PVsyst is the industry-standard software for solar energy yield simulation, used by EPC engineers, independent engineers (IEs), and project lenders on the vast majority of bankable solar project assessments worldwide. Getting your PVsyst model right is not just an engineering exercise — it directly determines whether your project will be financed, at what debt-service coverage ratio, and whether actual generation will meet financial projections over the 25-to-30-year project life. This guide covers every critical input parameter, common modelling errors, and how to produce a simulation that will survive lender scrutiny.
PVsyst (developed by PVsyst SA in Geneva, Switzerland) has dominated the commercial solar simulation market for two primary reasons: its parameterised loss model maps directly to physical phenomena that can be measured and validated in the field, and its output format — the detailed loss diagram and P50/P90 yield report — has been adopted as the standard bankability deliverable by project finance lenders and independent engineers globally.
Other tools exist: SAM (NREL's System Advisor Model), Helioscope, Homer, and PVLib (Python library). SAM is commonly used for US government procurement studies and portfolio analysis. Helioscope is popular for commercial rooftop design due to its speed and 3D shading modelling. However, for projects requiring lender-grade independent engineering review, PVsyst remains the required tool in most jurisdictions.
PVsyst version 7.4+ introduced bifacial gain modelling using the infinite sheds model and improved IAM (Incidence Angle Modifier) calculations, making it suitable for modern bifacial n-type module projects.
The quality of a PVsyst simulation is entirely determined by the quality of its inputs. Experienced independent engineers will closely audit every input assumption when reviewing a developer's PVsyst report.
PVsyst requires hourly Global Horizontal Irradiance (GHI), Diffuse Horizontal Irradiance (DHI), and ambient temperature data for a typical meteorological year. Approved data sources for bankable projects include:
For projects larger than 10 MW, lenders typically require irradiance data from at least two independent satellite sources, with on-site ground station measurement for 12+ months as a validation input. The inter-source discrepancy in annual GHI is usually 1–4%; when sources diverge by more than 5%, the project requires further investigation.
PVsyst uses the one-diode model for module electrical characterisation. Key parameters derived from the module datasheet are:
Always use the PVsyst-provided PAN file from the module manufacturer's official database or a PAN file created from flash test data, not the generic default. Manufacturers including LONGi, Jinko, JA Solar, and others maintain regularly updated PAN files in the PVsyst online database.
Shading losses are among the most variable and most impactful loss factors in a PVsyst simulation. For ground-mount utility-scale projects, far-field horizon shading and near-shading from row-to-row must both be modelled.
PVsyst's loss diagram is the heart of the simulation report. It traces irradiance energy from horizontal plane through to AC energy delivered to the grid, quantifying each loss step. Typical values for a well-designed ground-mount system:
Total system losses of 18–25% are typical for well-designed commercial and utility-scale PV systems, yielding a Performance Ratio (PR) of 75–82% for fixed-tilt systems.
The P50 yield is the median expected annual generation — there is a 50% probability that actual generation exceeds this value in any given year. The P90 yield is more conservative — there is a 90% probability that actual generation meets or exceeds this value. Project finance lenders typically size debt service on the P90 (or P75) yield to ensure loan repayment even in below-average irradiance years.
The gap between P50 and P90 is driven by interannual irradiance variability (typically ±4–8% standard deviation at the 1-sigma level) and model uncertainty. PVsyst's Monte Carlo uncertainty module can quantify total P90 uncertainty from combined sources.
| Uncertainty Source | Typical Contribution (1-sigma) | Mitigation |
|---|---|---|
| Interannual irradiance variability | 4 – 8% | Use 20+ year satellite data; on-site measurement |
| Irradiance data source uncertainty | 2 – 5% | Average two independent satellite sources |
| Module power uncertainty | 1 – 3% | Use STC-certified flash test reports at delivery |
| Soiling model uncertainty | 1 – 3% | Site-specific soiling measurements over 12 months |
| Model / software uncertainty | 1 – 2% | IE cross-check with SAM or Helioscope |
| Combined (P50 to P90 gap) | 6 – 12% | — |
Independent engineers reviewing PVsyst reports for lenders commonly flag the following issues:
| Output / Deliverable | What Lenders Check | Typical Acceptable Range |
|---|---|---|
| P50 specific yield (kWh/kWp) | Plausibility vs. GHI and PR | 1,100 – 2,200 kWh/kWp/yr (site dependent) |
| Performance Ratio (PR) | Consistency with loss diagram | 75 – 82% (fixed tilt), 78 – 85% (tracker) |
| P50 to P90 gap | Adequacy of uncertainty analysis | 6 – 12% depending on data quality |
| Year 1 vs. Year 25 yield | Degradation methodology | 25-year P50 ≈ 88–92% of Year 1 P50 |
| Inverter clipping loss | DC:AC ratio reasonableness | < 2% for bankable projects (< 1.3 DC:AC) |
One of the most common problems in solar project development is the gap between the module and inverter modelled in PVsyst during the development phase and the equipment actually procured. Changes in module supplier — even within the same manufacturer's product line — can shift the PVsyst P50 yield by 1–3% due to differences in temperature coefficient, bifaciality factor, and low-irradiance efficiency.
Econo Solar maintains current datasheets and PVsyst PAN files for all modules it supplies, including LONGi Hi-MO 7 and Hi-MO 9, Jinko Tiger Neo 78HL4, JA Solar JAM72D42, and others. Our procurement team can lock in module supply at the development stage so your PVsyst model and your delivery specifications are aligned from NTP to COD. Request a technical data package to support your energy yield study.
P50 is the median expected annual energy yield — there is a 50% probability that actual generation in any given year exceeds this value. P90 is the conservative estimate — there is a 90% probability that actual generation meets or exceeds this level. Lenders typically require DSCR (Debt Service Coverage Ratio) to be above 1.20x at the P90 yield to ensure loan repayment under below-average irradiance conditions. The difference between P50 and P90 (commonly 6–12%) quantifies the combined uncertainty from irradiance variability, data source accuracy, and modelling assumptions.
For bankable projects, use two independent satellite-derived sources such as Solargis and Meteonorm 8, or Solargis and SolarAnywhere. Average their annual GHI values as the P50 central estimate; the inter-source spread informs the data uncertainty component of your P90 calculation. For projects above 10 MW, lenders often require 12 months of on-site pyranometer data (Class A, ISO 9060) to validate satellite data. Where satellite-to-ground station bias exceeds ±3%, apply a correction factor to the long-term satellite data series.
Use PVsyst version 7.4 or later with the "infinite sheds" bifacial model enabled. Key inputs are the module bifaciality factor (from the datasheet — typically 0.70–0.80 for bifacial PERC, 0.80–0.85 for bifacial TOPCon), the rear-side ground albedo (0.20 for bare soil, 0.25–0.30 for concrete, 0.65–0.80 for white gravel), and the module's installation height above ground (rear clearance affects rear-side irradiance uniformity). For tracker systems, also input the tracker rotation limit and ground clearance at the lowest point. Typical bifacial gain modelled in PVsyst ranges from 2–8% depending on albedo and system configuration; independent engineers will scrutinise albedo assumptions closely.
Econo Solar provides module PAN files, datasheet packages, and bankable equipment documentation to support your PVsyst energy yield study.
Get a Free Quote