Planning new install - 5 degree metal roof aimed north, need advice

[bcroe]

of course bird poo doesn't wash off by itself.

It does here before long, in northern IL. Bruce Roe

[solarix]

Anything over 15 degrees stays quite clean in our area. Not sure exactly what angle actually is. 15 degrees is where I draw the line. Another factor is if you have anything sticky that blows around. Some pollen can be a problem and then of course bird poo doesn't wash of by itself.

[J.P.M.]

@Solarix mentioned that angles below 15 degrees from horizontal do not self clean from rainfall. I know you have done some analysis on the affect of dirt on panels. Care to speculate on where the sweet spot may be between tilt angles and panel soiling. I understand that different environments may have different soiling characteristics.

This is a theoretical question since I may have missed an opportunity when I had to remount some panels due to a reroof that my HOA did two years ago. They were originally installed flat because my HOA did not want them seen from the facility. I later realized that that kind of restriction was unenforceable but I was also limited on roof area. When I reinstalled them I considered giving them some small amount of slope but in my haste I did not bother with that detail. It is now a rental and the tenant pays for electricity so it is not a big financial issue for me.

Observations and measurements I've done have led me to several theories on panel fouling mechanisms that have led to further observations and measurements and then modified measurements.

But right now, big cooking/baking day for me. Got a roast to cook, a bunch of baking and some ice cream to make.

More on array fouling vs. tilt angle later if you or anyone else responds/wants more mind spoor from me.

Short/anecdotal comment til then: Horizontal panels turn into a sand box/dried mud pit sooner rather than later. Vertical panels stay about as clean as kitchen windows.

With respect to fouling rate vs. tilt angle only and maybe a bit distinct from of a general discussion about fouling rates in general: A bit less anecdotal but still speculative and in any case incomplete preliminary set of observations based on info/readings and observations of cooperative neighbors, what I've read in the journals, and what I think I might know about material transport mechanisms such as, for example, stream bed loading, angles of repose and other esoteric and seemingly unrelated phenomena: Between the two extremes, horizontal and vertical, the rate of fouling for an array might at least partially be some function of the cosine of the array tilt angle raised to some power < 1.0, but there may well be and probably are a dozen or more other factors that come into play, they may or may not be independent of one another, and some may be absent or greater in effect as f(site location, season, other).

[snowmonster]

I am also curious of the soiling issue. We do get some reasonably hard rains from time to time that I would think would help.

[Ampster]

..............
FWIW, my numbers for annual output pretty much agree with what you report. I'd note those #'s seem in line w/my comment that, to a 1st approx., the max. output of an unshaded sawtooth array on a horizontal roof is about the same as the output of a horizontal array on that roof.

@Solarix mentioned that angles below 15 degrees from horizontal do not self clean from rainfall. I know you have done some analysis on the affect of dirt on panels. Care to speculate on where the sweet spot may be between tilt angles and panel soiling. I understand that different environments may have different soiling characteristics.

This is a theoretical question since I may have missed an opportunity when I had to remount some panels due to a reroof that my HOA did two years ago. They were originally installed flat because my HOA did not want them seen from the facility. I later realized that that kind of restriction was unenforceable but I was also limited on roof area. When I reinstalled them I considered giving them some small amount of slope but in my haste I did not bother with that detail. It is now a rental and the tenant pays for electricity so it is not a big financial issue for me.

[J.P.M.]

The system is going to be a string inverter with optimizers. The Iron Ridge tilt legs will lift the panels off the roof even more so I don't think that will be a problem.

I started learning more about sun angle and azimuth today and then stumbled across this calculator - https://www.rbisolar.com/solar-calculator/ You can play around with time of day, location, panel size, etc. All of these online calculators seem to suggest a lower pitch than what I was calculating (even adjusting for my 5 degree sloped roof in the wrong direction).

As far as self shading goes, the sun angle here isn't even above 20 degrees until noon and drops below that again at 3:30. Combine that with only 2.55 hours of solar insolation on Dec 21, do I even really care about self shading at 9:00 AM or 10:00 AM on Dec 21?

Using the winter solstice solar elevation angles and hours away from solar noon will pretty much ensure no/min. self shading, but in reality because the POA irradiance on an array is less in intensity as well as frequency (cloudier days) from mid/late fall until say, mid Feb or so, decreasing the row pitch from the no shade ever scenario to one that loses some production away from solar noon during the winter months may actually result in more annual production by enabling a larger array to be placed on the roof. Using 12/21 0800 or 1600 hrs. angles is an example. There are lots of models like SAM for example that will generate a lot more detail and useful info for design than can be explained here.

After doing such things as more than a hobby but less than a job for 45+ yrs., I've learned that treating this subject in a linear mode isn't the best way for me to work on these types of things. Thinking in gradations and keeping a sinusoidal curve in mind helped me.

Any comment on wind loading considerations with respect to the design ?

I'd also respectfully suggest you think about what snow will do, where it will go once an array is in place, and the potential for damage from buildup/leaks/ice ponding and other things that will come with the array.

[snowmonster]

Tons of great advice, all of it. Thanks to JPM, solarix, Mike90250, littleharbor and Ampster for their usual great guidance.

I'm assuming that snowmonster's system would put the panels very close to the roof, so even if snow falls off the panels, it has no where to go.

Are solar panels engineered to be submerged in snow and stay submerged for a month, as might happen with a shallow angle, a bad snow storm and a few cold weeks? Is there any comparable issue to ice dams, where daytime melting followed by nighttime freeze might cause ice-stress?

Similarly, if the system uses microinverters or other electronics on the roof, is that engineered for snow submersion?

If the building heats the bottom of the snow, it might turn to pooling water, trapped by ice above. Is that an issue for panels or electronics?

The system is going to be a string inverter with optimizers. The Iron Ridge tilt legs will lift the panels off the roof even more so I don't think that will be a problem.

I started learning more about sun angle and azimuth today and then stumbled across this calculator - https://www.rbisolar.com/solar-calculator/ You can play around with time of day, location, panel size, etc. All of these online calculators seem to suggest a lower pitch than what I was calculating (even adjusting for my 5 degree sloped roof in the wrong direction).

As far as self shading goes, the sun angle here isn't even above 20 degrees until noon and drops below that again at 3:30. Combine that with only 2.55 hours of solar insolation on Dec 21, do I even really care about self shading at 9:00 AM or 10:00 AM on Dec 21?

[bob-n]

Tons of great advice, all of it. Thanks to JPM, solarix, Mike90250, littleharbor and Ampster for their usual great guidance.

I'm assuming that snowmonster's system would put the panels very close to the roof, so even if snow falls off the panels, it has no where to go.

Are solar panels engineered to be submerged in snow and stay submerged for a month, as might happen with a shallow angle, a bad snow storm and a few cold weeks? Is there any comparable issue to ice dams, where daytime melting followed by nighttime freeze might cause ice-stress?

Similarly, if the system uses microinverters or other electronics on the roof, is that engineered for snow submersion?

If the building heats the bottom of the snow, it might turn to pooling water, trapped by ice above. Is that an issue for panels or electronics?

[J.P.M.]

All great points I knew I would get from this site.

Agreed that the roof wasn't completely designed for solar as it has a 5 degree northern slope. I meant to say that structurally it can support the weight of an array.

I sat down and did some math/trig/geometry this AM and determined I can use 20 degree tilt legs (for an effective 15 degree tilt) and 5.5' row spacing with zero shadows from neighboring rows and still get 59 panels on the roof. This at least keeps the panels aimed in the right direction and reduced fouling. I know a 40 degree tilt is optimal but I would only have 2 rows of panels if that were the case. It is interesting to note that Renvu's row spacing calculator says I can get away with 27" spacing with landscape orientation but clearly there will be shading issues. It is an office building and the front of the building faces south - the appearance of panels on the front of the building isn't what we are going for.

I did mention the snow thing but we don't really get the crazy lake effect snow like you'll see in Michigan. I don't expect to have all that much production in Dec/Jan anyway so I'm not terribly worried about that.

PVwatts info:
Flat system : 24.0 kW system mounted flat on the 5 degree north facing roof = 27,567 kWh annually but fouling is certainly an issue and difficult to clean
15 degree southern tilt: 20.4 kW system with an effective 15 degree southern tilt (using 20 degree tilt legs) = 27,682 annually

It was interesting to me that there is only 384 kW difference in December between the two but I know a 15 degree tilt doesn't accomplish much.

Working on different configurations/row spacing/tilt angles now.

1.) The dead weight of most any array by itself (probably < ~ 5lbm/ft^2 of array or so) is not much of a problem for most any roof (often/usually a roof has a design dead load of ~ 40 lbm/ft^2 or more), but all required load combinations still need checking. The uplift and downward loads from wind on a well designed roof need checking and design consideration(s). Structurally, the common weak points or at least points of concern in a rooftop PV array design are the array supports, their attachment method(s) to the roof, and the resultant point loads and their effect(s) on the roof at those attachment points and on the roof in toto. If your jurisdiction requires permitting for PV, you may well run into a requirement to provide calcs/drawings to show that the design is fit for purpose with respect to those and possibly other loading conditions and combinations.

2.) If you already know this, ignore it, but using a 15 deg. off horizontal tilt on a 5 deg. north sloped roof, if you're designing for a row pitch that, for example, avoids ALL (self) array shading from the more southerly rows, between 0800 and 1600 hrs. solar time on the winter solstice, for the Renvu tool, you'll need to use a "sun angle" as Renvu defines it of about 8 degrees or so - that's your approx. solar elevation angle at 0800 or 1600 hrs. on ~ 12/21 at ~ 40.2 deg. N. lat.
That'll give a row pitch of ~ 220 " w/panels in portrait or ~ 139" in landscape. So, landscape orientation gets you a tighter row pitch but fewer panels per row in about inverse proportion to the ratio of panel L X W ratio and thus an approx. wash in max. allowable array area per roof area. I'd note that the landscape orientation may result in a lower design wind loading for about the same area.
Other stuff I've got more/less agrees w/the Renvu tool. One product/model is something called "SAM" from NREL. Given good input, it does a fair/good job on shjading, but it's a somewhat long learning curve.

3.) If you can live with some shading, using a 16 deg. "sun angle" will get you shade free from 0900 to 1500 on the winter solstice at your approx. location and doing so may not cost too much in terms of annual output, particularly if you're not counting or expecting much winter production. Looks like that row pitch drops to ~ 140" for portrait orientation and 86 " or so for landscape panel orientation. Again, the lower wind loading in landscape may be a consideration.

4.) FWIW, my numbers for annual output pretty much agree with what you report. I'd note those #'s seem in line w/my comment that, to a 1st approx., the max. output of an unshaded sawtooth array on a horizontal roof is about the same as the output of a horizontal array on that roof.

[solarix]

Some building departments will want you to prove that your standing seam roof has adequate fasteners (as in - lots more than usual) to withstand the additional uplift caused by a solar wing attached to the seams.

[snowmonster]

I thought that was what you meant by "support". One of the advantages of attaching solar to a standing seam roof is that you don't need to penetrate the roof. to attach the racking. Weight is only one design factor. Uplift is another that will increase as you increase the tilt.

That is why I chose that roof style. The uplift issue is in the works as well now that I'm looking at a 30 degree tilt relative to the roofline.

[Ampster]

.....

Agreed that the roof wasn't completely designed for solar as it has a 5 degree northern slope. I meant to say that structurally it can support the weight of an array......

I thought that was what you meant by "support". One of the advantages of attaching solar to a standing seam roof is that you don't need to penetrate the roof to attach the racking. Weight is only one design factor. Uplift is another that will increase as you increase the tilt.

[Ampster]

(Note to Ampster: This is an example of why the distinction between flat and horizontal might be useful to help avoid confusion)
....

Confusion?
There has only been one person on this forum who has been insisting that most roofs are flat. Perhaps that person is still confused. It is hoped that the following attachment, which was posted by a moderator some time ago, will clear up any confusion. I presume the red circle was added by the moderator for emphasis.

Hopefully that ends the need for further discussion. Sometimes one's beliefs do not correspond with generally held understanding of nomenclature in the construction industry.

[J.P.M.]

Would not a reverse tilt in snow country be a big snow trap?

More than a horizontal roof and among other factors, probably as some f(angle between roof slope and array slope, roof slope). A somewhat absurd example: A 45 deg. slope single row array elevated, say, 2m off a 45 deg. north tilting roof for example, but common sense hopefully intervenes at some point.

[snowmonster]

All great points I knew I would get from this site.

Agreed that the roof wasn't completely designed for solar as it has a 5 degree northern slope. I meant to say that structurally it can support the weight of an array.

I sat down and did some math/trig/geometry this AM and determined I can use 20 degree tilt legs (for an effective 15 degree tilt) and 5.5' row spacing with zero shadows from neighboring rows and still get 59 panels on the roof. This at least keeps the panels aimed in the right direction and reduced fouling. I know a 40 degree tilt is optimal but I would only have 2 rows of panels if that were the case. It is interesting to note that Renvu's row spacing calculator says I can get away with 27" spacing with landscape orientation but clearly there will be shading issues. It is an office building and the front of the building faces south - the appearance of panels on the front of the building isn't what we are going for.

I did mention the snow thing but we don't really get the crazy lake effect snow like you'll see in Michigan. I don't expect to have all that much production in Dec/Jan anyway so I'm not terribly worried about that.

PVwatts info:
Flat system : 24.0 kW system mounted flat on the 5 degree north facing roof = 27,567 kWh annually but fouling is certainly an issue and difficult to clean
15 degree southern tilt: 20.4 kW system with an effective 15 degree southern tilt (using 20 degree tilt legs) = 27,682 annually

It was interesting to me that there is only 384 kW difference in December between the two but I know a 15 degree tilt doesn't accomplish much.

Working on different configurations/row spacing/tilt angles now.