Detailed losses due to backtracking and backtracking management

[alfRCHD]

Hello,

I have a few questions about backtracking on slopes with PVsyst 7.2.3.

I know this is a very discussed topic in the PVsyst community, but I didn't manage to find all the answers to my questions.

I am currently working on different backtracking and topography scenarios to compare PVsyst simulations’ results. I have three 3D scenes: 2D layout, fixed slope, full topography and a scenario where I reduced the pitch by projecting the 3D scene with the terrain angle (*cos(alpha)). On each scenario, I have made a simulation with BT and another one without.

My first question deals with the detailed losses indicated by PVsyst:

Fixed slope Without Backtracking: GlobInc: +36.6%, NearShading: -5.52%, shading elec loss: -1.68%

Fixed slope With Backtracking: GlobInc: +31.5%, NearShading: -3.2%, shading elec loss: -1.02%

I understand that GlobInc is more important without BT for reasons of "non-optimal" sun angle, and that the near shading loss is more important without BT (mutual shading). However, it is the orders of magnitude that I do not understand. It seems that PVsyst considers that the gains in mutual shadings do not compensate for the losses in sun exposure in the case of the simulation with Backtracking.

More generally, the PR with backtracking is higher but the Egrid_total is higher without backtracking.

Do you have an explanation for these results? Would the solution be to increase the fraction for electrical effect?

(I get the same trend in simulations with full topography)

I have another question concerning the BACKTRACKING MANAGEMENT tool.

PVsyst tells us:

"You have to specify a pair of trackers acting as a refence for this calculation. For avoiding mutual shadings, you have to choose the pair with highest width/pitch value (GCR)."

I understood that the same angle optimisation via BT was going to be applied to all trackers based on the calculation of the pair of "leading trackers". Therefore, I tried to see the difference in "near shading factor tables" by taking two different pairs of leading trackers. The first is located on a steeper section of land than the second. This should result in different mutual shadings. It turns out that I get the same two near shading factor tables (the first pair is on a 16.5% slope and the other on an 11.5% slope).

Would you have an opinion on this point?

In other words, how do you consider the difference in altitude between the leading tracker and its neighbours in the calculation of the mutual shading factor?

Thank you very much for your answers and comments :D

----------------------------------

[Michele Oliosi]

Unless you are in the ideal case, a pure backtracking on a hill is impossible. Please see:

https://www.pvsyst.com/help/backtracking_onhill.htm

Therefore PVsyst does not have any slope awareness in terms of backtracking. The backtracking algorithm is a usual one, i.e. adapted to flat ground and regular rows of trackers. The only parameter that is considered by PVsyst is the horizontal pitch between trackers (and tracker sizes of course) so that the elevation difference will play no role there.

Therefore, with a constant slope or complex topography, you will have a sub-optimal orientation of your trackers, oftentimes leading to mutual shadings. This is realistic, unless you have a special slope-aware algorithm. (To our knowledge, constant slope situations do not really happen in reality ?)

Finally, note that PR is not an intuitive measure to compare backtracking with no backtracking, because the normalisation is on the POA irradiance, which is reduced in the case of backtracking (leading to a higher PR). Please see https://www.pvsyst.com/help/performance_ratio.htm