Module loss

6 posts / 0 new
Last post
pstuder
Module loss

Hi,

I am trying to gain a better understanding of "Module" loss in the loss diagram for PV systems. I was wondering what are the exact list of things that are taken into account in the module. If I were to compare "Module" to PVsyst losses, would they be IAM losses, PV loss due to temperature, and PV loss due to irradiance? It isn't very clear to me, and a list of what is taken into account for "Module" would be very helpful.

Thank you!

Paul Gilman

Hello,

The losses that SAM reports in the "module loss" category that you see in the loss diagram and in the "single value" category on the Data tab and elsewhere on the Results page includes all power losses between the irradiance incident on the plane of array (POA irradiance) and the module DC electrical output. The details of that loss value depend on which module model you use. For example, the CEC module model and Sandia module model use different methods for calculating the module output and model those loss mechanisms differently.

In general, the module loss category includes the following:

  • Incidence-angle-related losses (IAM) losses.
  • Spectral-related losses.
  • Losses associated with converting light to electricity.
  • Temperature-related losses.

Because of the way those losses are accounted for in the model calculations, there is not a straightforward way to report each loss as a reduction in the module's electricity output in kW. There are some ways to estimate the kW impact of each loss mechanism, but we have chosen not to report those estimated loss values because of the uncertainty associated with the estimates, and because there is not a way to report those losses for the different module models in a consistent way.

For example, for the CEC module model, the cell temperature is part of the equation that determines the value of the light current parameter of the single diode representative model (See Equations 9.26 and 9.11 in the SAM PV reference manual available on the Performance Documentation page). As you can see from the equations, there is not a simple relationship between the values of cell temperature, the light current parameter, and the module DC output. For the Sandia model, which is an empirical model rather than single-diode model, the cell temperature is part of several of the empirically-deterimined equations (See Equations 9.4 through 9.8 in the reference manual). 

Best regards,
Paul.

Paul Gilman

Here is some more detail on the module loss value reported by the detailed photovoltaic model that I have adapted from an email conversation.

Question 1

The following table shows parameters and results for the four module types.

 

MultiSi

MonoSi

CdTe

CIS

Efficiency (%)

16.1

21.1

16.3

16.3

Tnoct

43.3

45.7

46.3

49.3

Temperature coefficient Vmp (%/C)

-0.43

-0.352

-0.388

-0.380

Calculated module loss (%)

8.2

8.5

7.6

6.0

Energy yield (kWh/kW)

1617

1609

1638

1665

The modules in SAM that represent each type are:

  • Multicrystalline Si: Trina Solar TSM – 310 PD14
  • Monocrystalline: SunPower SPR-X22-475-om
  • CdTe thin film: First Solar FS-4117A-2
  • CIS thin film: Solar Frontier SF170S

In the table you can see that the module efficiency and NOCT temperature for each module from the inputs, and the module loss and energy yield that SAM calculated. I used the same layout configuration and solar resource for each module. There is not a correlation between module losses and the temperature coefficient, which indicates that there must be other parameters that are influencing module loss.  

The module loss calculation are a black box to me. I see from the discussion above that there are other parameters (IAM losses, spectral-related losses, etc.). Can you explain how these factors relate to the difference in module losses as given in the table above?

Question 1 Answer:

SAM uses the same IAM loss calculations for the different module types, so the IAM contribution to the module loss should remain constant for the four module types in your example. SAM does use differeint IAM calculations for the different module model options (CEC single-diode, Sandia empirical, Simple Efficiency, IEC 61853), so if your analysis involves different models, that will contribute to the difference.

Like the IAM loss, the spectral-related loss calculations are the same for the different module types. Each module model uses a different approach to modeling spectral losses.

The CEC single-diode model uses air mass modifier coefficients and a polynomial equation developed for polycrystalline cells (from the Sandia model), regardless of the module type. That means that that for thin-film modules the spectral-related losses may not be as accurate as for crytsalline silicon modules. This is a known shortcoming of this model.

The temperature and conversion losses are the two categories that cause the differences in module loss in your example. Temperature losses are important in general, but in this example, the differences in module characteristics dominate. The parameters like Vmp and Isc that determine the module fill factor (shape of the I-V curve) are the primary causes of the differences.

As an illustration, we created this SAM file (SAM 2017.1.17 436 KB) with two cases and a parametric simulation. We used the "user entered specifications" version of the CEC single-diode model to make it possible to change the module characteristic inputs. For the first case, we used the original parameters for the Trina module in your example. For the second case, we modified the Voc and Isc parameters, which did NOT change the nominal module efficiency (on the Module page) but DID change the energy prediction (Results page Summary tab) and the module loss output (Results page Losses tab). That is because those parameters change the shape of the IV curve.

The parametric simulation (click Parametrics at the bottom left of the SAM window) shows the effect of varying some of the module characteristics one at a time. We hope that helps illustrate the importance of the other characteristics that are different for the four module types in your example. The last characteristic that we varied is the Pmp temperature coefficient, and you can see that when that is the only variable that changes, the relationship is as you expected the more negative the temperature coefficient, the higher the module losses.

Question 2

The module loss is also very much affected by selecting 1 axis tracking with or without backtracking for a given module. For example, for the module Trina Solar TSM-315PD14 at location California Bakersfield. The module loss is as follows:

  • 1 axis tracking with backtracking: 7.2%
  • 1 axis tracking without backtracking – no shading losses active: 6.9%
  • 1 axis tracking without backtracking and active shading losses: 9.2%

The module loss increases considerably when not using backtracking and activating the self-shading calculations. In addition, will have 1.1% losses due to shading for GCR = 0.3. My question is related to this observed increased module losses without backtracking. Do you know what is the reason behind it?

Question 2 Answer:

You should expect the module loss to depend on the backtracking option because of the the losses relating to the incidence angle. During some times of day (typically early morning and late afternoon), the array with backtracking will be in a different position than the non-backtracking array, which will cause different incidence angles.

For self-shading without backtracking, SAM provides two options: Non-linear and linear.

For the non-linear option, you will see an effect on both the shading and module categories of the loss diagram. For typical crystalline silicon modules, self shading reduces the incident diffuse irradiance uniformly across the array, even when the array is partly shaded. That effect is accounted for in the shading loss. For the incident beam irradiance, partial shading has a non-linear effect on the array's DC output because of the arrangement of cells, modules, and diodes in the array. SAM accounts for the effect of partial shading with a DC loss that it includes in the module loss.

For the linear option, SAM accounts for the effect of self-shading by reducing the incident beam irradiance, which only affects the Shading category in the loss diagram.

You can see those losses in the time series results (On the Results page, see the Data tables tab or Time series tab) listed for each subarray:

  • Subarray n Self-shading non-linear DC factor (contributes to module category in loss diagram)
  • Subarray n Self-shading non-linear ground diffuse irradiance factor (contributes to shading category in loss diagram)
  • Subarray n Self-shading non-linear sky diffuse irradiance factor (contributes to shading category in loss diagram)
  • Subarray n Self-shading linear beam irradiance factor (contributes to shading category in loss diagram)

Best regards,
Paul.

tric

Dear Paul,

according to the information provided here and in the SAM manual, the CEC module model considers spectral losses based on modelling coefficients derived for Si modules. However, the CEC database contains the module type as a parameter type, and the user-specified CEC module input page in the SAM GUI also explicitly asks to specify the module type (Si, CdTe, etc.).
How is that module type information used within SAM?

Thank you and best regards,
Timo

tric

Dear Paul,

I only now found the part in the manual that comments on this as it was only explicitly mentioned in the section about the user-defined CEC module. In case anyone else stumbles upon this issue: "This parameter is used to guide the solution of normalized module coefficients, but is not directly used for power calculations once the coefficients are determined."

Best regards,
Timo

Paul Gilman

Thanks for that clarification, Timo.

Theme by Danetsoft and Danang Probo Sayekti inspired by Maksimer