Melting snow off PV..

icarus

I have never had to worry about my panels getting snow covered as either I am there to shovel them of, or I install panels flat on the wall for winter exposure so that they don’t get covered.  

Here is my dilemma...I have been commissioned to do a system design for a grid tied,  remote island location, in Northern MN.  As I said, they have the luxury of having grid power even though they are a mile or so off shore.  The design criterion is to off set summer seasonal use with winter production.  For 8-9 months there will be exactly ZERO loads on the system, summer time the loads are pretty substantial.  I will do the calcs to see how big a system would be needed to roughly offset the power needs.  

My question however, is that this system is going to be unattended and likely inaccessible  in all off the winter season so that no one will be able to routinely shovel or sweep the PV.  Given the design criterion, winter harvest is critical for the economic performance of the system.  So the question is how best to attack the problem?  My first inclination would be to simply put roof rated dielectric heat tape in some pattern around the PV.  I know from my own experience, even on a very cold (-30F) day, a bit of black panel exposure, and the panel pretty quickly clears itself.  

A number of problems come to mind with this strategy.  The first is making sure that sliding snow does not dislodge the heat line.  The second is having the system shut off when it is not needed.  In a normal heat tape situation, a simple T-stat suffices, but in this case, once the panels are clear, even it if it really cold, the heat line doesn’t need to be on.  It would seem to me that a link to the inverter such that based on time of day, if the inverter “should” be producing but isn’t, it could trigger the tape to come on, once the inverter produces X watts (by design) then it would shut off until the next cycle, reducing the parasitic energy cost to production.  

A couple of things to note, there is 3 phase power to the site, and I’m guessing at this point we might have 2-3 10KW arrays.  

So any ideas both on the hard ware as well as the software would be appreciated.  

Tony

mike95490

Put the panels on a pole mount, at a steep winter angle, and above the possible snow shed height.

 Possibly strap a paint shaker to the pole on a timer ?? Ha ha  just kidding about the paint shaker

BB.

Some utilities will "install GT Solar systems" elsewhere (perhaps onshore so the panels can be maintained?).

-Bill

icarus

Bill, Mike...thanks for the replies.

This is for an environmental non profit, and much of the value of a PV installation (to them) is the PR value, so off site doesn’t not suit thier purpose.  Pole mount is an interesting idea, and perhaps I will explore the option.  That said, there is no one available through out the winter to sweep them.  I would do my “normal” wall mounts on hinges, but the buildings all have trees or over hanging porches.

keep the ideas coming folks,

Tony

706jim

Zero loads during the winter? Place panels vertical and they will charge with no accumulation from snow.

Estragon

A somewhat "Rube Goldberg" solution
https://en.m.wikipedia.org/wiki/Rube_Goldberg_machine

Aim a Raspberry Pi zero with a camera module and sleepy-pi board at the array.  IIRC, the sleepy-pi will take ~6-30vdc to power the 5v Zero&camera.  Could be a little vertically mounted 12v panel & small powersport battery.

Have the sleepy-pi wake up the Zero at preset time(s) daily, take a pic of the array, and calculate the whiteness of the image.  If it's too white, trigger a relay to power the heat tape (and optional paint shaker).

Or you could use racking that can be set vertical for winter.

BB.

Then there is the question of how long will snow (on average) remain on the panels... If we assume "snow season" is around November through March:

https://www.weather.gov/dlh/duluthwinterclimate

And use PVWatts to figure out sun (I just pulled "Fox, Mn" out of a hat), fixed array, 50 degrees tilt from horizontal, 77% system efficiency (Grid Tied), numbers below based on 1 kWatt array (dollar value below based on $0.113 per kWH):
https://pvwatts.nrel.gov/pvwatts.php

Month	Solar Radiation ( kWh / m2 / day )	AC Energy ( kWh )	Value ( $ )
January	2.74	70	8
February	3.76	84	9
March	4.89	116	13
April	5.69	122	14
May	5.30	114	13
June	5.35	111	12
July	5.92	122	14
August	5.62	118	13
September	5.12	106	12
October	4.16	94	11
November	2.91	68	8
December	2.28	56	6
Annual	4.48	1,181	$ 133

And figure out the lost harvest assuming array covered by snow for November through March (may be a worst case/false assumption):

• 1,181 kWH total - (68+56+70+84+116 snow months) = 786 kWH non-snow harvest
• 786/1,181 = ~67% harvest
• 1/0.67 = 1.49x larger array assuming "snow cover months" (probably "overkill" for snow on panel guess?)

PVWatts also includes a spreadsheet download option... If the utility was both seasonal and time of use rate plan... You could download the hourly report and get a better handle on solar $$$ harvest on the hourly basis.

And there is looking at the Utility Rate Plan... Looks like there are something like 124 municipal electric companies in Minnesota... So I cannot even begin to guess there. Many utilities offer Summer vs Winter rate plans. In California, summer rates are significantly higher (irrigation, A/C)--Don't know about Minnesota--But it may be interesting dollar wise, to see if winter pricing will depreciate winter production more.

Tony, you have much more snow on panel experience than I (me, none). Any idea how long an array remains snow covered in the far north?

-Bill

BB.

Just for the heck of it, here is the 90 degree vertical harvest:

Month	Solar Radiation ( kWh / m2 / day )	AC Energy ( kWh )	Value ( $ )
January	2.78	71	8
February	3.48	78	9
March	3.94	95	11
April	3.84	82	9
May	3.08	65	7
June	2.97	60	7
July	3.36	68	8
August	3.61	75	8
September	3.87	80	9
October	3.72	84	9
November	2.86	66	7
December	2.35	57	6
Annual	3.32	881	$ 98

It does not increase winter harvest at all (from a solar perspective). And it "kills" summer harvest by something like 25%.

• 1/0.75 vertical array derate = 1.33x larger array assuming vertical all year round

Interesting... A vertical mounted array need a 1.33x larger system (fixed mount) vs a 1.49x larger array at 50 degree (optimum year round) mount. While I was going to write off the vertical mount as killing harvest and "not worth" the winter harvest--It turns out that having a year round snow free array vertical mounted is better.

I would guess that a mechanically configured 20-30 kWatt array moved 2x per year is not a huge interest to them. Could always look into some sort of 1 axis tilt motor setup (cost/maintenance wise).

And check the electric rates... If it is still ~$0.113 per kWH, then a 30 kWatt array in "optimum" conditions (no snow on panels) is something like (30 kWH array * $133 per kWH = ) $3,990 per year utility offset. Economically, I would guess that the installation and maintenance costs have to be kept on the low side.

-Bill

Estragon

November through March isn't out of the question.  Tend to be heavy dumps in late fall/early winter, then more or less the same snow blowing around until March, when we get another dump or two.  Last year I had to shovel a foot of snow off the dock and boat when closing up for freeze-up in mid-Oct.

My panels (at ~65°) seem to get covered in late Nov most years, and would stay that way until spring if not shovelled.  Snow on the ground can persist well into April/May, but I think the panels would clear in March.

Besides snow, the other problem is the low sun angle.  Trees, hills, etc to the south cast really long shadows around Xmas.

Dave Angelini

Since there is power, it is a no brainer to me. Build enough solar to offset the yearly power bill. Harvest only when it is not snowing or do not consider winter harvest. Especially if this is to be green and for "PR value"

Motors for tracking need to be attended. They will fail if left to their own. The concept of tracking the snow off will eventually fail if the snow turns to ice. I have seen a 100 degree elevation tracker freeze the snow and kill the motor. Heaters seem to be a PITA if the grid is there.

Good Luck Tony.

bill von novak

icarus said:

So any ideas both on the hard ware as well as the software would be appreciated.  

Go simple.  Do a 'standard' array optimized for latitude and weather, then do another vertical array for the winter.  Solar panels are cheap nowadays.  You can significantly overpanel the inverters if they are set up like that.

Estragon

bill von novak said:

icarus said:

So any ideas both on the hard ware as well as the software would be appreciated.  

Go simple.  Do a 'standard' array optimized for latitude and weather, then do another vertical array for the winter.  Solar panels are cheap nowadays.  You can significantly overpanel the inverters if they are set up like that.

That's what I did to keep batteries happy off grid, but I think the objective here is to sell into grid over the winter.

I assume they have/will check with the utility on feed-in terms & pricing on that size installation.

icarus

Ok...let me respond to all as I can...

First, the whole point of this exercise is to offset major summer usage with September-May harvest.  The PV watts numbers are interesting (thanks Bill) especially the wall angle mount.  If it were me, that is exactly what I would do, but I don’t get (much) of a vote.  The problem on the site is that there are few south facing walls that do not have covered porches.  They also are significantly shaded by trees that owners are loath to lose, roofs are better.  On all the off grid installations I have done or designed that are unattended in the winter have at the very least a mix of wall mount and roof mounts.  Some have hinged wall mounts which is a great solution...but not here.


Dave, your suggestion is good, except that convincing the owners to spring for that much PV is going to be well neigh impossible.  

I still think that heating the array ought not be such a problem, but the mere fact that no one has developed the technology in a consumer version tells me that it is more complicated.  There is a company known as “Blizzard solar” that has a technology apparently, but no product!  Someone once suggested that introducing reverse current into the PV to induce some heat, simple enough, except for optimizers and blocking diodes.  I am would be willing to try it on a beta panel, but not on someone else’s large array.

Once again, thanks for the input so far, and keep the ideas coming forward.

Tony

PS, we see our first snow in mid September, first accumulating snow by early October. By mid October (most years) any roof mount  PV would be covered, and would remain so, possibly until mid April, and even into May.

mcgivor

Some reading material 
http://www.keisolar.com/automatic-solar-panel-snow-removal/
https://tech.nikkeibp.co.jp/dm/atclen/news_en/15mk/031601972/?ST=msbe
http://www.scirustechnologies.com/product.html

BB.

You might also try contacting boB or Robin at Midnite (one of the founders)... I seem to remember many years ago, there was a plan to implement self heating of the solar array to melt snow and ice (probably pushing battery current through the solar panels by reversing solar array voltage and forward biasing the diode junctions).

But I never heard anything about it again (possibly planned for the Midnte Classic MPPT charge controller family????).

http://midnitesolar.com/

Sometimes boB drops by here and could comment (if he sees this thread).

-Bill

icarus

Thanks Bill, I will contact Robin or boB at midnight.

Mcgivor, thanks for the links.  I have written to two of them, the Nikkibp is a pv maker which would be a last resort.  I’ll report back with any information. Please keep the suggestions coming, as we are getting somewhere.

Tony

NANOcontrol

I don't know if the HAIN system actually mentions it, but you don't need any extra heaters to melt the snow.  The panels themselves make very good heaters.  Panels are just a series of diodes.  Exceed the forward voltage of them and they start to conduct. Like any diode it will heat up.  As long as you stay below the rated current of the panels it will be fine. It would be an interesting project. Disconnect the charge controller and use the battery to heat one or two panels. When those are sensed to be free, use them to heat the other panels. Or simply sense if the panels are making power by a certain time. If not back feed power from the grid for a certain time. Like they say, rinse and repeat till they are functioning.

Estragon

IIRC, Midnite did consider adding snow melt to Classics.  They didn't, according to boB, because it wouldn't work very well (or possibly at all if the Classic was fully panelled).

It could work if the number of panels to be heated was a smallish proportion of capacity, and ambient not too cold.  Off-grid, there would be the rather serious risk of depleting batteries while not successfully clearing to the extent needed to charge.

Dave Angelini

It will be interesting if you can find a way. Without proper funding I would walk unless these folks are dear to your heart. Unattended power systems have to also have funding for when it fails.

icarus

Dave...I hear you.  Mercifully, I have no dog in the fight.  The bottom line for me is to make some rational suggestions to them, and answer thier questions.  Any final design will be left to a local installer.  I am only acting as an advisory consultant...

That said, I do like the challenge of working in cold climates for remote sites, along with the requisite challenges.  These folks are not really dear to my heart, but for a variety of mother reasons I am their go to guy on a number of technical matters.

Nano...this is a grid tied system.  If you could provide any simple sketch drawing and or any more detail about how to energize a panel in reverse that would be helpful.  Secondly,  how would reversing the current affect either optimizers or RSS devices required by 2017 code?

Once again, great ideas folks, keep them coming...

Tony

NANOcontrol

No one mentioned optimizers.

"The design criterion is to off set summer seasonal use with winter production.  For 8-9 months there will be exactly ZERO loads on the system, summer time the loads are pretty substantial."

I can't wrap my head around that, storing winter production for summer use. Why not add just more panels in the summer.

BB.

That is an "advantage" of grid tied... "Unlimited" energy storage. A smaller system that generates 12 months of the year, and use power just during summer.

I did a calculation and if the system was just a non-snow season harvest, the system would have to be 1.49x larger. If it could harvest 12 months of the year (set panels to vertical to shed snow)--Then the system only needs to be 1.33x larger (panels set to optimum year round harvest, no snow, sizing of system = 1.0x).

If the system was sized for 3 month harvest, it would even need to be larger yet... And more costly.

Using the concept of reversing the voltage to the solar panels, then optimizers would be certainly confused, if not damaged (are optimizers one of the new ways to shutdown a solar array--Rapid Shutdown NEC requirement?).

I would guess that the voltage needed to reverse current flow th self heat the panels--Roughly 0.5 to 1.0 volt per "diode drop" (a solar cell simply being a forward biased diode when "heating"). That would mean a 60 cell panel (Vmp~30 volts) would need 30-60 volts to get substantial current flow.

And since diodes are notoriously for "poor" self current regulation (very small voltage between near zero current flow and damaging current flow, also temperature dependent), the "voltage source" would also have current regulation (something like 30-50+ volts output and 1 amp of current (30-60 volts @ 1 amp per parallel string) would be ~30-60 Watts of "heating (P=V*I) per panel. Put 10 panels in series (Vmp-array ~ 300 Volts and 1 amp current for heating) would be 300-600 Watts of self heating per ~2,000 to 3,000 Watts of array--Assuming panels are ~200-300 Watts each).

-Bill

Dave Angelini

NANOcontrol said:

No one mentioned optimizers.

I can't wrap my head around that, storing winter production for summer use. Why not add just more panels in the summer.

And yes it always comes down to money. That is the problem here with doing something that does not make sense and not having a budget to do it.

My latest thought for Tony is to have them invest in an energy company in the southwest that uses solar. At least if they are green they could hold their head up and say we are doing this for the planet.

Bill the unlimited energy storage concept of the grid is not working south of here today. I am off after I post to see why the batteries are dead and the grid has been down for 3 days.....

mvas

BB. said:

Just for the heck of it, here is the 90 degree vertical harvest:

Month	Solar Radiation ( kWh / m2 / day )	AC Energy ( kWh )	Value ( $ )
January	2.78	71	8
February	3.48	78	9
March	3.94	95	11
April	3.84	82	9
May	3.08	65	7
June	2.97	60	7
July	3.36	68	8
August	3.61	75	8
September	3.87	80	9
October	3.72	84	9
November	2.86	66	7
December	2.35	57	6
Annual	3.32	881	$ 98

It does not increase winter harvest at all (from a solar perspective). And it "kills" summer harvest by something like 25%.

• 1/0.75 vertical array derate = 1.33x larger array assuming vertical all year round

Interesting... A vertical mounted array need a 1.33x larger system (fixed mount) vs a 1.49x larger array at 50 degree (optimum year round) mount. While I was going to write off the vertical mount as killing harvest and "not worth" the winter harvest -- It turns out that having a year round snow free array vertical mounted is better.

I would guess that a mechanically configured 20-30 kWatt array moved 2x per year is not a huge interest to them. Could always look into some sort of 1 axis tilt motor setup (cost/maintenance wise).

And check the electric rates... If it is still ~$0.113 per kWH, then a 30 kWatt array in "optimum" conditions (no snow on panels) is something like (30 kWH array * $133 per kWH = ) $3,990 per year utility offset. Economically, I would guess that the installation and maintenance costs have to be kept on the low side.

-Bill

How is the Vertical Array at 881 kWhr / year BETTER than the 50 degree array at 1,181 kWhr / year ?

Avg Snow fall ...
Oct = 1"
Nov = 9"  <<< Avg Daily High is 40 degrees = lots of melting
Dec = 12"
Jan = 12"
Feb = 8"
Mar = 10" <<< Avg Daily High is 40 degrees = lots of melting
==============
Total = 52" 

9" of snow in Nov, is not very much snow.
10" of snow in Mar, is not very much snow.
And in both months, avg high is well above freezing = melting

If you go vertical, then you have wind load and snow drifts.
I must use the snow blower to clear a path in front of my array - the snow will pile up!
At 50 degrees tilt up from horizontal, the snow will slide off, when warm enough.

Estragon

The 50° array isn't 1181kwh/yr if it's covered by snow all winter.

mvas

At 50 Degrees tilt the snow will melt and slide off, even when the temp is below 32 F.
I live in an area that gets 50" of snow avg / year - that is not a lot of snow.

Minneapolis, Min ...
November has avg Highs approx 40F = melting
March has avg Highs approx 40F = melting

Even in Dec, Jan & Feb when the High Temp stays below freezing, I get melting and my panels are clear.

Pretending that we get 0 Watts of production from Nov 1st until March 31st ( as shown in message # 7 ) is not the truth, and helps nobody.

BB.

It was a worst case assumption... And I said I did not know how "sticky" the snow was in the region (and was not even sure of the region/lake effect snow/etc.). It was more of a demonstration of how somebody could use data to make quick estimates of harvest by month (or even by day, if somebody wanted) vs array angle (assume fixed array, or if somebody wanted a summer/winter angle/etc.).

And different people have different experiences... Your panels clear. Another poster (a bit farther north) has experienced late November through March array snow cover...

I was not presenting a "truth"... Just data, an analytical process, and even more questions.

-Bill

Dave Angelini

I have seen snow freeze to ice and even on a tracker going 90 degrees vertical to zero horizontal.

Solid ice for days until a human showed up. This was up the hill from me at latitude 37.

Keep the data coming. That investment in a south west power plant is how I would go.

Dave Angelini

If a budget was not an issue another way would be to use bifacial modules on a tall adjustable series of arrays.
They are making a comeback in commercial and at least 4 manufactures that I know of.

Pricey and a history of delam issues but definitely power on the backside. Wind on an island is an enemy of any tall vertical array unattended. Did something like this in Panama and the old guy who bought the system was informed that it would not last more than 10 years if he took care of it. He said, "Dave I only need 5 years"

icarus

Keep those cards and letters coming in folks!  I am inclined (at this point!) to do a beta test with a few panels with some heat line heat tape.  As Dave notes, snow that turns to ice will not come off until spring unless prodded.  A band of heat line across the top of each panel array should melt it enough to clear the panel quickly on a sunny day.  I’ll try it this winter on my own off grid roof system (that I usually sweep or shovel, and see how quickly it clears after a snow fall) That might give me a reasonable beta?

Tony

BB.

Embed a temperature probe between the heat tape and the panel (?).

I wonder about heat tape (plus some insulation to 'force heat' toward panel front) sort of ends up with some conflicting heat flow requirements (Want to heat panel in winter, but cool in summer/under direct sun).

As long as you are experimenting, how about a ~20-40%+ methanol+water mix and sprayer bar at top of panel ("real" spray rig would need to be UV resistant, possibly compressed air purge, insect resistant--do bugs like methanol?--Methanol is illegal in California except for winter/mountain snow areas, I think):

https://novosolution.ca/images/Freezing-Points-Methanol.pdf

Maybe look into Aircraft de-ice type 1 (actual de-ice) or Aircraft de-ice type 4 (green slime to let snow/ice slip off as aircraft?

-Bill

PS: https://en.wikipedia.org/wiki/De-icing