Bifacial panel field testing?

[sensij]

To be clear, the above conditions are not the same as deliberately, and as part of a design, having diffuse or specular reflectors directly in front of/behind/around a surface or array. That's different and not considered here.

A close read of these two papers can help show why J.P.M.'s statements are true, but that bi-facial panels can still be employed in a cost-effective manner.

http://www.researchgate.net/profile/Eduardo_Pigueiras/publication/232903070_Estimation_of_Radiation_Incident_on_Bifacial_Albedo-Collecting_Panels/links/551016310cf2ac2905afcfa4.pdf

http://www.researchgate.net/profile/Richard_Perez2/publication/222444334_Ground-reflected_radiation_and_albedo/links/09e415087db85e0417000000.pdf

[J.P.M.]

A close read of these two papers can help show why J.P.M.'s statements are true, but that bi-facial panels can still be employed in a cost-effective manner.

http://www.researchgate.net/profile/Eduardo_Pigueiras/publication/232903070_Estimation_of_Radiation_Incident_on_Bifacial_Albedo-Collecting_Panels/links/551016310cf2ac2905afcfa4.pdf

http://www.researchgate.net/profile/...0417000000.pdf

Thank you for the references.

I have the Perez paper.

As for the first ref., I'd remark that the conditions used for reflectors are not commonly found in practice, and that 80% reflectivity is difficult to attain and maintain for diffuse surfaces, or even specular ones as a practical matter.

A bigger issue for me with the Krenzinger - Lorenzo paper is shown in fig. 2, where the mid day backside reflected irradiances are about 60% of the frontside for the "D" configuration. Given the dimensions of their model as listed, I'm not sure the view factors will allow that high a level, or close to it, at least as I learned to calc multireflective view factors. I'd also wonder about the use of what is described as "small cells" to measure front and back radiation.

Even if the view factors could be that high, maybe that high level of back side irradiance is possible for the model and rig they used, being, in effect, a form of a diffuse concentrator. However, in the past, I've been unable to do as well with mirrors in a very similar configuration. Now that was about 30 years ago, and I was using a hand held $60 solar meter in my back yard, but the best I could get, if memory serves, was about 20-30% backside enhancement and that was pushing it - and to repeat - using mirrors, not diffuse reflectors, bottom and back.

I'm also not sure I understand the authors' "Figure of Merit" and is use, or superior value over other methods. Guess I've got some additional rereading to do.

So, while not having data to share to buttress my opinions, I'll stick with my claims of a few % boost for bifacial panels over std. types in most practical situations, the higher figure for highly reflective (and I'd suggest relatively unusual and therefore somewhat unrealistic) conditions.

If vetted and reviewed data is presented that would cause me to reexamine my views, I'll do so. I have no trouble changing my opinions in the face of new data/evidence. When I do so, I try to keep in mind that data presented in formats such as presented here, while usually valid, technically competent and usefully informative, is not gospel, is subject to being incomplete, limited in application, or inappropriate or unrealistic for some conditions/situations.

[DanKegel]

1) The SolarEdge optimizers don't report live data simultaneously, and there is a randomizing factor in the timing to help avoid data collisions. I have 12 panels, and it typically is takes close to 10 min for all of them to report. 10 min is enough time that detecting a performance difference on the order of a couple percent is very hard to do, because the sun position, temperature, cloud cover, etc etc can change by more than that over that amount of time.

2) There are no canned reports or API access to panel level data. You can create a chart in the portal and download it with down to 15 min resolution (resolution is a function of how much time you want to include in the chart), but the chart interface is really aggravating to me so I haven't tried to set it up to capture the data.

So, just to confirm, per-panel performance data with something like 10 minute resolution is available to every SolarEdge
customer if they're willing to use the portal to generate a chart? Can you send me one of those charts so I can see if I
can parse it? (If it's a .pdf, there's a chance I can rip the data out of it.) Thanks! Maybe I can program a script to push the "give me a chart" button once a day, parse out the output
for the two side-by-side panels, and upload them to pvoutput.org.

BTW I wrote up what I hope to do at

http://kegel.com/energy/solar/retrofit2/

including pictures of the spot suitable for a reverse tilt mount,
with a shot of the shading from a palm tree at 8am.
I think it's shade-free from 9am to at least 4pm.
I'll have to watch that shading closely to see if it affects one panel more than the other.

[sensij]

Here is an example of an export of a panel power chart. Similiar data (and format) is available for energy, current, voltage, and optimizer voltage. I don't know why it only wants to give me data up to 5 pm today, but the rest will probably fill in at some point tonight. I stretched the report so you have yesterday's complete data for the day in there.

Charts can be saved, I'm sure someone with the required skills could automate it.

ChartExport 06_07_2015 4_17 AM.zip

Edit:
@J.P.M. Thanks for the comments. On track record / reputation alone, Ineichen and Perez are clearly much more established than the authors of the earlier paper. The nice thing about (most) published work is that enough information is provided that suitably trained or experienced readers can assess how practical the results may be, or determine that assumptions or experimental setup choices may limit their applicability. I was happy to find unbiased* work that has not been obviously written by vendors for their own benefit.

*all authors get funding from somewhere, of course...

[J.P.M.]

Here is an example of an export of a panel power chart. Similiar data (and format) is available for energy, current, voltage, and optimizer voltage. I don't know why it only wants to give me data up to 5 pm today, but the rest will probably fill in at some point tonight. I stretched the report so you have yesterday's complete data for the day in there.

Charts can be saved, I'm sure someone with the required skills could automate it.

[ATTACH]6989[/ATTACH]

Edit:
@J.P.M. Thanks for the comments. On track record / reputation alone, Ineichen and Perez are clearly much more established than the authors of the earlier paper. The nice thing about (most) published work is that enough information is provided that suitably trained or experienced readers can assess how practical the results may be, or determine that assumptions or experimental setup choices may limit their applicability. I was happy to find unbiased* work that has not been obviously written by vendors for their own benefit.

*all authors get funding from somewhere, of course...

You're welcome. I've met Prof. Perez a couple of times long ago and far away. Not a mentor, but I've learned a lot from his work and work ethic.

[DanKegel]

Here is an example of an export of a panel power chart. Similiar data (and format) is available for energy, current, voltage, and optimizer voltage. I don't know why it only wants to give me data up to 5 pm today, but the rest will probably fill in at some point tonight. I stretched the report so you have yesterday's complete data for the day in there.

Charts can be saved, I'm sure someone with the required skills could automate it.

Thanks. Was that generated by clicking on the "Export to Excel" button mentioned in
http://www.solaredge.com/files/pdfs/...user-guide.pdf ?

The sample you gave was for power (watts), and if I can download a day's worth of that
once a day, I'm ok.
Do you know if there's a demo account I could use for a day or so?
I'd like to try writing the script to automate downloading that data.

[sensij]

Yes, that was the export to excel output. Check your PM's.

[adoublee]

I've been tracking a 5.46kW system with Sanyo (now Panasonic) bifacial modules mounted south facing at 30deg tilt on a white TPO roof for about 5 years. These were/are top of the line and spendy panels. Output is indeed much closer to marketing literature than I would have expected. If you want to go to the effort of making a prediction of kWh yield based on whatever you want (like real whether files), I would play along. Just pm me if you want the location, I don't mind if any other data is posted. Data would be as reported by SMA 7kW inverter. 2011 is first full year, and panels were integrated into a larger tripower inverter mid-2014.

[DanKegel]

I've been tracking a 5.46kW system with Sanyo (now Panasonic) bifacial modules mounted south facing at 30deg tilt on a white TPO roof for about 5 years. These were/are top of the line and spendy panels. Output is indeed much closer to marketing literature than I would have expected

Do you have individual panel data, or just output of whole string? I'd love to hear what happens to your output if you cover the backs of the panels.

[J.P.M.]

I've been tracking a 5.46kW system with Sanyo (now Panasonic) bifacial modules mounted south facing at 30deg tilt on a white TPO roof for about 5 years. These were/are top of the line and spendy panels. Output is indeed much closer to marketing literature than I would have expected. If you want to go to the effort of making a prediction of kWh yield based on whatever you want (like real whether files), I would play along. Just pm me if you want the location, I don't mind if any other data is posted. Data would be as reported by SMA 7kW inverter. 2011 is first full year, and panels were integrated into a larger tripower inverter mid-2014.

Do you have an accurate way to measure insolation including albedo ?

[DanKegel]

How accurate a measurement of albedo do you need?
http://vixra.org/pdf/1110.0035v1.pdf suggests you can get within a few percent using a digital camera and a piece of white paper.

[adoublee]

Do you have individual panel data, or just output of whole string? I'd love to hear what happens to your output if you cover the backs of the panels.

Multi-string data (total inverter output). It is more practical and useful to try and evaluate how much energy we get out of it over time than how much more power there is at any moment in time.

Don't forget the effect of bright white clouds to the North of the panels increasing diffuse radiation to back of panels. There are many new variables and some new decisions to be made when trying to maximize backside contribution (think inverter sizing rules of thumb). For the Sanyo product, STC nameplate data specifically excluded backside contribution.

[J.P.M.]

Multi-string data (total inverter output). It is more practical and useful to try and evaluate how much energy we get out of it over time than how much more power there is at any moment in time.

Don't forget the effect of bright white clouds to the North of the panels increasing diffuse radiation to back of panels. There are many new variables and some new decisions to be made when trying to maximize backside contribution (think inverter sizing rules of thumb). For the Sanyo product, STC nameplate data specifically excluded backside contribution.

Time integrated total output is valuable, but adds more uncertainty due to more variables and how their variation over time affects performance.

[DanKegel]

I threw together a formal experiment writeup at http://kegel.com/energy/solar/retrofit2/

(EDIT) I realized the temperature coefficients of the LG305 and the GxB300 are so different
that I should only compare measurements from GxB300s when calculating back boost,
so now I'm planning to install two GxB300's side by side in a reverse tilt
(along with two flush mounted GxB300's next to all the LG305's).

I also found that Tyvek's a good high-albedo material, so I'll also test with that beneath the panels as a best-case condition.

I've updated the web page to match.

Feedback welcome.

[DanKegel]

Is that 2-3% for a dark background, or a light one?

Wow. Did I just like completely forget his whole post when I wrote that?

Rereading his post I see it was about 2% for a dark background, and 3% for a 30% albedo background.

adoublee said his Panasonic/Sanyo panels were mounted over a white TPO roof, and got surprisingly close to the sales claim.
Panasonic claims a 29% increase mounted over a 95% albedo surface, see

http://panasonic.net/ecosolutions/solar/hit/

Older data sheets claimed 26%:

http://www.solarselections.co.uk/blog/wp-content/uploads/2012/12/Panasonic-HIT-solar-panels-vbhn201de01-vbhn205de011.pdf

If he's getting even half of that, it seems quite a bit larger than JPM is predicting.