Bifacial Gain Expectations by Installation Type
Key Takeaway
Bifaciality is a module factor from a stated test. On three Weltrus TOPCon parameter tables it is printed as 80% plus or minus 10 percentage points. Bifacial gain is the extra energy a particular installation makes from light on the rear. Rooftop arrays are often low. Elevated ground-mount arrays are often higher. High-albedo surfaces can be higher still. Those sentences are patterns for a yield model. They are not a Weltrus warranty and not a promised percent.
Table of Contents
When One Gain Percent Is Pasted onto Every Site
A datasheet says bifacial. A slide then prints a single extra-energy percent and uses it for a membrane roof, a tracker, and a white-gravel yard. The three sites do not see the same light on the back of the module. The roof rear faces a dark surface a short distance away. The tracker rear sees the ground between rows, and also sees the torque tube. The gravel yard reflects more, until the gravel soils or the coating wears.
Buyers feel the error twice. First in the energy model, where a field percent on a roof overstates kilowatt-hours and can hide a design that only works if the rear contribution is real. Then in the contract, if that percent is described as something the module supplier stands behind. A bifaciality line is a laboratory ratio. An annual gain is a site output. The limited warranty, unless it expressly adds rear-side energy, is neither of those documents.
The practical split is simple. Ask the factory for the bifaciality factor, the test that defines it, and the tolerance. Ask the designer for albedo, clearance, row pitch, rear shading, and the inverter’s clipping behavior. If either answer is missing, the gain cell in the model is an assumption. Label it as an assumption.
Keep the Module Factor and the Site Gain Apart
On the Weltrus parameter tables for the 465-490 W, 540-560 W, and 645-670 W TOPCon bands, bifaciality is 80% with a tolerance of plus or minus 10 percentage points. This note reads that tolerance as a factor from 70% to 90%. It is the share of rear-side performance relative to the front under the test the datasheet names. It is printed the same way on those three bands. It is not printed here as a family value for every wattage from 100 W to 700 W. A model outside the three bands needs its own line. The datasheet for that model controls.
A simple unshaded estimate of the instantaneous power increase is the bifaciality factor multiplied by the ratio of rear irradiance to front irradiance. The estimate ignores non-uniform light on the rear, mismatch between cells, the angle of the rear light, soiling on the back, and hours when the inverter is already at its limit. It is a way to see how large the term can be before those losses. It is not a forecast and not a Weltrus measurement.
The tolerance belongs inside that product, not as a footnote. A buyer who types 80% and ignores the plus or minus 10 percentage points has picked the middle of a range the table itself prints. Run the low end and the high end. Then apply the site losses. The number that survives is the one that belongs in a sensitivity case, still labeled as a model, not as a guarantee.
Front-side nameplate stays the commercial watts. Standard test conditions are 1000 W per square meter, 25 degrees C cell temperature, and the AM1.5 spectrum, measured on the front unless the datasheet says otherwise. Rear gain is extra energy in the yield model. Where these three tables print first-year loss of 1.0% or less and then 0.35% per year, 29 steps after a 99.0% end of year 1 land at 88.85% of nameplate at the end of year 30. That floor is a share of nameplate under the warranty’s test. Do not add a bifacial gain percent on top of it and call the sum a warranted yield. Rear-side watts enter the warranty only if the warranty text adds them.
Temperature still applies to operating power. The coefficient of maximum power on these three tables is -0.26% per degree C. A hot roof lowers front-side power while the cell is hot. That drop is not degradation, and it is not a reason to inflate bifacial gain to “make the energy back.” Model the temperature on the front, model the rear irradiance for the installation you are actually building, and keep the two terms in separate cells.
Example Ranges by Installation Type
The table states how to treat three common installations. The words “often low” and “often higher” describe the direction engineers usually see. They are not Weltrus test results, not a promised band, and not a substitute for a site model. The arithmetic under the table uses example rear-to-front ratios so the bifaciality tolerance can be checked with a calculator. Those ratios are illustrations. Replace them with the irradiance your yield model produces for the drawing you will build.
| Installation | What the rear usually sees | How to treat gain in the model |
|---|---|---|
| Commercial roof, module close to a dark membrane | A near surface with low reflectance, often with the rear partly blocked by the roof itself | Often low, sometimes near zero. Do not import a ground-mount percent |
| Elevated ground-mount over soil, grass, or ordinary gravel | Ground between rows, plus shade from rails, piles, or a torque tube | Often higher than a flush roof. Still a site output. Height and row pitch move it |
| High albedo: pale gravel, a white coating, or seasonal snow | More reflected light, for as long as that surface stays in place | The high side of example models. Snow is a season, not a year. Coatings age |
Two example ratios, labeled as arithmetic
Take a rear-to-front irradiance ratio of 0.05 as a stand-in for a close, dark roof in an example model. Multiply by bifaciality. A factor of 0.70 gives 3.5%. A factor of 0.80 gives 4.0%. A factor of 0.90 gives 4.5%. The tolerance band on the module factor moves this particular example by about one percentage point, before rear shade and mismatch take more off.
Take a ratio of 0.15 as a stand-in for an elevated ground-mount example, not as a survey of any Weltrus project. The same three factors give 10.5%, 12.0%, and 13.5%. The gap versus the roof example is the site term, not a different Weltrus bifaciality. Both bands in this paragraph use the same 70% to 90% module factor. Realized annual energy is usually lower than the raw product because part of the rear is shaded, the rear light is uneven, and some sunny hours are clipped by the inverter. Clipping matters more when the front side is already near the inverter limit and the rear adds power the inverter cannot take.
A high-albedo case is the same formula with a higher rear ratio, held only for the hours that surface exists. Fresh snow can raise the ratio for days or weeks. A year-round model that types the snow-week ratio into all 12 months will overstate annual energy. Pale gravel and a white roof coating last longer than snow and still change when they soil. If the coating is part of the energy case, the maintenance plan has to say who restores it. The module factor does not maintain the roof.
What the three module tables do not add
Outline and mass differ across the bands: 1762 x 1134 x 30 mm at 28.0 kg, 1990 x 1134 x 30 mm at 25.0 kg, and 2382 x 1134 x 30 mm at 34.0 kg. A longer module on a tracker changes how much of the rear is shaded by the torque tube and how the row is divided. That is a geometry question for the tracker drawing. It is not a reason to assign the 645-670 W band a higher bifaciality. The factor printed for all three is the same 80% plus or minus 10 percentage points. Efficiency bands differ (23.3-24.5% on the first, 23.9-24.8% on the other two) and describe the front-side window. They are not a rear-gain rating.
Open-circuit voltage and short-circuit current still belong in the electrical design: 38.20 V and 15.92 A, 43.19 V and 16.09 A, 51.40 V and 16.16 A. Rear irradiance can raise operating current above the front-side figure. Size the fuse to the current the datasheet requires for bifacial use, not to a gain percent in an energy model. A gain percent is energy. A fuse is current. They are different columns.
Selection Boundaries
- Do not publish a single Weltrus gain percent. The factory figure you can cite from these tables is the bifaciality factor and its tolerance. The gain is the output of a model that names the installation.
- Do not move a ground-mount percent onto a commercial membrane, or a snow-week percent into an annual model. Write the installation type in the same cell as the percent.
- Do not apply 80% plus or minus 10 percentage points to a model that is not in the three bands. The datasheet for that model controls. A 100 W to 700 W catalog span is not proof that every code carries this factor.
- Do not infer glass-glass or a transparent backsheet from the factor alone. Rear construction changes the optical path. If the datasheet does not state it, ask. Do not fill it in.
- Do not treat IEC 61215 or IEC 61730 as a bifacial yield test. They are type tests. A bifacial energy number still needs the site.
- East-west roofs, carports, and tracker geometries need their own rear-shade model. The three rows in the table do not cover every mounting drawing.
When you compare quotations, ask each bidder for the bifaciality factor, the tolerance, and the test definition, and ask your designer for the rear-to-front ratio the model used, the albedo, the clearance, and whether clipping was included. If those six items are not written down, you do not yet have a bifacial energy case. You have a bifacial adjective.
Send the Installation Type With the Model Code
Tell Weltrus whether the array is a close roof, an elevated ground-mount, or a high-albedo surface, and send the model code. The reply can confirm whether the 80% factor applies to that code. The gain percent remains a site calculation, not a warranty line.
Frequently Asked Questions
Is bifaciality the same number as bifacial energy gain?
No. Bifaciality on the three Weltrus TOPCon tables in this note is a module factor: 80% with a tolerance of plus or minus 10 percentage points, read here as 70% to 90% under the datasheet test. Energy gain is the extra energy a site produces from light on the rear. The factor scales the rear irradiance. It is not itself the gain percent.
Does Weltrus warrant a universal bifacial gain percentage?
No. Rooftop arrays are often on the low side because the rear faces a close, dark surface. Elevated ground-mount arrays are often higher. High-albedo surfaces can be higher still. Those patterns are engineering examples, not a warranty and not a Weltrus measured yield. The limited warranty power floor applies to the test the warranty names, usually front-side nameplate, unless the warranty text adds the rear.
How does the plus or minus 10 percentage point tolerance move a sample calculation?
A simple unshaded product is bifaciality times rear irradiance divided by front irradiance. At a rear-to-front ratio of 0.05, factors of 0.70, 0.80, and 0.90 give 3.5%, 4.0%, and 4.5%. At a ratio of 0.15 they give 10.5%, 12.0%, and 13.5%. Both ratios are example inputs, not site measurements. Realized energy is usually lower after rear shade, mismatch, and inverter clipping.
Can one gain percent be used for every Weltrus module from 100 W to 700 W?
No. The 80% plus or minus 10 percentage point factor is printed for the 465-490 W, 540-560 W, and 645-670 W tables. A model outside those bands needs the bifaciality line on its own datasheet. Even inside the bands, the gain still changes with the installation.
Does a white roof coating or fresh snow lock in a higher annual gain?
Only while that surface is actually there. A coating that darkens, gravel that fills with soil, and snow that lasts a few weeks all change the annual number. Put the albedo you measured, or the albedo you are willing to maintain, into the yield model. Do not treat a short winter week as the year.
