What is the best way to overcome the problem of shade from a chimney?

[FB.]

It'd be interesting to know others' thoughts.

Is it viable to have all the panels arranged in landscape as close to the ridge and/or as far to the right as possible? I reckon the roof is wide enough to take rows of six along the ridge and at least four per row on the unshaded right. The sun would have to be very low in the sky to affect those on the ridgeline.

 

[SolarCity]

Is it viable to have all the panels arranged in landscape as close to the ridge and/or as far to the right as possible? I reckon the roof is wide enough to take rows of six along the ridge and at least four per row on the unshaded right. The sun would have to be very low in the sky to affect those on the ridgeline.

Can only get 5 in a row in landscape and the layout receives more shading this way.

 

[FB.]

I'd stick with that layout then.

I would worry about an "upper" and a "lower" string because in autumn, winter and spring that chimney is going to play games with the leftmost five panels on the upper row.

If we number the panels 1-15, starting at the top left as #1, the top right as #9 and with #10 being the one on the bottom row closest the chimney, I think I'd have the eight on the right (6,7,8,9,12,13,14,15) on one MPPT and the seven on the left (1,2,3,4,5,11,12) on another.
I appreciate that #5 has less impairment than #12 and you might want the least-impaired panels on one string, but I think in the afternoons #5 will be more affected than #12 as the sun shines between the chimney and the neighbour's roof (your first image shows what I mean).

The second string is going to be a bit of a mess and rarely reach its full potential, but a scan function for optimum voltage should be able to keep it running close to three-quarters potential on sunny days (and no noticeable losses on the many cloudy days).

 

[SolarCity]

Actually, I think you'd be surprised. Which is why PV Sol is such an excellent piece of software.

You can try different string set ups and it will recommend a best option. No guess work. I've been trying to post a video to show you what I mean.

 

[Gavin A]

As I said before - Power-One probably split into 8 and 7, set the full MPPT scan function to 5 minutes, and maybe adjust the low voltage setting down to allow for lower string voltages on the shaded side when shading has knocked out a few of the strings within the panels.

string systems underperform badly in shaded conditions when the installer doesn't know what they're doing, and either spec an unsuitable inverter, or don't set it up properly.

As your numbers indicate, solar edge would likely outperform the power-one, but not really by enough to justify the extra expense in this situation IMO.

 

[FB.]

Actually, I think you'd be surprised. Which is why PV Sol is such an excellent piece of software.

You can try different string set ups and it will recommend a best option. No guess work. I've been trying to post a video to show you what I mean.

My concern - and reason for suggesting a left/right split - is there are two different causes of shading, and the shading is at different times of day.
The left-most panels are affected by the chimney in the afternoons and evenings at any time of year.
The lower row of panels are affected by the neighbour's roof around mid-day in winter.

I worry that mixing left and right might end up with a complex mix of shading almost all day. But by splitting off the right-most eight panels they benefit from no shade at any time of day during spring, summer and autumn (when about 90% of annual generation occurs).

But as already said by others: two strings and getting the inverter to scan the voltage should keep losses to manageable levels. However you configure it I would expect it to meet the SAP-2009 estimate for a 3.7-4kWp array that isn't shaded.

 

[SolarCity]

This video shows the shade cast on the roof over a year. It's a long watch but an interesting one.

[video=youtube;49dUf4sYIus]http://www.youtube.com/watch?v=49dUf4sYIus[/video]

It's worth nothing how the shade cast by the chimney affects the array at different times of the year. Shade cast in the winter if far less of a worry than shade cast in summer and you'll notice that the chimney tends to affect the lower row of panels only in the summer months.

 

[telectrix]

What is the best way to overcome the problem of shade from a chimney?


Knock it down???

come on geordie. knock it down is a plumber's solution. as electricians we would carefully demolish it brick by brick, stacking said bricks neatly in the garden for the client to make a water feature from it, being careful not to shade his PV array.

 

[compmad1]

Any chance of some pictures? (maybe screenshot captures from google maps)

The shading might not be as bad as you fear and there are lots of clever tricks which could be used to get the best from the panels even if they suffer hard shadow at certain times of day.

Thanks. I did upload a picture from Google Earth, but being new to the forum, I don't think I am allowed to at the moment. Solar City have put up some good pictures with the PV Sol Expert software. Thanks for all the advice you are giving.

 

[FB.]

Ah, it faces SSE which makes the neighbour's roof shadow on the lower panels worse than expected.

I still wonder if landscape layout (two rows of five or six in landscape close to the apex) - or at least exploring its possibilities with the software - would help the upper row's bypass diodes gradually adjust the panel output so that maybe even without voltage scanning the loss of just one-third of a panel could be sufficiently close to the optimum voltage that the inverter would find it or not be far from it even if tracking a false peak.
The chimney knocking out one-third of a panel in a string of eight is only going to change the voltage going through that string by 10V (out of 240V) which won't be far from the optimum of the curve.

 

[SolarCity]

No, it faces just 10 degrees off south towards west.

 

[FB.]

No, it faces just 10 degrees off south towards west.

Ah, you set the view from the West, rather than the convention for maps which is South!

Looking at the video I wonder whether the changing shade conditions (and the angle from which the shade comes) are sufficiently gradual that a single string of fifteen panels might work surprisingly well because the panels will be affected one-by-one and by only 2% as each bypass section drops out (giving time for the MPPT to adjust).

In summary: I think the shading looks worse than it actually will be in the real world. It will be a nuisance but probably not enough to affect the array's ability to reach its SAP target except in really cloudy summers when we all will struggle! I think sensible design and a quality inverter would deal with most of the problems. I wouldn't be over-keen on micro-inverters because of the hassle if they go wrong (especially if the company behind the warranty is no longer trading at some point in the future).

 

[SolarCity]

Yes, I suppose that is the beauty of the software. No guesswork - just accurate models.

 

[SPARTYKUS]

Not a solar chap but really enjoyed this thread, faskin ating.

 

[Worcester]

@SolarCity, just watched the video..
I would normally have chosen a string inverter (Power One or SMA both with their shading options switched on) configured with the heaviest shaded panels in a single string, configured similar to how FB suggested.
However this summer we did some playing around with string inverters and Solar edge and noticed that at certain times of day the string inverters where up to 30% down on power compared to the SE when we had shading similar to that from the neighbour's roof, i.e only covered a part of the bottom of the panel. - We were actually staggered at the result and so double checked it.

SE is a far better solution than micro-inverters in most cases - See my comments in a previous string about start up voltages on micro-inverters. However you still have to weigh up the extra cost of the Solar Edge over the potentially improved output..