Join our upcoming webinar "Can you rely on measured data?" on October 27, 2026 at 10:00 AM CET. Register here
We compared six tools using the same six test sites — chosen for climate diversity, from arid desert to humid tropics — and the same four PV system configurations at every site. Wherever possible, we also matched input data and model selections, so the comparison isolates differences in methodology rather than configuration.
| PV simulator | Type |
|---|---|
| Solargis Evaluate | Commercial, cloud-based |
| Solargis Prospect | Commercial, cloud-based |
| PVsyst | Commercial, desktop |
| System Advisor Model (SAM) | Free, desktop (NREL) |
| pvlib | Open-source Python library |
| SolarFarmer | Commercial, desktop and cloud (DNV) |
Quantitative results cover pvlib, PVsyst, and SAM against Solargis Evaluate. Solargis Prospect and SolarFarmer appear only in the methodology-based comparisons.
The comparison covers the full simulation chain: solar and weather inputs plus scene setup, optical simulation (irradiance, shading, module losses), and electrical simulation (DC and AC conversion through to the grid connection).
At the grid connection point — the number that determines a project's reported yield — median bias against Solargis Evaluate stays within about 1% for all three simulators:
Note: Solargis Evaluate is used as the reference point because the comparison needs one. This doesn't mean Evaluate is the more accurate simulator of the group — it means the study measures how the others diverge from it, not which one is closer to the truth.
That final agreement is closer than it looks. Each simulator gets there through two large deviations of opposite sign that mostly cancel out: an upward drift through the optical chain, reversed by a downward step at DC conversion. A tool that agrees with Solargis Evaluate to within 1% at the grid connection point can still differ by 2–3% earlier in the chain — and agreement also narrows sharply on projects with tracking, complex terrain, or difficult horizons.
Beyond the numbers, the tools differ structurally in what they model and how:
| Capability | Solargis Evaluate | pvlib | PVsyst | SAM | SolarFarmer |
|---|---|---|---|---|---|
| Built-in solar and weather database | Yes | No | No | No | No |
| Native resolution | 15-min, 1-min | Hourly or sub-hourly | Hourly or sub-hourly | Hourly or sub-hourly | Hourly or sub-hourly |
| Shading computation | 3D ray tracing | View factor | View factor | View factor | View factor |
| Built-in soiling model | Yes | Yes | No | No | No |
| Built-in snow model | Yes | Yes | No | Yes | No |
| Integrated reporting and analysis | Yes | No | Yes | Yes | Yes |
