Two frame axis - backtracking in both directions

[Falberto76]

Hi,

We have to simulate an old version of horizontal tracker, which has two horizontal axis, one axis oriented in N-S direction and whose rotation determinates the tracking along the E-W direction (+/-60°) , and another axis which optimize the tilt  (range -15/+45) alond the daily sun movement. the two axis support frame of PV modules.

We have been told by the backtracking algorithm company that the tracker performs backtracking on both the axis and we're trying to get from them the tracking astronomical map of the trackers. 

We have difficulties to find a reliable way to simulate the energy yield on PVSYST since the SW allows to consider backtracking only at 1 of the two axis. We have operating data too of existing plants with the same tracker and performed some simulation with PVSYST with backtracking only at 1 of the two axis in order to verify if the simulation roughly approximates the field behaviour.

The complexity is given also by the fact that we should also simulate the same trackers with different PV modules, due to a revamping strategy which interests such trackers. 

Here the questions.

1) Assuming we're able to get hourly values of tracker position in both axis, is it possible to calculate  GTI on such custom orientations and get back an excel file?

2)Assuming the above (or either finding another reliable  tool which implements the Perez transposition  on the tracking surface), an idea could be to obtain the annual GTI and then roughly calculationg the yield by using the Performance ratio calculated with PVSYST on the same system, but with only one tracker allowed to to do backtracking. This may imply calculation errors but anyway will approximate the energy behavior and - correct me if 'm wrong - the  other losses from GTI surface collected energy to energy output at the meter should not be too different. Is this plausible or there are any ways to do the calculation?

Another way could be to calculate the system as a single axis (ie considering the primary axis only, i.e the one which has the greater rotation along the day), let's say the N-S axis which turn the E-W rotation, and assume a constant tilt, or a limited tilt to be sure the tracker does not go in mutual shading, and calculate with this configuration the yield. Then the yield could be adjusted by calculating the GTI on the tracking plane which follows backtracking in both directions.

Thank you for any hints or suggestion

 

[Michele Oliosi]

17 hours ago, Falberto76 said:

1) Assuming we're able to get hourly values of tracker position in both axis, is it possible to calculate  GTI on such custom orientations and get back an excel file?

17 hours ago, Falberto76 said:

Another way could be to calculate the system as a single axis (ie considering the primary axis only, i.e the one which has the greater rotation along the day)

At the moment it is technically possible to simulate with custom orientations, but cumbersome and very time consuming. We will improve this in the future, but it will take quite some time to implement a custom tracking angles feature. With the current version, you should use the batch mode, and a "Fixed Tilted Plane" orientation. Please define the 3D scene with just one array. With the batch mode, you should simulate different orientations, each time creating a hourly output file. Finally based on the tracker orientations given by the manufacturer you can select for each hour in which file to look. Sorry, as said this may be complicated.

You can indeed maybe apply the same strategy with a single axis tracker and run multiple variants with different eg. tilt. It's worth a try.

17 hours ago, Falberto76 said:

but with only one tracker allowed to to do backtracking.

I am not sure why and how you would have only one tracker to backtrack. Could you give more details ?

17 hours ago, Falberto76 said:

roughly calculationg the yield by using the Performance ratio calculated with PVSYST on the same system

I believe "rough" is indeed the right term. The uncertainty may be larger than the gain of using backtracking over tracking. Not sure you could trust this result. Sorry, since this case is not very common I don't have that much experience. In this case the best would be to try different options and compare.

[Falberto76]

Hi Michele, thank you for your feedback.

Just a small clarification for the posted point: "but with only one tracker allowed to to do backtracking." -> actually I was meaning only "one tracker axis to backtrack".

Sorry for the language, it was written fastly.

As a matter if fact we need to assess such type of trackers, since there are hundreds of MWp installed in Italy. 

 

Based on your feedback we understand in the batch simulation we will implement as parameter the axis orientation given by the controller algorithm and (maybe) is it possible to calculate the batch only for the specific time or an entire year shall be performed? Then we take as output the GTI and rebuild the annual GTI from the batches calculation. Down in the sankey diagram we could apply the other losses taken from the same b.o.p system configuration. Of course this will have an higher uncertainty than normal calculation.

We will revert to the post after getting some progress in this task. We're also actually thinking to start from basic equations but may be it is very time consuming.

 

thank you!

[Michele Oliosi]

@Falberto76 you can launch the simulation for shorter periods, but the same period for all the batch runs if I recall correctly. You select the option when choosing the simulation dates in the "Advanced simulation" window. Do tell me if that doesn't work.