PV placement

Effects

I'm looking for a bit of advice regarding PV placement on my roof. In the attached drawing it shows a raised section of roof light with the PV panels to the left. Originally the architect had them showing wider, as can be seen in the elevation.
I mentioned my worry about shading on the panels causing a drop in efficiency. That's why he's moved the panel away from the edge of the roof light. Am I right to be worried about this? Don't modern PV panels have a bypass to avoid shaded sections of the panel? I was also worried about the angle, as it's on the flat, but the proposed installer has said it's only an efficiency drop of about 5%. Also based on another thread here it seems they are placed on the flat a lot more these days.
Any thoughts?

Effects

Woops. Fixed that broken link.

Are there some height restrictions there? If not it may be better to mount the panels on a frame that angles them to face the SSW direction for maximum exposure.

Effects

We're trying to keep a neighbour on side. We have already reduced the height of the roof light section so it's below their first floor eye line. I know they won't want to be looking out at the rear of a PV panel! I'm sure angling SSW can be done to a certain extent. I'm not sure if we can tilt them though. Overshadowing is my main concern. Visiting an installation nearby tomorrow by a recommended installer so it'll be good to ask him about everything.

quentingargan

The biggest problem with panels flat is that they are not self-cleaning. You need some sort of slope - at least 5 degrees. You can get a small bucket like device that mounts panels at a slope of 15 degrees. Not elegant, but it is hidden by the panels, and a lot cheaper than making a frame up. If you PM me I can send a link.

The loss of efficiency is pretty much as set out in the chart in table H2 of the DEAP manual.

There are bypass diodes and if you mount the panels in landsacpe (rather than portrait) this will reduce the effect of shading by a parapet. The strings of cells are laid on so that the panel is usually split into three sections. You may lose one section, but if you had shading in portrait format you would lose al three sections.

freddyuk

As I see it you are going to get shading from the roof light at all times of the year and in winter possibly all day!
I would place the panels at either end of the roof with clear sky. I would get them as high as possible.
Any shading will kill the output. True it is better to shade the "long" side never the short side but having them horizontal and shaded is a complete no no if you have options. The inbuilt diodes prevent hotspots and potential failure when shading is a problem they do not actually improve the output at all.
What is wrong with the rest of the roof space?

Maybe you could mount them on a horizontal frame that is level with the top of the skylights, at least that would solve the issue of "hiding in your own shadow", but what about other shading from other structures.

quentingargan

Oh - just spotted the compass rose. How about on a 15 degree mounting facing ESE, up against the roof light? The ideal would be on the SSW at an angle between the parapet and the roof light as suggested by Dolan Baker.

Are you still at the design stage?, if so another suggestion would be to swap the positions of the solar panels and the skylights, this would allow you to place the high end of the panels level with the top of the skylights and the other end lower, giving you a slight angle.

Effects

Effects wrote: »

I mentioned my worry about shading on the panels causing a drop in efficiency.

It's not a drop in efficiency it's a drop in output. And it's very high as the lit cells power the shaded cells. Output decrease is almost exponential to the shaded portion.

Effects wrote: »

Don't modern PV panels have a bypass to avoid shaded sections of the panel?

Yes but they only bypass strings of cells so if the shadow doesn't line up to the cell layout they don't do a whole lot. Repeated shading will cause damage to cells over time if it's always in the same place.

Effects wrote: »

I was also worried about the angle, as it's on the flat, but the proposed installer has said it's only an efficiency drop of about 5%.

If an installer said that to me I'd fire him on the spot. You can expect a 20% decrease in daily harvest in the Summer and >300% decrease in the Winter and a marginal gain in cloudy weather and longer exposure, I've seen it oversimplified to a 40% annual reduction.

quentingargan

Sir Liamalot wrote: »

If an installer said that to me I'd fire him on the spot. You can expect a 20% decrease in daily harvest in the Summer and >300% decrease in the Winter and a marginal gain in cloudy weather and longer exposure, I've seen it oversimplified to a 40% annual reduction.

Instinctively I would have agreed with you, but Table H2 from SEAI suggests 11% overall loss from having horizontal -vs- 30 degrees facing south over the year. I must run a PVSyst simulation on that sometime and see what the real figure is...

You're right about the shading though.

I'm not swayed by SEAI either Q. Where are they getting their numbers? If they have two side by side arrays one flat and one optimised on the same lekytronics and cable I'll eat my hat (and upload a vid. ) 30° sounds shallow to me too.
I have a module on manual tilts with batteries tied through a shunted watt-meter I'm well aware of the real figures. Obviously cloud and exposure will skew the longterm findings but I'm entirely sceptical it's a mere 11% decrease. {Edit: now on the eating of hat I have to admit that come summer solstice power is abundant so I 'd not bother angling it, the gains are of the practical kind the rest of the year which could spanner the entire argument if the watt used is twice as valuable as the watt exported...}



This is the best explanation I've seen, It's in regard to light metering and luminous flux density.

Skip to 5mins for the relevant segment and all told by 7mins50.

What's the best roofing material/mounting system to cool the rearside of the panels?

freddyuk

Just place a light(watt/m2) meter on the module and move it from horizontal to 45 degrees and you can see the difference in potential output. Once the sun disappears the effect is much less marked and horizontal becomes potentially a better option.
If you are building a house from scratch you build it to accommodate the solar input not try and fit it in afterwards which leads to disappointing results. Has the architect ever done a solar powered building as he could easily amend this design to satisfy all the issues.
There is even an Irish company that does clear backed PV panels to allow maximum light through the panel so these may be used as roof lights but to build anything with an inherent shading problem is simply not on.

freddyuk

freddyuk wrote: »

Just place a light(watt/m2) meter on the module and move it from horizontal to 45 degrees and you can see the difference in potential output.

DC Clamp meter works too, once they're commissioned, bit more useful around the house.

freddyuk wrote: »

..but to build anything with an inherent shading problem is simply not on.

+1

freddyuk

The watts per sq meter is what makes the modules perform so that is "input" and the readings are very accurate and high resolution so easy to see the rapid deterioration as you come off the perfect angle. Clamp measuring volt output is lower resolution so not quite so obvious if you looking for the perfect angle! Still works though.
If you want to compare 2 different panels with same light input and then compare the output with clamp meter you know which is better. Both essential tools!(for MCS certification).
Decent clamp meter absolutely should be in the tool box but DC current versions are a bit more pricey so most don't have them.

Even decent clamp meters aren't that accurate under 1A, I just use mine for quick reference and relative values. ie. set it to the highest output and then check the shunted meter.

The light meter is probably best for dry testing which has it's own calculation margins for errors...here's a white paper on the subject. Admittedly I only skimmed it, looked at the graphs and the conclusion while contemplating which hat I'd prefer to eat.

quentingargan

I ran a simulation on PVSyst to compare the potential performance of panels at 10 degrees facing west with panels facing south at 35 degrees. Over the year, the total energy produced would fall by 16%.

Shading clearly is a bad idea for any array, but going off the optimum angle and orientation doesn't have as much effect as you might expect, and I expect that Table H2 in the DEAP manual is quite accurate.

If that is the case, horizontal would be better than any orientation east via north to west. That may be the case. If you have an east-facing array, the shallower the pitch is, the better because it picks up diffuse light all day long.

In hindsight my bias is for battery charging so I'd be aiming 50° to get the most of the early Spring late Autumn rays.

Have you ran PVSyst compared to your own system Q. Did you find it accurate?

The panels Freddy mentioned earlier are called glass on glass panels which are translucent. I'd expect the inherent failing of using them as a roofing material is the reduced output compared to a panel with the rearside ventilated as the heat from the house will reduce their performance.
I believe heat de-rating is in the order of 1% per 2°C over 25°C.

quentingargan

Sir Liamalot wrote: »

In hindsight my bias is for battery charging so I'd be aiming 50° to get the most of the early Spring late Autumn rays.

Yes - although the yearly average is down only 16%, the winter daily average gets totally hammered. The summer daily average is interesting for self-consumption - lower peak but a longer spread. The two charts attached have three lines for 4kw systems; 30 degrees facing South, 15 degrees facing south and 10 degrees facing east and west (8 each way)

What's the Y axis Q?

quentingargan

Sir Liamalot wrote: »

What's the Y axis Q?

Y axis is watts at that time of the day. PVsyst gives you an expected yield for a particular day and I picked mid-summer and mid-winter. I'd need to re-open the project to get daily averages. Didn't save it.

The attached charts show the yield each month for each system.