kWh/kWp/yr). Verify that figure falls within 5% of an independent PVGIS or PVWatts run using 13–14% system losses and flush roof mounting.