Prospect PV simulator optical simulation

In this document

This document describes the optical simulation stage of the Prospect PV simulator in Solargis Prospect. It explains how solar radiation is translated into effective Global tilted irradiance (GTI) at the PV cell level, accounting for all relevant optical losses.

Overview

The optical simulation is the first computational stage of the Prospect PV simulation chain, following the collection of simulation inputs. It takes the solar radiation and meteorological inputs together with the PV system geometry and uses them to calculate the effective Global tilted irradiance (GTI) on the surface of each PV module — the foundational quantity for all subsequent electrical calculations.

This stage determines how much solar radiation actually reaches each PV cell after accounting for the geometry of the sky, horizon shading, inter-row shading within the PV field, soiling, angular reflectivity, and spectral mismatch. The Prospect simulator applies the same core optical models as the Argus PV simulator — including the Martin and Ruiz angular loss model and the Lee & Panchula spectral correction model — but uses a view-factor approach for near shading rather than Monte Carlo raytracing. This makes the Prospect optical simulation orders of magnitude faster, which is essential for the portfolio-scale and screening use cases the simulator is designed to serve.

Because GTI must be evaluated at multiple points across each unique table in the field, the Perez model is split into two parts:

Components that depend only on sun position and radiation inputs are computed once per time step.
Components that depend on the specific point geometry on the table are computed per sample point.

This avoids redundant recalculation and keeps the optical simulation computationally efficient even at fine spatial sampling.

Processes included in this stage

The following processes are applied sequentially during the Prospect simulator optical simulation:

Sky irradiance model (Perez model, solar position pre-calculated as input)
Far horizon shading
Near shading — view-factor model with direct and diffuse components
Direct (hard) shadows
Diffuse shading
Angular reflectivity losses (pv-angular model)
Soiling losses
Spectral correction (Lee & Panchula model)

Optical simulation

Sky irradiance model

Global horizontal irradiance (GHI) and Direct normal irradiance (DNI) are decomposed into their basic irradiance components — Direct horizontal irradiance (DHI) and Diffuse horizontal irradiance (DiffH). These derived components serve as inputs to all subsequent optical calculations.

The Perez model then calculates the standard sky model coefficients F1 (circumsolar brightness) and F2 (horizon belt brightness) as a function of the basic irradiance components and sun elevation. From these, the theoretical GTI (without losses) and the pv-angular coefficients — beam, diffuse, and reflected — are pre-calculated once per time step. The difference between GTI and GTI_eff (the effective GTI accounting for angular reflectivity) is that each GTI component is multiplied by its corresponding pv-angular coefficient.

Sun position (azimuth and refracted elevation) is provided as a pre-calculated input array rather than computed within the optical simulation.

For full details on the Perez sky model, angular losses, and spectral correction, see Incident irradiance.

Far horizon shading

The far horizon profile — provided as a series of elevation angles evenly spaced clockwise from north — is applied as a pre-computed shading mask. For each time step, direct radiation is set to zero for any sun position that falls below the horizon elevation at that azimuth. This removes irradiance blocked by distant terrain features before near-field calculations begin.

Near shading — view-factor model

Near shading quantifies the reduction in GTI caused by adjacent PV tables within the field. The Prospect simulator uses a view-factor approach, processing direct and diffuse shading separately.

Direct (hard) shadows

Direct shadows are calculated as hard shadows only — the sun is treated as a point source with perfectly parallel rays. Shadows from tables within the 3 × 3 neighborhood are also calculated. Testing whether a point on the table is in shadow reduces to checking whether it falls within any of the projected rectangles.

Diffuse shading — hemisphere view-factor model

Diffuse shading accounts for the portion of the sky dome and surrounding surfaces that is obstructed by nearby tables, even when those tables do not cast a direct shadow. Each unique table is sampled on a configurable grid of points (default 10 × 10). For each sample point, a hemisphere of 48 × 60 = 2,880 segments is evaluated to determine what fraction of the hemisphere is covered by sky, ground, nearby panels, far terrain, and the horizon belt. These fractions are the view-factor coefficients used in the GTI calculation.

Note: For fixed-tilt systems, diffuse shading conditions are time-independent and the hemisphere calculation is performed only once. For tracker systems, hemisphere calculations must be repeated for all time steps and all unique tables, as the table orientation changes continuously. To manage this computational cost, the hemisphere grid is set to 5 × 5 points for tracker systems, giving a total of 25 hemispheres per unique table per time step.

Angular reflectivity losses

The angular losses model accounts for optical losses caused by the angle of incidence (AOI) on the module surface. Once the pv-angular coefficients (beam, diffuse, and reflected) are pre-calculated for a given time step, GTI_eff is obtained by multiplying each GTI component by its corresponding angular coefficient. These losses increase at low sun angles and affect the total radiation available for conversion at the PV cell.

For full details, see Incident irradiance — angular losses.

Soiling losses

A soiling correction factor is applied as a multiplicative loss to GTI after horizon shading, before angular reflectivity losses are calculated, attenuating the effective irradiance reaching the module surface due to dust and particulate deposition. The factor is configured as a monthly or yearly average value in the simulation parameters.

For full details, see Soiling loss model.

Spectral correction

The First Solar spectral correction model (Lee & Panchula) adjusts GTI_eff for the spectral mismatch between incident sunlight and the module's spectral response. The correction factor is a function of optical air mass — calculated from site elevation and refracted sun elevation — and precipitable water content. Separate correction coefficients are applied for cSi, CdTe, CIS, and aSi module types. The spectral correction factor is applied by default and can be disabled via the disableSpectralCorrection simulation parameter.

For full details, see Incident irradiance — spectral correction.

The optical simulation output — the spectrally corrected per-cell GTI (GTI_eff) — feeds directly into the electrical simulation stage, where it is converted into DC power output.

Documentation Index