Transition from PVsyst to PlantPredict

This guide is for engineers moving a project—or a full modeling workflow—from PVsyst to PlantPredict. It covers how to bring your PVsyst files into PlantPredict, how the two tools’ structures and loss trees map onto each other, and the specific differences that affect results.

---

​

1. Importing from PVsyst

​

Project import

PlantPredict can import a complete PVsyst project from a ZIP file. The PVsyst project (PRJ) becomes a PlantPredict project, and each calculation variant (VC#) becomes a separate prediction. Modules (PAN), inverters (OND), and weather data (MET) embedded in the project are imported automatically—you do not need to upload them separately. To import: go to Import Project → Import PVsyst ZIP and select your ZIP file.

​

Individual file imports

If you are migrating components piece by piece, each PVsyst file type has a dedicated import path: For PAN file imports, PlantPredict re-solves the single-diode parameters after import to ensure internal consistency—see PAN file re-solve in Section 3 for what to expect.

​

Shade factor table export

To validate shading results against PVsyst, PlantPredict can export a Beam Irradiance Shade Factor Table (10° sun height × 20° azimuth grid) from the 3D Scene page via Export Shade Factor Table. This is available for fixed-tilt systems and trackers using Standard Backtracking or True Tracking, with Binning Type set to Detailed.

---

​

2. What corresponds to what

​

Plant topology

PVsyst’s “system” (array + inverter grouping) maps to a PlantPredict Inverter plus its associated DC Fields. PlantPredict adds Arrays and Blocks as aggregation layers above that. The DC Field is the lowest modeling unit and aggregates identical modules and strings behind a shared tracker or fixed-tilt array. See the Power Plant Builder pages for the full hierarchy.

​

Loss tree

The table below maps common PVsyst loss diagram lines to PlantPredict’s P50 Loss Tree. Loss percentages are not directly comparable between the two tools even when the underlying physics is identical—see Loss-factor denominators in Section 3. PlantPredict’s electrical shading response is configurable per simulation: None, Linear, Fractional Electrical Shading, or Step Fractional Electrical Shading (v12.17.0+). The Step Fractional model accounts for bypass-diode circuits with configurable horizontal partitions per bay, more closely representing the non-linear electrical response of partially shaded modules than the legacy Fractional model. See Electrical Shading Effect.

​

Simulation settings

The most common PVsyst-to-PlantPredict setting mappings on the Simulation Settings page:

---

​

3. Differences to be aware of

​

Azimuth convention

PlantPredict measures all azimuths clockwise from geographic North on the 0°–360° range: 0° = North, 90° = East, 180° = South, 270° = West. PVsyst uses 0° = South with a signed ±180° range. The importer converts azimuth values automatically, but always re-verify module azimuth and tracker axis azimuth after import, and use the PlantPredict convention for any values entered manually.

​

Loss-factor denominators

PVsyst reports each loss line as a percentage of the value before that loss is applied, so losses compound multiplicatively and a given mechanism appears larger when earlier losses have already reduced the signal. PlantPredict reports each loss as a percentage of a single fixed reference quantity (e.g., Global POAI, DC Power at STC, or Total AC Power from Inverters), so losses are additive against a constant denominator. Total plant output is consistent between the two conventions, but line-item percentages are not directly comparable.

​

Circumsolar default

PVsyst (v7+) handles circumsolar as a distinct component within the Perez model (CircInc), applied directionally rather than as isotropic diffuse. PlantPredict offers a binary choice: Direct or Diffuse (default: Diffuse). Because the two tools decompose circumsolar differently, they will disagree on the Transposition vs. IAM line items even when total POA is identical. When benchmarking, set PlantPredict’s Circumsolar Treatment to Direct on Simulation Settings to most closely match PVsyst’s behavior.

​

Horizon shading scope

PlantPredict’s far (horizon) shading applies to beam only: when the sun is below the interpolated horizon profile, beam irradiance is set to zero. Sky-diffuse irradiance and ground-reflected irradiance are not reduced by the horizon. PVsyst applies far shading to beam, diffuse, and albedo components. At sites with a significant horizon elevation, PlantPredict will predict more annual diffuse and reflected energy than PVsyst, and the gap grows with horizon elevation. See Horizon Shading.

​

Solar position algorithm

PlantPredict uses the NREL Solar Position Algorithm (Reda & Andreas, 2004) with ±0.0003° accuracy. PVsyst uses the US Navy algorithm. The difference in sun position is negligible for energy totals but contributes to small systematic differences in time-series comparisons.

​

String sizing temperature

PlantPredict sizes strings using an ASHRAE temperature bin by default. PVsyst uses −10 °C. This produces different string counts for the same module and inverter when sizing helpers are used; verify string lengths after import.

​

Inverter power factor for sizing

PlantPredict assumes a 0.95 power factor when sizing inverters. PVsyst assumes 1.0. This affects the apparent DC/AC ratio when comparing plant designs sized in each tool.

​

PAN file re-solve

After importing a PAN file, PlantPredict re-solves the single-diode parameters ($I_{ph,ref}$, $I_{0,ref}$, $\gamma_{ref}$, $R_{s,ref}$, $R_{sh,ref}$) using Levenberg–Marquardt to ensure all parameters are internally consistent. The resulting values will differ slightly from PVsyst’s stored values. This is expected—not an import error. See Module Parameter Generation for the full field-by-field mapping. Additionally, PlantPredict adds per-module DC wiring resistance to $R_{s,ref}$ at run time; the module-file value is the pre-wiring value only. If the source PAN file already baked wiring into $R_s$, confirm that wiring is not counted twice.

​

Temperature coefficient units

PVsyst PAN files store $\beta_{V_{oc}}$ in mV/°C and $\alpha_{I_{sc}}$ in mA/°C. PlantPredict stores both in %/°C. The PAN importer handles this conversion automatically. If you are entering coefficients manually, convert to %/°C before entry.

​

Subhourly clipping

PlantPredict runs predictions natively at whatever timestep is in the weather file and applies no sub-hourly correction factor. In PVsyst v7.3 and earlier, sub-hourly clipping losses are estimated via an approximation algorithm applied to hourly input data. PVsyst v8.0 introduced a correction derived from actual sub-hourly irradiance data, and v8.1.0 moved to full native sub-hourly simulation with time-step-adapted Perez coefficients. If you are comparing against PVsyst v7.x with an hourly weather file and a high DC/AC ratio, expect clipping to be handled differently between the two tools.

​

Module temperature model

PlantPredict models cell temperature using a steady-state heat balance (default), with NOCT, Sandia, and DC Field Defined as alternatives. PVsyst v8.1.0 introduced a transient thermal model that accounts for module thermal mass, replacing its prior steady-state approach. When comparing against PVsyst v8.1+, expect systematic differences in cell temperature during periods of rapidly changing irradiance (cloud transients, morning ramp-up, evening ramp-down), which will affect DC output at those timesteps.

​

Perez coefficient set

PlantPredict ships nine selectable Perez coefficient sets; the default is the PlantPredict set. PVsyst typically uses its “Perez, modified” set. When benchmarking, confirm the Perez Coefficients dropdown on Simulation Settings matches the dataset used in PVsyst. Mismatched coefficient sets produce transposition differences that have nothing to do with the underlying physics.

​

P90 / uncertainty methodology

Both tools compute P-values by combining named uncertainty components via root-sum-of-squares into a single sigma, then applying a normal-distribution z-score to P50. The component taxonomies differ, so uncertainty inputs do not transfer directly between the two tools. PlantPredict’s Uncertainty Analysis uses five components: interannual variability, irradiance measurement accuracy, monitoring-period representativeness, spatial variability, and modeling accuracy. PVsyst uses a different set: weather data annual variability, PV module model and parameters, inverter efficiency, soiling and module quality loss, and long-term degradation. These do not map one-to-one and must be re-entered independently in each tool.

---

​

References

• Reda, I., & Andreas, A. (2004). Solar Position Algorithm for Solar Radiation Applications. NREL/TP-560-34302. DOI: 10.2172/15003974