Modeling modules with different powers in a single MPPT

[Gustavo Pianovski]

I'm working on a project with 3 different modules (705, 710 and 715 Wp) connected to a single MPPT inverter:

 

Ideal Condition (No Losses)
	Module	705 Wp	710 Wp	715 Wp
	Config.	A	B	C	Total
Module	Imp (A)	17.55	17.59	17.63
	Vmp (V)	40.2	40.4	40.6
	P (Wp)	705	710	715
Electrical Configuration	String/Inv	8	36	37	81
	Mod/STR	30	30	30
Arrangement	Vmp String (V)	1205.1	1210.9	1216.7
	Imp String (A)	140.4	633.2	652.3
	P config. (kWp)	169.2	766.8	793.7	1729.7

 

An alternative modeling approach is to change the .OND to 3 MPPTs and use the MULTI-MPPT tool. This will allow me to model all three module types in the same inverter.

Will making this change to .OND cause other problems?

 

I did the simulations below to test:

 

• Simulation 1: 3 inverters with original .OND (1 MPPT per inverter). Each inverter is configured with a different module, totaling 5,176 kWp.
  • E_grid = 11261 MWh
     
• Simulation 2: 3 inverters with modified .OND (3 MPPTs per inverter), only to allow modeling in PVsyst. Each inverter is configured with 27 strings of the three types of modules, totaling 5,176 kWp. The PVsyst MULTI-MPPT tool is used to group the 3 MPPTs on the same inverter.
  • E_grid = 11261 MWh

The energy result (E_Grid) is the same in both cases. I understand that PVsyst doesn't consider any differences between the models. Therefore, we must calculate a loss due to this electrical configuration.

 

I understand that this electrical configuration will result in a greater mismatch. One way I found to calculate the electrical losses of this configuration is to consider that "the array voltage is always limited by the lowest voltage string." In this case, the voltage of the 710Wp and 715Wp strings will be equal to the voltage of the 705Wp string. I calculated the power of each configuration (P config. (Wp) = Vmp String (V) X Imp config (A)):

 

Edited July 9 by Gustavo Pianovski

[Gustavo Pianovski]

Continuation...

 

Actual Condition (COM Losses: string voltage limited by the smallest)
	Module	705 Wp	710 Wp	715 Wp
Module	Config.	A	B	C	Total
	Imp (A)	17.55	17.59	17.63
	Vmp (V)	40.2	40.2	40.2
Electrical Configuration	P (Wp)	705	706.6	708.2
	String/Inv	8	36	37	81
	Mod/STR	30	30	30
Arrangement	Vmp String (V)	1205.1	1205.1	1205.1
	Imp config (A)	140.4	633.2	652.3
	P config. (kWp)	169.2	763.1	786.1	1718.5

 

In the first table (ideal condition), the total power is 1729.7kWp. In the second table, the total power is 1718.5kWp. The difference between these two power outputs is 0.65%. I'm considering applying this loss to the mismatch simulation. Is this analysis correct?

 

I tested a second methodology: using the Detailed Study tool in the Mismatch Losses tab in PVsyst. In this analysis, the Voltage Difference between the 715Wp and 705Wp strings is 11.5V (1216.7V – 1205.1V (Table 1). Applying a voltage difference to the string of 11.5V, PVsyst returns a mismatch of only 0.05% (Figure below). Why is the loss calculated by PVsyst so low?

 

What is the best way to model this project and a loss to be applied due to the power difference of the modules?

[Luca Antognini]

Hi!

If you could provide the full detailed about the module (the missing Isc and voc values) that would be easier to reproduce your example.

The simulation you want to describe consists of different strings connected in parallel into a single MPPT, which will create some small losses of voltage mismatch between the strings (they don't have all the same Vmpp).

In this view, none of the two first simulations setup you describe can include this effet, as they both represent strings connected either to different MPPT or different inverters. They won't suffer from voltage mismatch and lead the same results as you pointed out. And there are no ways to define exactly the configuration you are describing from the system window (as this is quite an unusual one).

The correct approach is indeed to add these losses later in the detailed losses menu, by evaluating them roughly (they are rather small in the end) independantly of the main simulation.

Now, for the difference between your calculation and PVsyst calculation, I see first that in your table you update the voltages to lower values but keep the current unchanged. However, if you reduce the voltage from MPP, it will increase the current and therefore the power is slightly higher than in your evaluation. And consequently the losses would be smaller. It is really not surprising that PVsyst gives you such a small value when you have such as small difference of voltage among strings.

Another impact you should consider is that the MPP of your configuration will be closer to the one of the strings formed of 710W and 715W pannels than the one of the string composed of the 705W, because they have much more strings in parallel (36 and 37 vs 8). Therefore there is more to loose to move the voltage for those one than for the 8 strings of 705W pannels.

If you want to see this effect, you can also check it with another tool : Tools / Electrical behaviour of PV Arrays / heterogeneous arrays

(note: the limit of the tool is 20 strings but it is not a problem for us).

You can see that the relative losses are smaller than 0.1%. And this is already a pessimistic scenario here.

So In the case of your system you could consider neglecting completly those voltage mismatch losses.