Melting snow off PV..

icarus

Or...now that I think about it, how about simply running dielectric heat tape under the panel, against the back sheet?  That way you have no shading issue, the sliding snow won’t interfere, and I wouldn’t think that thermostatically controlled heat line would run any hotter than a good summer day so over heating the back sheet shouldn’t be an issue.  That way I could run multiple strands as needed, and one could simply drill the frames, run heat line through grommets.  

In point of fact, with a well regulated heat line tape, you could have it run in anticipation of snow, making the melting even easier since the snow couldn’t accumulate.  It doesn’t take very many BTUs to keep snow of a drive way, shouldn’t take to many WH to keep snow from accumulating on the panels.

Thoughts?

Tony

BB.

I was wondering about heart transfer to the front of the panel vs out the back of not insulated.

And if insulated, summer heat kept in panel of insulated.

If using bi-facial panels, there would be rear shading of cells...

Bill

Horsefly

Our system is at 9,000 ft elevation in Colorado, and no one is up there basically from early November until late May due to the roads being impassable. I have a logger on the battery, and once the roads are OK I can load the data into a spreadsheet and examine it (some here probably remember some of the graphs I've posted with the data). 

We've only had the system through the past two winters.  The panels are ground mounted at a 40 deg angle.  The battery voltage is logged once per hour. I realized pretty quickly that I can tell when the panels are covered with snow. On normal days, I can see the charge controller voltages climb at least up to float, as most nights the batteries don't drop enough to cause bulk charging the next day. If the battery voltage barely moves through the next day, it's obvious that there isn't enough solar to move the needle. 

So here's the results: During the first winter, there was ONE 3 day period that there was virtually no solar production. This past winter, there was ONE 2 day period with no production. In both winters there was also a dozen or so days when the production was down, but not really by all that much. 

Remember, this is at 9,000 ft elevation in the mountains of Colorado. The winter of 2017-2018 was pretty mild, but conversely the winter of 2018-2019 was extremely wet in that part of Colorado (well, actually in all of Colorado). 

So with no one there to witness, I'm doing a bit of speculation based on the observations, and this is what I think:

1. I know it snows up there, and I know in previous years it has gotten as deep as five feet. So I know the panels do get snow on them.
2. It's very sunny there, even in the winter. If a storm does dump snow, it is somewhat likely to be sunny skies a day or two after the big snow. I think it doesn't take much sun to hit the frame of the panels and start warming them up. Even more so once a few of the cells become exposed.
3. I think the snow much mostly slide off within a day of most storms rolling out. 

So at least in my case, I would agree with those that say the impact of snow on my solar production has been minimal. Although I was worried about it when I was doing the design, it appears to be insignificant.

I should also point out that we have very low humidity in Colorado. We sometimes joke that we don't see the snow melt because it just evaporates. That not only contributes to clearing the panels quickly, I think it also prevents the kind of ice buildup people see in the mid-west and the great white north, where the humidity is higher.

Estragon

It seems to me my waterline backup heat line won't get much warmer than ~40°f, reducing current as it approaches that temp.  

I'm a bit north and west of International Falls MN.  Fluffy winter snow isn't much of a problem.  The problem is some fall/early winters which feature a sticky snow which form a covering layer of snow stuck to panels, effectively shielding the panel from sun until spring.  At ~65°tilt, I've found ~6-8" of snow stuck to the top of the array in Jan.  It takes a pretty long spell of warm weather to melt that, and it rarely happens here.  It pretty much takes a shovel.

Once a horizontal part of the array gets sun, the rest melts well (assuming room below for snow to slide off).

The summer lack of backside ventilation cooling could be an issue, but it likely  wouldn't take much more than a strip of rigid insulation maybe 2" inches wide x an 3/4" thick to heat the glass and substrate enough to create a water layer.

Assuming the array gets winter sun on the bottom, and room at the bottom for snow, maybe a heat trace on the lower back of the array makes sense?

mvas

icarus said:

Or...now that I think about it, how about simply running dielectric heat tape under the panel, against the back sheet?  That way you have no shading issue, the sliding snow won’t interfere, and I wouldn’t think that thermostatically controlled heat line would run any hotter than a good summer day so over heating the back sheet shouldn’t be an issue.  That way I could run multiple strands as needed, and one could simply drill the frames, run heat line through grommets.  

In point of fact, with a well regulated heat line tape, you could have it run in anticipation of snow, making the melting even easier since the snow couldn’t accumulate.  It doesn’t take very many BTUs to keep snow of a drive way, shouldn’t take to many WH to keep snow from accumulating on the panels.

Thoughts?

Tony

How will you know when there is snow on the PV Panels and then activate the heaters?

How many watts do you need to melt all of the snow on one 400 watt PV Panel?
100 watts ?
200 watts ?
300 watts ?
400 watts ?   

Doesn't that depend upon how low the ambient temperature is and the wind speed?
How will you know when the snow, if any, is 100% melted?

What if it snows at 1" per hour and the heaters can only melt 1" per hour?
You could spend all 4 peak hours of a short winter day melting snow,
and COSTING 40 kwHr = 4 hours x 10kw of heat energy per array,
and generating nothing ( 0 kWhrs ) = ouch !

How will you "smartly" control 10 kw of heaters per array?
If you add electric heaters, make sure you have enough heat at the bottom edge, to prevent an ice dam.
Does adding heat to a cold PV Panel void the warranty?

Personally, I would not add heaters, I let the sun and gravity do their job, for FREE ....

Estragon

I hoped the sun and gravity would do their job for free, but most years it just hasn't worked out that way.

mvas

Estragon said:

I hoped the sun and gravity would do their job for free, but most years it just hasn't worked out that way.

How much snow do you get per year?
On average, how many days does the snow stay on your panels?
What is the angle of your PV Panels?

Keeping costs to the bare minimum is so very important with Solar, to get a reasonable break-even.

I did not make my ground mount high enough and eventually the pile of snow would reach the panels.

icarus

mvas said:

icarus said:

Or...now that I think about it, how about simply running dielectric heat tape under the panel, against the back sheet?  That way you have no shading issue, the sliding snow won’t interfere, and I wouldn’t think that thermostatically controlled heat line would run any hotter than a good summer day so over heating the back sheet shouldn’t be an issue.  That way I could run multiple strands as needed, and one could simply drill the frames, run heat line through grommets.  

In point of fact, with a well regulated heat line tape, you could have it run in anticipation of snow, making the melting even easier since the snow couldn’t accumulate.  It doesn’t take very many BTUs to keep snow of a drive way, shouldn’t take to many WH to keep snow from accumulating on the panels.

Thoughts?

Tony

How will you know when there is snow on the PV Panels and then activate the heaters?

How many watts do you need to melt all of the snow on one 400 watt PV Panel?
100 watts ?
200 watts ?
300 watts ?
400 watts ?   

Doesn't that depend upon how low the ambient temperature is and the wind speed?
How will you know when the snow, if any, is 100% melted?

What if it snows at 1" per hour and the heaters can only melt 1" per hour?
You could spend all 4 peak hours of a short winter day melting snow,
and COSTING 40 kwHr = 4 hours x 10kw of heat energy per array,
and generating nothing ( 0 kWhrs ) = ouch !

How will you "smartly" control 10 kw of heaters per array?
If you add electric heaters, make sure you have enough heat at the bottom edge, to prevent an ice dam.
Does adding heat to a cold PV Panel void the warranty?

Personally, I would not add heaters, I let the sun and gravity do their job, for FREE ....

Don’t know...don’t know...don’t know!  That is the function fo the beta test.  As Estragon points out, it is not the light, fluffy snowthat is the problem, but the wet snow followed by cold early in the season, that can leave the PV covered until spring.  I know, I have seen it!

The whole point of this is to design a system, that while being left unattended on a remote island location, can off set summer loads with fall/winter/spring harvest.  To give up 4-6 months of harvest waiting for the sun and gravity to do the job for “free”, would make theproject much less feasible.

Tony

icarus

Estragon said:

It seems to me my waterline backup heat line won't get much warmer than ~40°f, reducing current as it approaches that temp.  

I'm a bit north and west of International Falls MN.  Fluffy winter snow isn't much of a problem.  The problem is some fall/early winters which feature a sticky snow which form a covering layer of snow stuck to panels, effectively shielding the panel from sun until spring.  At ~65°tilt, I've found ~6-8" of snow stuck to the top of the array in Jan.  It takes a pretty long spell of warm weather to melt that, and it rarely happens here.  It pretty much takes a shovel.

Once a horizontal part of the array gets sun, the rest melts well (assuming room below for snow to slide off).

The summer lack of backside ventilation cooling could be an issue, but it likely  wouldn't take much more than a strip of rigid insulation maybe 2" inches wide x an 3/4" thick to heat the glass and substrate enough to create a water layer.

Assuming the array gets winter sun on the bottom, and room at the bottom for snow, maybe a heat trace on the lower back of the array makes sense?

The place in question is on Rainy Lake, just outside International Falls/Fort Frances.  I am located (much of the time now, not permanently any more) 150 KM east, north of the Quetico park.

T

PS.  Maybe you want a job coming out to shovel off the system?  That would be a solution!  Keep in touch,

Tony

icarus

One keeps digging, and one keeps finding stuff!  There are snow sensors used for driving circuits either on roof melting applications, or pavement melting applications.  Couple these with t-stats and the duty cycle/WH use of heat line heat tape is pretty small.  Heat line ( a Canadian company) heat tape uses ~5 watt/foot at 50f.  If you ran two strings of heat tape under a ~300 watt panel in portrait orientation, you might use ~6’ of tape, or 30 watts.  My guess is that would keep most of the panel free from falling snow, and would certainly cause it to melt and slide as soon as there was sun on the PV.  The trick would be to install a snow sensor at the location on the array that keeps it’s snow longest, (or use multiple strings of heat line, with multiple sensors). Feels like more winter testing for me?

Tony

Estragon

mvas said:

Estragon said:

I hoped the sun and gravity would do their job for free, but most years it just hasn't worked out that way.

How much snow do you get per year?
On average, how many days does the snow stay on your panels?
What is the angle of your PV Panels?

Not sure about total snowfall, but IIRC I used ~2-3' snow loading for structural.  Not that much really.

Panels are at ~65° - fairly steep (vs year-round optimal ~45°).

Average days covered is meaningless.  They either get covered and frozen around Nov, or they don't.  If they do get covered, they stay that way until manually cleared, or spring.  If it's cold enough in early snows, the snow doesn't stick and the panels stay clear.

Estragon

Rainy Lake would be pretty much identical climate as mine (Lake of the Woods).

Last year was a fairly bad one for slush on the ice.  I barely made it in Jan to clear panels, and aborted a planned trip in Mar.  It was a wet fall and froze up with high water.  LWCB let it drop over the winter, which avoided spring problems, but made lots of cracks over winter as the water dropped.

Maybe if they bought me an ice boat, we could talk about me clearing their panels 

littleharbor2

Dave Angelini said:

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.

I'm picturing the Donner Party. 

mvas

icarus said:

mvas said:

icarus said:

Or...now that I think about it, how about simply running dielectric heat tape under the panel, against the back sheet?  That way you have no shading issue, the sliding snow won’t interfere, and I wouldn’t think that thermostatically controlled heat line would run any hotter than a good summer day so over heating the back sheet shouldn’t be an issue.  That way I could run multiple strands as needed, and one could simply drill the frames, run heat line through grommets.  

In point of fact, with a well regulated heat line tape, you could have it run in anticipation of snow, making the melting even easier since the snow couldn’t accumulate.  It doesn’t take very many BTUs to keep snow of a drive way, shouldn’t take to many WH to keep snow from accumulating on the panels.

Thoughts?

Tony

How will you know when there is snow on the PV Panels and then activate the heaters?

How many watts do you need to melt all of the snow on one 400 watt PV Panel?
100 watts ?
200 watts ?
300 watts ?
400 watts ?   

Doesn't that depend upon how low the ambient temperature is and the wind speed?
How will you know when the snow, if any, is 100% melted?

What if it snows at 1" per hour and the heaters can only melt 1" per hour?
You could spend all 4 peak hours of a short winter day melting snow,
and COSTING 40 kwHr = 4 hours x 10kw of heat energy per array,
and generating nothing ( 0 kWhrs ) = ouch !

How will you "smartly" control 10 kw of heaters per array?
If you add electric heaters, make sure you have enough heat at the bottom edge, to prevent an ice dam.
Does adding heat to a cold PV Panel void the warranty?

Personally, I would not add heaters, I let the sun and gravity do their job, for FREE ....

Don’t know...don’t know...don’t know!  That is the function fo the beta test.  As Estragon points out, it is not the light, fluffy snowthat is the problem, but the wet snow followed by cold early in the season, that can leave the PV covered until spring.  I know, I have seen it!

The whole point of this is to design a system, that while being left unattended on a remote island location, can off set summer loads with fall/winter/spring harvest.  To give up 4-6 months of harvest waiting for the sun and gravity to do the job for “free”, would make the project much less feasible.

Tony

And that is why I asked these very important questions - they need to be answered, or at least addressed

I think, 6 months is quite the exaggeration, given ...
a) Daytime avg High temps don't fall below freezing until Dec 1st and  then
b) Daytime avg High temps go above freezing by Mar 1st

Can some years be colder & snowier ?
Yes 
But then other years will be warmer and less snowy
These are the long term averages at International Falls, MN, regardless of what you saw one winter.

Yes, the density of snow varies ( infinity ) from about 20:1, to about 5:1 = snow depth vs melted water depth.
20:1 snow is light and fluffy - translucent
5:1 snow is heavy and thick - opaque
And there is every density in-between

I found that it takes about 100 watt-hours to melt 1kg of snow

assuming the 300 watt pv panel is about 1.5 m^2 area and ( 1" of snow ) is 2.54 cm thick ...
 0.04 m^3 volume of snow = 1.5 m^2 x 0.0254 m

assuming very dense snow at 5:1 ...
 0.008 m^3 of water = 0.04 m^3 snow x 1/5 density

water is 1000 kg/m^3 ...
8 kg = 0.008 m^3 x 1000 kg/m^3

If it takes about 100 watt-hours to melt 1kg of snow, so then 800 watt-hours* of electricity to melt 1" of heavy snow
It may be possible to melt the lowest layer of snow and have the slab of snow slide off, or not?
Efficiency will be less than 100%, especially when it is very cold and/or very windy.

Worst Case:
If it snows 1" every other day, for 104 days ( or 3.5 months ) then ..
You use 800 watt-hours to melt the 1" of snow
You collect 800 watt-hours the next day
NET - you have not collected any EXTRA energy for those 3.5 months ?

Best Case:
Many inches of snow fall in a day, or two, or three and then a week of clear skies?


At the minimum, this provides a starting point ...

Note:
* = assuming all of my math is correct

Estragon

In the five winters since my panels were installed, IIRC there was only one in which I didn't need to manually clear the arrays of stuck on snow.

The whole array wouldn't need to be melted.  Even if a fairly small area gets sun on the dark panel, unless badly drifted, in the panel will eventually clear.  In Nov/Dec we tend to get long runs of gloomy weather, short days, and low angles, which allows the snow to build and freeze to a ~6-8" crusty coating.

If cleared in January, they seem to reliably stay clear until spring.  FWIW

icarus

mvas said:

icarus said:

mvas said:

icarus said:

Or...now that I think about it, how about simply running dielectric heat tape under the panel, against the back sheet?  That way you have no shading issue, the sliding snow won’t interfere, and I wouldn’t think that thermostatically controlled heat line would run any hotter than a good summer day so over heating the back sheet shouldn’t be an issue.  That way I could run multiple strands as needed, and one could simply drill the frames, run heat line through grommets.  

In point of fact, with a well regulated heat line tape, you could have it run in anticipation of snow, making the melting even easier since the snow couldn’t accumulate.  It doesn’t take very many BTUs to keep snow of a drive way, shouldn’t take to many WH to keep snow from accumulating on the panels.

Thoughts?

Tony

How will you know when there is snow on the PV Panels and then activate the heaters?

How many watts do you need to melt all of the snow on one 400 watt PV Panel?
100 watts ?
200 watts ?
300 watts ?
400 watts ?   

Doesn't that depend upon how low the ambient temperature is and the wind speed?
How will you know when the snow, if any, is 100% melted?

What if it snows at 1" per hour and the heaters can only melt 1" per hour?
You could spend all 4 peak hours of a short winter day melting snow,
and COSTING 40 kwHr = 4 hours x 10kw of heat energy per array,
and generating nothing ( 0 kWhrs ) = ouch !

How will you "smartly" control 10 kw of heaters per array?
If you add electric heaters, make sure you have enough heat at the bottom edge, to prevent an ice dam.
Does adding heat to a cold PV Panel void the warranty?

Personally, I would not add heaters, I let the sun and gravity do their job, for FREE ....

Don’t know...don’t know...don’t know!  That is the function fo the beta test.  As Estragon points out, it is not the light, fluffy snowthat is the problem, but the wet snow followed by cold early in the season, that can leave the PV covered until spring.  I know, I have seen it!

The whole point of this is to design a system, that while being left unattended on a remote island location, can off set summer loads with fall/winter/spring harvest.  To give up 4-6 months of harvest waiting for the sun and gravity to do the job for “free”, would make the project much less feasible.

Tony

And that is why I asked these very important questions - they need to be answered, or at least addressed

I think, 6 months is quite the exaggeration, given ...
a) Daytime avg High temps don't fall below freezing until Dec 1st and  then
b) Daytime avg High temps go above freezing by Mar 1st

That may be your environment, but we see snow in mid September, often have daytime highs only in the mid 30sF in October.  Snow does not start melting in southern exposed areas here until after the solstice, and then we routinely have snow into May
Can some years be colder & snowier ?
Yes 
But then other years will be warmer and less snowy
These are the long term averages at International Falls, MN, regardless of what you saw one winter.

I’ve live here (not I Falls, but 150 km east for over 60 years, and have a pretty good almanac 

Yes, the density of snow varies ( infinity ) from about 20:1, to about 5:1 = snow depth vs melted water depth.
20:1 snow is light and fluffy - translucent
5:1 snow is heavy and thick - opaque
And there is every density in-between

I found that it takes about 100 watt-hours to melt 1kg of snow

That assumes you are melting ALL the snow.  In fact, melting a small are of the PV, exposing it to the sun will result in fairly rapid melting from the sun heating the panel, and the liquid water flowing under the snow, releasing the snow load.  If the snow is NEVER allowed to accumulate, then there is that much energy required?

assuming the 300 watt pv panel is about 1.5 m^2 area and ( 1" of snow ) is 2.54 cm thick ...
 0.04 m^3 volume of snow = 1.5 m^2 x 0.0254 m

assuming very dense snow at 5:1 ...
 0.008 m^3 of water = 0.04 m^3 snow x 1/5 density

water is 1000 kg/m^3 ...
8 kg = 0.008 m^3 x 1000 kg/m^3

If it takes about 100 watt-hours to melt 1kg of snow, so then 800 watt-hours* of electricity to melt 1" of heavy snow
It may be possible to melt the lowest layer of snow and have the slab of snow slide off, or not?
Efficiency will be less than 100%, especially when it is very cold and/or very windy.

Worst Case:
If it snows 1" every other day, for 104 days ( or 3.5 months ) then ..
You use 800 watt-hours to melt the 1" of snow
You collect 800 watt-hours the next day
NET - you have not collected any EXTRA energy for those 3.5 months ?

Best Case:
Many inches of snow fall in a day, or two, or three and then a week of clear skies?


At the minimum, this provides a starting point ...

Note:
* = assuming all of my math is correct

I think your thinking is basically on track, and your math is probably right.  That said, I think you are too pessimistic.  The only way I am going to know is to do a beta test.  What I do know for a fact, is that the panels are likely to be fully covered in snow by ~November 1, not to be exposed again until ~April 1, that’s 5 months of zero production.

I am not arguing with your logic, but simply questioning your assumptions.

Thanks,

Tony

icarus

Estragon said:

Rainy Lake would be pretty much identical climate as mine (Lake of the Woods).

Last year was a fairly bad one for slush on the ice.  I barely made it in Jan to clear panels, and aborted a planned trip in Mar.  It was a wet fall and froze up with high water.  LWCB let it drop over the winter, which avoided spring problems, but made lots of cracks over winter as the water dropped.

Maybe if they bought me an ice boat, we could talk about me clearing their panels 

Ice boating for us would be a problem...too much snow on the ice.  Our problem with the lake ice, is slush.  Where I live, we are 25km off a plowed road, then 15k down the lake by snowmobile.  The slush on the lake can be deadly unless you have a packed trail.  I can’t get to my island unless I am very careful if I have left for a month or so.

On the site on Rainy, they do have a plowed ice road, so getting there is piece of cake.  Tony

Estragon

When I aborted last spring, I was actually on an ice road with a tracked SxS.  There was maybe an inch of ice over several inches of water/slush, so kept breaking through the ice layer into water, climb back onto ice, break through again, etc.  Had already spent 1.5hrs digging out after getting stuck in slush before hitting the ice road, and still had 2-3 miles of untracked lake to go, so gave up.

With a plowed ice road, I assume there are permanent neighbours?  30kw of panels might make a tempting target with easy truck access 

mvas

Snow Loss per National Renewable Energy Labs ( NREL ) ...

Location = International Falls, MN
PV Tilt = 60 Degrees
------------------------------------------------
Jan = 86%
Feb = 51%
Mar = 01%
Apr = 07%
May = 00%
Jun = 00%
Jul = 00%
Aug = 00%
Sep = 00%
Oct = 00%
Nov = 18%
Dec = 82%
=========
Total = 245% or 2 1/2 months


This chart shows that Snow Loss is about 2 1/2 months of production - mainly Dec & Jan.
I cannot find any data that supports ... "4 - 6 months avg of Snow Loss" 

icarus

mvas said:

Snow Loss per National Renewable Energy Labs ( NREL ) ...

Location = International Falls, MN
PV Tilt = 60 Degrees
------------------------------------------------
Jan = 86%
Feb = 51%
Mar = 01%
Apr = 07%
May = 00%
Jun = 00%
Jul = 00%
Aug = 00%
Sep = 00%
Oct = 00%
Nov = 18%
Dec = 82%
=========
Total = 245% or 2 1/2 months

  
This chart shows that Snow Loss is about 2 1/2 months of production - mainly Dec & Jan.
I cannot find any data that supports ... "4 - 6 months avg of Snow Loss" 

I think your data point is about right, for a 60degree tilt.  I am arguing (or agreeing with  Estrogon) that our real world experience is actually a bit worse than your data, mostly because of icing.  I can state with assurance from years of experience that my south facing roof mounted PV will be buried under 4’ of snow in March if I have been away since December as I often am these days.

As I said, a beta is in order, I need to organize it however.

Tony

icarus

Estragon said:

When I aborted last spring, I was actually on an ice road with a tracked SxS.  There was maybe an inch of ice over several inches of water/slush, so kept breaking through the ice layer into water, climb back onto ice, break through again, etc.  Had already spent 1.5hrs digging out after getting stuck in slush before hitting the ice road, and still had 2-3 miles of untracked lake to go, so gave up.

With a plowed ice road, I assume there are permanent neighbours?  30kw of panels might make a tempting target with easy truck access  . 

Slush is the killer!  Most folks have no concept of lake ice slush.  Even the most powerful snowmachines are not up to plowing through.  I never go any where without snowshoes and provisions to spend the night when there is slush...and there almost always is, and one can never really tell where it is, and how think it is.  I have seen 6” of snow on top of 12” or more of slush, with water on top of the blue ice.

My solution is...an airboat!

Tony

mvas

icarus said:

I think your data point is about right, for a 60degree tilt.  I am arguing (or agreeing with  Estrogon) that our real world experience is actually a bit worse than your data, mostly because of icing.  I can state with assurance from years of experience that my south facing roof mounted PV will be buried under 4’ of snow in March if I have been away since December as I often am these days.

As I said, a beta is in order, I need to organize it however.

Tony

How high will the bank of snow be on the south-edge of your solar array?
I get 50" of snow per year ( Nov - Mar ), but never 50" deep, except in front of the solar array
So, I clear a path with my 4' wide PTO-driven Snow Blower attached to the front of my John Deere 4x4 tractor.
If I did not remove that bank of snow, that slid off and piled up, then my PV panels would be buried, too.
You got to have some place for all of that snow, that is piled up on top of the PV Panels, to go ...

icarus

Don’t know.  At this point, I am not sure which building(s) any array will be on.  Drifting on certain facets, areas certainly exceed 4 feet regularly. My personal wall mount panels are 6’ off the ground and they are always clear, but the other side of the house can be drifted up to the windows.

Tony

Estragon

Before I got the SxS, I would snowshoe out pulling a small plastic sled.  One January there was slush and quite cold.  With each step in the slush, some would freeze to the snowshoe, boot, and sled.  By the time I got to the cabin, it was like dragging a grand piano with a couple of cinderblocks tied to my feet.  My son literally puked from exhaustion.

My arrays face a bit west of south because of terrain to the east.  The worst drifting generally comes with northwest wind and drifting in the lee of obstructions (southeast), so the face of my arrays tend not to be prone to much drifting.  If they had to face SE instead, drifting would likely be worse.

My main arrays are on a flat structure ~1' from the south edge.  Closer to the edge would limit drifting even better, but it's ~15' high in spots with not great ladder base.  My little vertical winter float arrays are on a SW facing hill ~2-3' off the ground and (so far) haven't drifted in.

mvas

Framed vs Frameless ?
https://www.energy.gov/eere/articles/let-it-snow-how-solar-panels-can-thrive-winter-weather

framed = snow covered
frameless = clear 

Also, will monocrystalline / black cells absorb more heat energy faster than polycrystalline / blue?

mvas

If your array is two or more rows high then ...
add extra space between the rows to allow the snow to fall through the slot,
instead of sliding down on to the lower row.

mvas

There are plenty of electronic snow sensors that can automatically activate a heat strip.
But where to place the heat strip and what wattage per any given ambient temperature & wind speed combination?
Heating only along the top, in cold weather, will cause an ice dam along the bottom.

A different route ...
Sense the snow / ice along the bottom edge of the PV Panel and only then activate the heater until clear.

icarus

My intuition is a heat band a few inches up from the bottom, and another just above midline...then let the sun work.

Tony

mvas

Snow melting options ...

a) Always melt the snow flakes as they fall, in real-time, day or night -OR-
b) During the day, melt the snow as it falls but let the snow accumulate over night = too cold -OR-
c) Always, wait for the snow to stop falling and then energize the heater only during daytime, turn off when clear

Implementation of the 3 strategies varies from very easy to more difficult
The amount of heat energy required will vary per strategy.

Melting all of the snow, as it falls, uses the most energy but is simple & reliable to implement
Getting the whole slab to slide off, uses the least energy, but is more complicated to implement

garynappi

icarus said:

>>SNIP<<

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.  >>SNIP<<

I actually think it will be a problem and not a subtle one. The variables for an unattended defrosting / deicing / snow melt are pretty substantial. 

There's light, temperature, time of day and presence of snow or ice to consider. Light goes away as a function of time and cloud passage, temperature is variable all day with or without snow on a PV. For example, on a clear night during winter it would naturally be dark and cold without any snow on the panel. Using the controller inputs and outputs for data collection would be (IMO) nearly useless.

A possible solution might be...If you had two separate light sensors one on the face of the panel, the second under the panel, you "could" develop an "offset" between the two based on the time of day and come up with a normalized figure and an algorithm to determine if the panel is receiving the light it "should" be seeing. Then if the offset is too wide, the heating mechanism could be triggered which by itself will be an issue because even if it were "clear" it would present an efficiency issue year round. Heat tape if applied to the metal perimeter of the panels may not be sufficient to remove all of the snow or ice from the PV.

In my hypothetical example above, if both light sensors had zero offset like they would at night, the deicing / snow removal would not be possible until the daytime. 

Also positioning the active / passive electronics as well as powering them would need addressing. 

I think the camera solution as suggested above and some coding, may be possible but how to de-ice the camera lens or housing it's in becomes an issue :-) 

I think the presence of snow is the big issue and overall whatever de-icing / snow melt mechanism deployed might be best run on a simple timer, run for x hours with a frequency every x hours regardless whether or not the panels are snow or ice covered? Simple :-)