In this document
This document describes the Geras PV simulator used in Solargis Monitor and Solargis Forecast. It explains how the simulator calculates global tilted irradiance (GTI), PV module temperature (TMOD), and PV power output (PVOUT), summarizes each stage of the simulation chain, and presents the verification of the GTI calculation.
Usage in Solargis platform |
|---|
This simulator is used in Solargis Monitor and Solargis Forecast data services, delivered through the Monitor & Forecast API and Push Delivery. |
Overview
The Geras PV simulator converts solar radiation and air temperature into the AC power output of a PV system for every time step of a data series. It is a fast, empirical simulator designed for repeated calculations on large numbers of sites, with each PV system described by a small set of parameters.
Geras relies on established models: the Perez anisotropic sky model, the Martin and Ruiz angular loss model, the Lee and Panchula (First Solar) spectral correction model, and the Huld power-rating model for PV modules. The same simulator processes historical, operational, and forecast data. Differences in PVOUT between periods therefore reflect changes in the input parameters (solar and meteorological data, PV configuration), not a change of PV simulation model.
Unlike the S2 PV simulator used in Solargis Evaluate, Geras does not model the 3D layout of the power plant or calculate current-voltage (I-V) curves. It uses view factors and empirical loss factors, trading detail for the speed that monitoring and forecasting services require. For an overview of all Solargis simulators, see PV energy yield simulation.
Aspect | How the Geras simulator handles it |
|---|---|
PV system per simulation | One mounting geometry, one PV module type, and one inverter type per request. All inverters are considered identical. |
Shading | Far horizon shading from the terrain model or a user-defined horizon. Inter-row shading of direct solar radiation. Near shading by objects is not modeled. |
Terrain | Terrain slope is used only for inter-row shading of fixed-tilt systems. 3D terrain is not modeled. |
PV modules | Generic technology classes: crystalline silicon (CSI), amorphous silicon (ASI), cadmium telluride (CdTe), and copper indium selenide (CIS). Bifacial modules are not modeled. |
Soiling and snow | Fixed annual or monthly loss factors, not dynamic models. |
Geras simulation stages
The simulation chain is organized into three sequential stages, each building on the results of the previous one.
Stage 1: Simulation inputs
Global horizontal irradiance (GHI) and direct normal irradiance (DNI) come from the Solargis satellite-based model at a spatial resolution of about 90 m. Air temperature at 2 m (TEMP) comes from reanalysis and numerical weather prediction (NWP) models, post-processed by Solargis to a spatial resolution of about 1 km. Forecasts use the Solargis nowcasting model and a blend of NWP models. For details, see Origin of the data for Monitor and Forecast.
Position: Latitude and longitude define the sun position for each time step.
Horizon: With terrain shading enabled, the far horizon is calculated from the Shuttle Radar Topography Mission (SRTM) digital terrain model (about 90 m). A user-defined horizon overrides it and can represent nearby obstacles.
Terrain slope: Slope and slope azimuth of the ground, used only for inter-row shading of fixed-tilt systems.
Ground albedo: By default, monthly values from the Solargis satellite-based albedo data.
PV system configuration: Mounting geometry, installed power, module technology, inverter, losses, and layout. Parameters not provided by the user are replaced by defaults.
PV system parameters and default values
Parameter | Default |
|---|---|
Mounting geometry | Horizontal plane |
Installed power [kWp] | No default; required for PVOUT |
Installation type | Free-standing |
Module technology | No default; required for PVOUT |
Temperature coefficient of power [% / °C] | CSI -0.438, ASI -0.18, CdTe -0.297, CIS -0.36 |
NOCT [°C] | CSI 46, ASI 44, CdTe 45 (API) or 48 (Push Delivery), CIS 47 |
Start-up date | No degradation |
Degradation, first year and annual [% / year] | 0.8 and 0.5 respectively |
Relative row or column spacing | None (no self-shading) |
Self-shading | Off |
Shading influence type | UNPROPORTIONAL_1 for CSI, PROPORTIONAL for other technologies (Push Delivery) |
Inverter efficiency [%] | 97.5 |
AC power limit [kW] | No limit |
DC losses: mismatch, cables, soiling and snow [%] | 1.0, 2.0, 2.5 |
AC losses: cables, transformer [%] | 0.5, 1.0 |
Terrain shading | On in the API; off in Push Delivery |
Note: Some settings are available in one interface only. The API supports a power-dependent inverter efficiency curve and module surface reflectance. Push Delivery supports a ground albedo override, and 1-minute and long-term summarization.
Tip: Terrain shading is enabled by default in the API, but not in Push Delivery. Keep it enabled for realistic PV simulation. A user-defined horizon is applied only when terrain shading is enabled.
Stage 2: Optical simulation
The optical simulation calculates the solar radiation on the module plane and the effective irradiance that enters the PV cells.
Module orientation: The simulator determines the orientation of the module plane: fixed tilt, one-axis (horizontal, inclined, or vertical), or two-axis tracking. Trackers follow the sun within their rotation limits. With backtracking, they rotate away from the ideal angle near sunrise and sunset to avoid shading between rows.
Sky irradiance model: Solar radiation on the module plane is the sum of the beam, diffuse, and ground-reflected components. The diffuse component follows the Perez anisotropic model, which describes diffuse sky radiation as an isotropic sky background with a brighter circumsolar region around the sun and a brighter band along the horizon. The sky background and ground-reflected components are weighted by view factors, which express the share of the sky and of the ground that the module plane sees.
Horizon shading: When the sun is below the horizon line, the horizon blocks the beam and circumsolar components and reduces the visible sky for the isotropic sky background. The GTI output includes horizon shading. Angular losses are not part of GTI; they are applied in the PVOUT calculation.
Angular losses: The module front glass reflects part of the solar radiation, more at high angles of incidence. The Martin and Ruiz model calculates these losses with typical surface properties for each module technology. The result is the effective irradiance Geff, the solar radiation that enters the PV cells.
Spectral correction: The Lee and Panchula model corrects Geff for the spectral mismatch between sunlight and the module spectral response, using air mass and precipitable water.
Stage 3: Electrical simulation
The electrical simulation converts the effective irradiance into AC power at the grid connection point, accounting for all electrical losses along the way.
Module temperature: Module temperature is calculated with the steady-state Ross model, with the heating coefficient derived from the nominal operating cell temperature (NOCT):
Where:
Tmod: PV module temperature [°C];
Tair: air temperature at 2 m [°C]
k: heating coefficient [°C·m²/W]
NOCT: module temperature at 800 W/m², 20 °C air temperature, and 1 m/s wind speed [°C]
Δk: mounting correction [°C·m²/W]
The model neglects wind and thermal inertia, which is adequate for 15-minute and hourly data.
DC power of PV modules: DC power is calculated with the power-rating model by Huld et al., which describes PV module efficiency as a function of solar radiation and module temperature. Its coefficients are fitted to measurements of commercial modules, separately for each technology.
Where:
PDC: DC power of the PV array [kW];
Pinst: installed power at Standard Test Conditions (STC) [kWp]
G' = Geff / 1,000 W/m²: normalized effective irradiance
T' = Tmod - 25 °C: module temperature difference from STC
ηrel: relative efficiency from the Huld model, a function of G’ and T’, equal to 1 at STC;
D: cumulative degradation at the given time
Inter-row shading: With self-shading enabled, the share of the module area in direct shadow of the neighboring row is calculated from the sun position, module orientation, relative spacing, and, for fixed-tilt systems, terrain slope. It applies to fixed-tilt systems and to horizontal and inclined one-axis trackers without backtracking. Only direct solar radiation is shaded.
Shaded cells limit the current of a whole string, so partial shading reduces power by more than the shaded area. The Geras simulator scales the shadow effect by a shading influence from 0 % to 100 %, which depends on the wiring of cells and modules.
Shading influence types
Type | Shading influence |
|---|---|
PROPORTIONAL | 20 % |
UNPROPORTIONAL_1 | 40 % |
UNPROPORTIONAL_2 | 60 % |
UNPROPORTIONAL_3 | 80 % |
In the API, the type names do not contain the underscore, for example UNPROPORTIONAL1.
Degradation: If a start-up date is provided, module power decreases by the first-year degradation rate and by the annual rate in the following years. Degradation D is applied to every data point at the native time resolution, according to the age of the system at that time.
DC losses: Mismatch, cable, and soiling and snow losses are applied as factors:
Where Lmm are mismatch losses, Ldc are resistive losses in DC cables, connectors, and fuses, and Lss are soiling and snow losses, given as one annual value or 12 monthly values. All losses are fractions.
Inverter conversion and AC power limitation:
Where ηinv is the inverter efficiency, either constant (Euro or California Energy Commission (CEC) weighted efficiency) or a power-dependent efficiency curve, and PAC,max is the AC power limit of the inverters or of the grid connection. The limit reproduces inverter clipping.
AC losses:
Where Lac are losses in AC cables to the grid connection point and Ltr are transformer losses.
Note: PVOUT represents a fully available system without outages or curtailment. You can apply unavailability afterwards, for example by setting PVOUT to zero in unavailable periods.
Geras simulator outputs
Geras simulator output | Description |
|---|---|
GTI | Global tilted irradiance on the module plane, with horizon shading and without angular losses [W/m², Wh/m², kWh/m²] |
TMOD | PV module temperature [°C] |
PVOUT | PV power at the grid connection point: power [kW] for sub-hourly time steps, energy [kWh] for hourly and longer time steps |
Module angles | Angle of incidence, tilt, and azimuth of the module plane for each time step |
Tip: To simulate a power plant with several orientations or inverter types, send one request per part and sum PVOUT. GTI is specific to each part and is not summed.
Verification of the GTI calculation
The GTI calculation of Geras was cross-checked against the GTI simulator of Solargis Evaluate, a separate implementation of the same approach. The tests covered about 25 sites and all mounting types except the vertical one-axis tracker with backtracking:
The vast majority of GTI values agree within ±2 W/m², most of them within ±0.5 W/m².
For fixed-tilt systems, the two simulators match.
For trackers, rare differences occur at sunrise and sunset, because Geras interpolates the angle of incidence and horizon belt between sky segments.
The accuracy of GTI and PVOUT is therefore driven mainly by the solar input data, documented in Validation of solar irradiance data, and by how accurately the request describes the PV system.
Further reading
Publications:
"Modeling daylight availability and irradiance components from direct and global irradiance": R. Perez, P. Ineichen, R. Seals, J. Michalsky, and R. Stewart
"Calculation of the PV modules angular losses under field conditions by means of an analytical model": N. Martin and J. M. Ruiz
"Spectral correction for photovoltaic module performance based on air mass and precipitable water": M. Lee and A. Panchula
"Design techniques for flat-plate photovoltaic arrays": R. G. Ross Jr.
"A power-rating model for crystalline silicon PV modules": T. Huld, G. Friesen, A. Skoczek, R. P. Kenny, T. Sample, M. Field, and E. D. Dunlop
Solargis Knowledge Base:
"PV energy yield simulation": Solargis
"S2 PV simulation chain": Solargis
"GTI simulator overview": Solargis
"Incident irradiance": Solargis
"Origin of the data for Monitor and Forecast": Solargis
"Satellite-based albedo data": Solargis
"Validation of solar irradiance data": Solargis