Sandia Model

The Sandia model for the modelling of PV Modules has been developed at the Sandia National Laboratories since 1992.

It is fully described in the report SAND2004-3535 (2004)¹ directly available on the web.

For a given PV module, the Sandia model defines approximately 40 parameters established from continuous outdoor measurements over several days representing various weather conditions.

Sandia Laboratories provides a database of measured data comprising approximately 500 items, but corresponding to only about 100 actually measured modules, since most entries present actual measurements renormalized to the specified STC values. The Sandia model cannot be applied to other modules than those which have been fully measured using this procedure.

The Sandia model does not describe full I/V characteristics but only 5 points on the I/V curve: (0, Isc), (Vmp/2, Ix), (Vmp, Imp), (Vmm, Ixx), and (Voc, 0) (where Vmm = (Voc + Vmpp)/2).

Therefore, for use in simulation, PVsyst must complete the I/V curve. For a given operating condition (G, T), we first calculate the 5 points from the Sandia model equations, then establish the one-diode model passing through these 5 points.

The Sandia model defines a spectral correction, as a function of the Air mass only (no dependence of the Kt). Therefore we understand that this correction should be applied to the beam component only. The spectral balance (gain for crystalline, loss for amorphous) is reported on the loss diagram.

Comparisons with the one-diode model

When opening a module in the PVsyst database, if an equivalent module is available in the Sandia database, it is referenced in the "Remarks" (and in the choice lists), giving you the option to select which model to use.

• On the "Additional Parameter" page, the parameters of the model are visualized, along with eventual differences with the PVsyst model. Here you can also choose a Sandia module to be associated to the module.
• The button Sandia-Diode compare opens a window with a close comparison of the results of both models.

If the STC values match, you can request adjustment of the one-diode parameters to best match the Sandia model. You can view the discrepancies between the two models at 25°C or across all temperature conditions.

For crystalline modules, it is usually possible to achieve very good agreement between both models, keeping in mind that during simulation, the Sandia model applies a small spectral correction (gain) that is not accounted for here.

For amorphous modules there are very significant differences, which we don't understand in detail.