Has the software taken into account the low-light performance on the rearside ？

[Chen]

Dear sir：

  Has the software taken into account the low-light performance on the rearside of bifacial module？ thanks

[Chen]

Because the  irradiance of the front side or the rear side of PV module will be different, in this case, there will be a difference in conversion efficiency between the front and back sides,with respect to 1000 W/m². 

 

[Michele Oliosi]

Yes, actually the irradiance of front side + backside*bifaciality factor are put together, which gives an effective irradiance evaluation. Only after that, the PV conversion model applies. Indeed, the light reaches the same cells, be it from the front or the back. The same low-light performance is therefore taken into account for front and back (up to the bifaciality factor).

[Chen]

I understand. What I mean is, if the irradiation intensity on the back of the module is relatively low while that on the front is relatively high, so does it imply that the low-light performance of the rear side is worse? 

[Luca Antognini]

Clarifying Definitions to Avoid Confusion

At PVsyst, the term "low-light performance" is usualy defined in the context of a PV module illuminated solely from the front side.

Let’s consider the following example:

• At a reference irradiance of 1000 W/m², the module has an efficiency of 22%

• At a lower irradiance of 200 W/m², the efficiency drops slightly to 21.34%

We define low-light performance as the relative difference in efficiency:

(21.34 - 22) / 22 × 100 = -3% (relative) 

This means the efficiency at the lower light level is 3% lower relative to that at full irradiance.

This is the standard definition used in PVsyst.

Important Note: A more negative number (e.g., -4%rel or -5%rel) does not necessarily imply worse performance. In fact, it often indicates that the module performs better at high irradiance (1000 W/m²), which is an observed trend in modern PV modules, largely due to reduced series resistance

---

Bifacial Modules and Rear-Side Illumination

When it comes to bifacial modules, if only the rear side is illuminated (for the sake of the discussion), the module’s efficiency is proportional to the front-side efficiency, scaled by the bifaciality factor f.

This factor applies equally at both high and low light levels. Therefore, the relative low-light performance of the rear side is identical:

(f × 21.34 - f × 22) / (f × 22) = -3% (relative) 

---

Link to the Bifaciality Factor Definition

To relate this to my colleague’s earlier explanation: the bifaciality factor is defined as the ratio of power output when the module is illuminated from the rear versus the front. This difference in power is primarily due to optical property differences between the two sides of the solar cell. Indeed:

• The cell materials are different on the front and rear sides

• Any differences in charge transport or extraction are negligible (These effects are much much smaller than the typical uncertainty in datasheet bifaciality (e.g., 80% ± 5%))

Thus, the bifaciality factor can effectively be understood as a scaling factor for the photogenerated current.

In practical terms, the total irradiance received by a bifacial module can be calculated as:

Irrad_Tot = front-side irradiance + (rear-side irradiance × bifaciality factor) 

The PV module performance modeling can then proceed without needing to differentiate which side the light is coming from—the model is agnostic to the direction of light incidence and so is the low-light performance.

[Chen]

ok，I agree with your viewpoint, but do you think this calculation is reasonable?

if the fornt side of module receive irradiance of 1000 W/m², and the rear side of module receive irradiance of 200 W/m²,

 

front side：(f × 22- f × 22) / (f × 22) = 0% (relative)  
rear side：(f × 21.34 - f × 22) / (f × 22) = -3% (relative) 

[Luca Antognini]

Careful, the 200W/m2 I indicated was not refering to the rear side irradiance. It was a lower irradiance level, irrespective of whether the front or the rear recieves it.

I would agree with your calculation only in the case where the front side doesn't recieves any light. Otherwise you cannot decompose the efficiency in two parts from front and rear, it doesn't add up linearly. You need to consider the full I-V curve under the effective irradiance of both front and read sides together. 

-----

Let's take other numbers to not create confusion again.

If in certain scenario, the front side recieves 1000 W/m2 and the rear side e.g. 50W/m2, the module recieves an equivalent of 1000 + f*50 W/m2 of total irradiance. Let's say f=0.8, it is 1040W/m2.

If the irradiance level is lower, let's say 5 times lower, the front side recieves 200 W/m2 and the rear side 10 W/m2. Thereforer, under low light, the module recieves an equivalent of 208 W/m2 in total.

The curve of performance of the module as function of irradiance should be understood as function of this effective total irradiance. (Indeed, in the norms, when bifacial module are measured, the rear side is masked to not absorbed any light from the rear.) You can visualize this curve in the .PAN file:

So whether the light comes from the rear or the front, it is the Incident global light that matters. For this specific curve, if the Incident global

Incident global = front-side irradiance + (rear-side irradiance × bifaciality factor)

is equal to 200 W/m2, the efficiency will be -3.0% lower than if the Incident global was 1000 W/m2.

So any module PV module, bifacial or not, under given illumination condition will operate somewhere on this curve. If the module operates

• above the reference of 1000 W/m2, you have an efficiency degradation with respect to STC,
• between 500-1000 W/m2 (for this particular example), you have an efficiency boost,
• below 500 W/m2 you have an efficiency degradation.

It doesn't matter if it is bifacial or not.

However, the bifaciality has this influence that the module will operate at a higher illumination level than a monofacial counterpart (in the hypothesis they have exactly the same electrical properties), so at another position on this curve. But as you can see, it could well either mean an improved performance!

PVsyst of course take this into account during simulation and you can find it in the loss diagram. For example if you use the "Demo Commercial Oakland" with VC6, you can see it there

 

 

If for the sake of the test, you modify the .PAN file and set the bificiality factor to 0, you get then:

Therefore, for this specific example, on the yearly production you don't see a specific loss difference due to the irradiance level operating further from STC between a bifacial and monofacial module. This is because this effect is small here (I would call it a third order correction).

If you had a system with better Albedo for example, the comparison could be different.