Number of sheds (rows) and accuracy

The basic principle of the bifacial 2D models in PVsyst is to analyse the irradiance reflected from the ground between the rows of collectors or trackers, as well as the irradiance coming from the sky on the rear face, reaching between the rows.

In this model, the first and the last rows are not accounted for in the same way. When there are many rows, this difference may be neglected. But with very few rows, one should be aware that this may lead to inaccuracies.

Effect on the front side

The usual calculation of mutual shadings for "unlimited sheds" orientations is based on a simple geometric calculation in 2D. This simple calculation is not valid for the first row, because it is not shaded by the other rows. Therefore, this generic shading factor is multiplied by (Nrows-1)/Nrows.

The reflexion contribution on the front side is calculated from the inter-row ground points. For the first row this contribution is underevaluated, because the ground points are fully irradiated. This is indeed an approximaton. By the way, this contribution is never calculated in usual non-bifacial systems, as we do not avail of the "height above ground" and thus of the tools (ground points) for this evaluation.

Effect on the rear side of the last row

The reflected irradiance is coming from the ground points behind the last row, which are treated in the same way as the interrow ground points. There may be an additional contribution of the terrain beyond these ground points, which is not taken into account. In usual cases, this contribution should not be very important, as it is coming with a low view factor and high incidence angle.

Moreover the "shading" of the would-be next row for the back side sky irradiance is not present.

This is the reason why we discourage using this 2D model for experimental bifacial systems made of 2-3 rows only. This contribution becomes negligible with high number of rows (it scales as 1/Nrows).

Vertical plane systems

With such systems, it is indeed common to have one single row (for example a sound barrier along a road). We have checked that with the hypothesis of the bifacial model, the backside irradiance is quite similar to the normal front side calculation.