The shading loss for every panel, from the building’s own 3D model.

Shading is calculated against what is in the 3D model: the roof and walls of the house, other building parts and the obstacles you add from above in step 3. The obstacles are dormers, chimneys, vents, trees and poles or flagpoles, each with its own height and dimensions. A tree has both a height and a crown radius.

A neighbouring house that casts shade can be drawn as a building part of its own, and then shades just like the house’s own parts. Panels on a pitched roof lie in the plane of the roof and do not shade each other. Rows on racks on a flat roof do, so the row spacing shows up in the loss.

A large share of the year’s irradiation comes from the sky, not directly from the sun. So for each sample point the designer works out how much of the sky is visible, weighted by the angle at which the light falls on the panel. A dormer or wall next to the panel takes away both direct sun and a piece of sky, and both are deducted.

The loss per panel is the shaded share of the year’s irradiation – direct sun at the points that lie in shade and the diffuse light from the hidden part of the sky – divided by the panel’s total irradiation over the year.

The detailed simulation goes one step further. For each panel, the shading on the panel’s three bypass sections is calculated against a grid of sun positions, and for every hour of the weather year the shading is read from that grid. The terrain horizon from PVGIS blocks the direct sun when it is low.

The string power is then calculated with a simplified model of the panels’ current and voltage, including bypass diodes. A shaded section limits the current or is bypassed, and the string operates at the point that gives the most power. This shows the loss of having a shaded panel in the same string as unshaded ones. With optimisers, each panel is calculated on its own.