Tweet
I have looked forever without success. I'm looking for a calculator to determine the cost/benefit of going all flat or minimal tilt so that more panels can be installed in the same space vs optimizing the tilt for annual production. We have a lot of flat roofs and I don't see anyone addressing this specific issue. The roof angle compared to the south should also be addressed. If the panels are flat, then that their orientation toward the sun doesn't manner, but if I use 10-15% tilt, what is the % loss at say 20% off from due south? I have a suspicion that the most optimum is to ignore the southerly orientation and use 10% tilt with almost no spacing, or a single rack of 10-15% spacing, and have the last row at the optimum angle, maybe optimum for winter.
-
-
For variations in hourly, weekly, monthly or annual array output as f(panel or array orientation), PVWatts is your friend, but it won't help much with shading/interferences.
For that portion of what I believe you may be asking about, look into a model called "SAM" (PVWatts' big brother on steroids).
If you are referring to optimum row spacing vs. the tilt angle of tilted rows on a (mostly) horizontal flat roof, a general rule is that, because of shading interference of rows farther from the equator by rows closer to the equator, to a pretty good first approximation, tilting rows in a sawtooth arrangement on a (mostly) horizontal roof will gain little in annual maximum output over an array oriented parallel to the roof. That is, again to a 1st approximation, the maximum annual output of an array on a (mostly) horizontal roof - or any flat, horizontal surface for that matter is pretty close to the annual output of an array whose orientation is parallel to that surface.
Put another way, tilting rows of panels on a limited area (mostly) horizontal surface such as a roof won't result in more annual output from that limited (mostly) horizontal surface than if that limited (mostly) horizontal surface were covered with PV or other solar panels in a fashion that is parallel to that (mostly) horizontal surface.
Shading of adjacent rows is the culprit.
However, tilting rows of arrays can reduce the array surface area and so reduce the cost of the panels at the expense of more and perhaps beefier racking. Doing so may result in a lower LCOE (Levelized Cost Of Energy).
So, adjusting panel tilt and row pitch on a (mostly) horizontal roof in a fashion that maximizes energy collection per square area of solar collector can minimize array area, but the available annual energy collection is pretty much limited to something close to the output of a maximized area of an array that's parallel to the roof.
The cost benefit of all that is a problem in solar process economics. -
Yes if you can afford the extra panels, just covering a limited
flat space will collect at least as much energy as panels set
for a better angle. This due to eliminatng shading issues as
the sun moves. But all flat can have anothe benefit for some
areas. It will collect more energy under clouds. I am constantly
reminded of this, as I see my panels supplying the inverters in
the 50% to 100% capacity range under some clouds. Bruce RoeComment
-
But it's a bad idea for other reasons, a lot of which have to do with array/panel fouling. Horizontal, or low tilt angle orientations quickly turn panels into mud caked evaporation ponds.
As for collecting energy under clouds, well, the amount of energy collected under clouds by an array in a solar energy viable climate is pretty small. Even less so if the climate is somewhat cloudy with a seasonal or annual clearness index of about 0.50 or less.Comment
-
Instead of "all flat" how about 10 degree alternating E/W tilt per row? No southerly tilt. That eliminates that fouling issue and has the same production as a non-fouled all flat system. I can put maybe 28 panels on the flat part of my roof like this, but only 20-22 if they are tilted at the optimum angle. Again, it would seem best to have the last row at either optimum tilt for the year or either summer or winter. I used PVwatts and there was very little change in production by keeping the panels at 200 degrees vs 180 degrees (which is due south in PVwatts). However, with no southerly tilt, it wouldn't matter anyway. The maximum loss by having 0 tilt vs an optimum annual tilt was -16% (0 vs 37 degrees tilt). Furthermore, the loss in roof space from shading is about 15% for every 5 degrees of tilt, so it clearly isn't worth saving the extra 16% in system cost to tilt the panels at an optimum angle unless there is over capacity of roof space. Sound right?Comment
-
How are you calculating the tilt angle vs. row pitch ? The two are related in that the greater the tilt angle, the greater the row pitch to minimize shading (or maximize panel exposure - same thing, but it's non linear and also somewhat dependent on cloud cover variables and usual diurnal cloud cover patterns as well as panel azimuth and a few other variables.
Most of the array orientation and other design decisions are not dichotomous, but also not linear with a lot of variables winding up dictating a lot of tradeoffs in design. For example, Low tilts will foul quicker than high tilts, but less in climates with regular precipitation.
As for E-W orientations, and aside from whether or not an array is continuous or sawtooth, to maximize the specific output of an array (that is, the most output per square area of panel, in this case on an annual basis) one orientation for the whole array rather than multiple orientations will maximize that number. Design considerations and space limitations may dictate multiple orientations, but in and of itself, multiple array orientations will always decrease the specific annual output of a specified total array area. There may be other reasons for doing so, but to split an array orientation with the idea of getting more output from the same area of panels is a mistake. That means split orientations for the same total area will almost always result in a greater required array area for the same annual output. In a roundabout way, that's a consequence entropy, but that's off topic.
However, and as you note, at low tilts that orientation difference won't matter too much - but know that the array will need more frequent cleaning to maintain the same performance and an array with a higher panel tilt angle. At higher tilts, it'll be a PITA to figure out shading and will also start to take a greater toll on specific output as the higher the tilt angle, the greater the specific output penalty for split orientations.
One advantage of sawtooth arrays on (mostly) horizontal surfaces is that the required row spacing leaves room for access to all the panels. Often and usually single orientation arrays on roofs are arranged with no thought for access to all the panels for service and packed in cheek by jowl. In the days when string inverters were just about ubiquitous, that limited access was much less of a problem. Now, with micros and optimizers being the preference of installers and with them their order of magnitude increase in failure probability, access becomes more of a priority.
In theory at least, and to a pretty good long term average first approximation, the azimuth and tilt of an array or any panel that maximizes annual output (but maybe and probably not annual array revenue or worth of that output under an NEM arrangement, particularly if it's a T.O.U. rate schedule) is probably not dead nuts south (but is indeed one orientation).
With that in mind, if it was me and I was putting an array on a (mostly) horizontal and (mostly) flat roof, and assuming the roof is otherwise unshaded by other structures/trees etc., I'd first use PVWatts to find the ideal array orientation - tilt and azimuth - that maximizes annual specific output (that is, maximum annual output per square area of array), use that tilt and azimuth and place the array on the roof. Then, I'd design for the row pitch that minimizes the shading on adjacent arrays - that pitch being the one that maximizes the total annual array output while keeping the chosen tilt and azimuth constant. The best tool I've used for that purpose is one of the features of a model called "SAM" which is PVWatts big brother on steroids. Not for the feint of heart but useful if one takes the time to learn how to use it.
Then, because it's all about most bang for the buck with me, I'd do a life cycle cost analysis on the proposed project and iterate things until I got the lowest LCOE (Levelized Cost Of Energy) of the project and compare that to the LCOE of not doing the project at all. Then, compare the two alternatives and choose the one with the lowest LCOE. But that's just me.
BTW, that here is one array orientation that maximizes annual specific output or maximizes output for a given array area is the reason and proof that multiple array orientations require more surface area to produce the same output as one orientation of the same total surface area. There may be reasons for multiple orientations but maximizing array output per square area and so probably project cost effectiveness ain't one of them.Comment
-
Other factors of near flat are for each owner to evaluate. InOriginally posted by fraser
9 years here panels have never been washed, though 3 had
to have pine sap SCRAPPED off.
With 40% more panels, you can expect 40% more energy
output under clouds. Is that energy pretty small? Today under
pretty heavy clouds I saw 4.1KW from a 15KW inverter plant.
That is actually a lot more than it sounds like, BECAUSE it comes
all day long, not just during several hours near solar noon.
Another day I saw 10KW under less dense clouds, still obscured
the sun. Bruce Roe
Comment
-
Bruce: If you would indulge me, I forgot, how many STC kW of panels do you have installer - not the inverter sizes - the array's STC rating ?
Also, while I believe I understand but don't totally agree with your logic of multiple array orientations, from what I've learned over the last 45 + years or so, I don't think it's possible to get a semi steady rate "all day long" - or at least while the sun is up - out of an array when it's cloudy. If you get close to the same rate at, say 0800 solar time as you get at solar noon, that rate will be so small as to not be worth talking about anyway. Reason: a larger than usuyal portion of the irradiance will be blocked by heavier than usual overcast. Also, and under less heavy but still overcast skies, the irradiance, while much reduced, still has some directional qualities to it, meaning that more of the diffuse irradiance reaching the ground will still be coming from the direction of the sun, meaning that split arrays will gather less energy under cloudy or less than clear conditions than split orientation arrays. At least that's the way I learned it and also the way I've confirmed it by measurement.
I'd also be interested to read what your specific output (kWh/day per installed STC kW of panels) - not inverter capacity - for a totally cloudy day in the winter and for a totally cloudy day in the summer.
Lastly, with 40% more STC wattage - not strictly 40% more panels - in the same orientation as before the increase in STC wattage, you can most likely expect a 40% increase in output under all atmospheric conditions be they clear, totally overcast and every condition in between. In the same location and orientation larger arrays perform no better in terms of output per installed STC kW than smaller arrays.
Respectfully,Comment
-
J.P.M. Thanks for the input! So, I used PVwatts, inputting different tilt angles and azimuths to find the losses for annual productions. As I said, the optimum tilt - around 37degrees, only produced 16% more annual production vs a 0 degree array. But every 5 degree tilt, 15% more space is required between the rows, so the breakeven is at about 5% tilt, which isn't enough to prevent fouling. The point is optimizing for solar panel production is not the goal, nor should it be, unless you are not limited for space for the amount of annual power needed, which applies to most residential applications. So, with space being the constraining factor, not solar panel production efficiency, then, a flatter system seems to output more annually, simply because more panels can be installed. I remeasured and I can get up to 30 panels on my roof with a 10degree E/W tilt (or 15, doesn't matter much at these low angles). Just need enough angle to minimize fouling, as you indicated. However, I am not sure I understand the reasoning that a single orientation produces more for a given space. For example, I can have 3 rows of near flat panels, but can't put in a 4th row. However, I can put in a highly tilted row (toward the south) at the very north part of the array because of it's lower footprint, which wouldn't cause any shading on the flat portion AND would actually marginally increase the flat panel production from reflection (assuming 90% light absorption and 10% reflection). I honestly can't see how this wouldn't be optimal on almost every roof except for those that can 'just' fit in the last row without tilting, leaving no space for a final titled row on the north end (south end in southern hemi). I do see your point on leaving enough access spacing between every 2 rows. But, respectfully, with the cost of solar installations dropping and the power of the panels increasing, I think "best bang for the buck" is not the best way to optimize solar installations anymore. Instead, overpowering the system and eliminating all electricity costs is cheaper overall, especially if it permits additional energy consumption like EVs. I even found that to be cheaper than making energy efficiency home improvements, much to the power company's chagrin.Comment
-
That peaky curve of array energy output peaking at solar noon, is really
the composite of (at least) 2 curves, the dominant one being the relative
alignment of the array to sunshine. The other is the amount of atmosphere
the sun must penetrate to get to the earth, I believe this is much lessor
variation over the day, except at the extremes.
The clouded output largely is not subject to the first, the second is broader.
I am not putting out any hard numbers, difficult to find those calibrated
clouds. But I am often amazed at just how much energy is produced when
I cannot even see the sun, compared to the initial generic installation.
Running a 15KW inverter plant, with a DC:AC ratio of 2.333:1. If you take
into account 95% conversion efficiency, it is effectively 2.22:1. Panel
aging... Bruce RoeComment
-
Yes. BruceOriginally posted by J.P.M.Comment
Comment