Simulate Terrain-Following Tracker Systems

[Valerie Chan]

Is there any guidance for modeling terrain-following tracker systems, like Nextracker XTR?

[Michele Oliosi]

Because these terrain-following algorithms are proprietary, at the moment PVsyst cannot model them. PVsyst will at the moment always apply the same tracking angles to the whole scene, to the exception of tracker axis tilt differences which will induce some partially adaptive behaviour.

For very irregular topographies, you may change manually the parameters used by PVsyst to backtrack. In the 3D scene > backtracking management, you can manually adjust the top and bottom frames of the tables (increase them), which makes the backtracking a bit more conservative.

Finally, some people use a method in which they create a new MET file using the tracker angles they have received from the tracker manufacturer. I believe people at Nextracker may help you with this, or some people on the forum may be able to use this procedure ?

[FernandoAlbendea]

3 hours ago, Michele Oliosi said:

Finally, some people use a method in which they create a new MET file using the tracker angles they have received from the tracker manufacturer.

Good afternoon, Michelle. If I am not wrong, this means that it would be possible to include real tracking angle data (i.e. for a whole year) plus the GHI and DHI, and simulate with that meteo file?. If that's the case it would be very interesting to know how it could be carried out. 

Thanks!

[Michele Oliosi]

Actually the trick used is a bit more complicated. It involves calculating the POA irradiance that trackers would yield and then simulating with that POA irradiance. This is because at the moment it is not possible to import tracking angle data through the MET file directly, one needs to trick their way out of the problem.

[FernandoAlbendea]

But, even if we estimate/measure the POA Irradiance (also known as GTI or GlobalInc) and imported it to PVSyst it would:

1. Transform into GHI and DHI using the Hay Model.

2. Transform again into GTI assuming perfect -and the same- tracking for all the trackers.

Therefore I see some limitations:

a) if the POA is recorded on site, PVSyst won´t use that exact values of GTI and Diffuse for the simulation.

b) If backtracking 3D is used on site, there is no way to simulate it, not even with POA data.

c) Simulations are forced to use Hay model, which computes a lower PR.

Please let me know if I am missing something on this regard, and if there is a way to solve this potential limitations. Thanks!

[Michele Oliosi]

Honestly, I think the strategy was different there were more operations involved. But as mentioned you should ask the manufacturer directly. Currently I am not able to describe the full procedure.