Solar PV Performance Analysis: 3.2 Shading

At some stage during the year almost all solar projects will have shading, either from nearby objects (e.g. buildings, antenna, trees), or from far objects (e.g. mountains) or from interrow shadings (e.g. one module shades another nearby module).

In the northern hemisphere, the sun is lowest in the sky (winter solstice) around the 21st December and in the southern hemisphere around the 21st June. When the sun is lowest in the sky this will lead to the highest shading losses as there are typically more horizontal obstacles than vertical ones.   

When designing the site, the engineering team should of course plan to minimise shading losses by doing a full survey of the site, noting down any obstacles (chimneys, guard rails, walkways, buildings) and leaving a buffer area around them to minimise shading impact. Also string wiring should be considered to limit any shading losses and optimisers can be used to further mitigate losses. A PV Syst simulation including shading scene analysis should then be done as a reference for site performance.

So when we talk about underperformance from shading, we’re specifically referring to the losses above the budgeted financial model losses. E.g. in the below example if the financial model has 4% shading losses and we work out that actually there are 5% shading losses, then we have 1% underperformance from shading.

Working out the exact shading losses is not easy because they vary throughout the day and the year, but we can estimate shading losses by simply comparing the output of non-shaded inverters vs the output of shaded inverters. So the losses are the difference in output between the non-shaded “reference” inverter(s) and those that are shaded. Taking a reasonable sample of days over the year and performing the same calculation allows for a good general estimation of shading losses.

Some shading such as trees are quite easy to fix by pruning or removing the tree. This is the same for ground mounted sites, where the vegetation needs to be well managed to never shade the bottom row of modules.

Above we have shading caused by a crane. The crane should be moved when not in use to avoid more shading than necessary.

Here we have shading from a nearby building, this is more difficult to solve permanently but the effects can be minimised by rewiring the strings linearly to limit the losses to the string closest to the building.

In this case the modules could have been installed further up the roof slope to increase the buffer zone from the surrounding building.

Typically in the monitoring system we can see shading based on production curves that start late or finish early. In the below example we see two strings have shading both in the morning and in the afternoon. We can check the string layout to see where these strings are physically installed and whether or not we can remove the shading object.

To work out the % impact of shading on those strings, we can compare the output of non-shaded strings (nice bell curve) with the shaded strings (output dips in the morning and afternoon)

As a side note we also need to be aware of shading on the pyranometer. If the pyranometer is shaded then this will artificially inflate the PR and lead to wrong assumptions!

In summary :

During the design stage:

• A full site assessment needs to be done to ensure any obstacles are noted and that modules are placed where possible to avoid these obstacles (particularly for rooftop projects)

• The string layout should be designed to minimise shading of any objects noted during the site assessment. For example if there is unavoidable shading, putting linear string wiring can help to minimise that shading

During the operational phase:

• Any temporary sources of shading (e.g. trees, cranes) need to be minimised

• The budgeted shading in the financial model should be compared to actual shading losses as a feedback loop to improve on designing for future projects.