Does this site survey justify pointing the array way away from south?

[williaty]

We're at the point of actually needing to start setting anchors into the ground for our solar install here at our house. Unfortunately, I did the initial site survey mid-winter, guessing about the obstruction from vegetation and also guessing about where I'd actually be able to put anchors into the ground for the panel mounts. OUPS came out and marked the underground utilities today and my preferred site for the array is out of the question. In trying to find an alternative, I also realized I'd really missed a lot of obstructions in the winter due to lack of leaves. So given that I'm really constrained by where I can put anchors in the ground, there's one site where just looking around I could see a lot more sky than anywhere else. When I did a new survey (using Spyglass on iOS), I realized that due south is basically clobbered by 2 large cherry trees that are in a place I can't access to cut down. I've attached a chart that has the path of the sun through the sky each month vs the treeline. Given that the East sky isn't great, and the South sky is completely clobbered, it seems like the array needs to face southwest (roughly 225*) to maximize the clear sky available from the only location that's even slightly viable. However, "ignoring" half the day just feels really wrong if I'm trying to catch sun! What do you guys think of this?


solar_insolation_sun_path_chart-mod copy.jpg

[bcroe]

With those difficulties, a ballasted array might allow you to put it in
the best place without any digging. In summer the sun will be close
to straight up, rising and setting somewhat to the north. Winter may
be a big loss, but it already is anyway. good luck, Bruce Roe

[solarix]

Optimum orientation depends on what your utility rate structure is. If you have a time of use plan where the afternoon "on-peak" rates are high - then aiming your array westward to some degree is best. The array may produce somewhat less kWhs but save you more dollars.

[williaty]

Oh, this is a battery system, no grid tie at all. I'm trying to see how much of my house I can switch to solar over time, as well as ensuring the freezers stay cold while the utility electric is out.

[bob-n]

Don't forget that trees grow 1' to 2' per year, so those obstructions will get worse with time.

[J.P.M.]

Run PVWatts for your location and at any considered orientations.

Use the hourly output option.

Deduct hourly outputs for times/hours when the array will be shaded.

1- (what's left divided by what PVWatts started with) will probably be a rough 1st approx. of % annual output penalty from shading.

Depending on the type of shade, solid as from a wall or partial as from a deciduous tree in the winter, you'll have after the math is what's probably a somewhat conservative (that is, somewhat pessimistic) SWAG of what you might achieve in terms of annual output.

Just remember that tree shade, even from leafless trees can and usually will extract a larger penalty on output than the simple % of shade it produces on an array.

BTW, back in the day, I had some small contribution(s) to the graph you include. Long story.

[peakbagger]

Solar modeling does not include all local site conditions. As an example I live on a ridge adjacent to a large river valley. I get sun far earlier in the AM as its far rarer to have a morning fog bank compared to folks in the valley. I also tend to get more afternoon cloud buildup at my location. So I tend to go a bit farther south then standard modeling may indicate.

[J.P.M.]

Solar modeling does not include all local site conditions. As an example I live on a ridge adjacent to a large river valley. I get sun far earlier in the AM as its far rarer to have a morning fog bank compared to folks in the valley. I also tend to get more afternoon cloud buildup at my location. So I tend to go a bit farther south then standard modeling may indicate.

It sure doesn't, but if the OP has a SWAG of shade potential/penalty from a shade model/tool, laying that over a solar model's unshaded output might be a useful start to getting a rough shade estimate which, depending on how "rough" is defined/used, might be a decent start to an analysis.