Two Questions for the seasoned pros of this great forum

[scrambler]

Hi there, I have the following two questions if anyone has the answer.

1- Thermal expansion gap between panels:

Installation usually calls for at least 3/8” between panels to allow for thermal expansion and venting.

Usually mounts create the space between panels in the same row, so it is mostly about leaving a gap between rows.

I was wondering if anyone had actually measured the space between rows in the winter early morning and in the hottest hours of a summer day, and noticed an actual change, and if so how much?


2- Combining Strings in parallel.

Let say you have 20 panels in two strings of 10 (panels 33V Vmp and 9.5Amp Imp). This means each string has a max of 330V and 9.5 Amp, and if they were to be combined in parallel, the result would have a max of 330V and 19Amps.

Assuming a string inverter with two Mppts, each being able to accommodate the full values, Would the best configuration be to use one string on each Mppt, or to combine them on a single Mppt?

Does combining them reduces the efficiency of the system, improves it, or makes no difference?

Side question: If every panel has a Tigo TS4-R optimizer, does that change the above in anyway.


Thank you for your input

[Mike90250]

As long as strings are SHADED equally (all shade, no shade) they should be able to be combined to the limits of a single MPPT channel.

If there is any difference in illumination or heating of the panels, you will get more efficiency with dual MPPT. ( in my opinion )

[scrambler]

Thanks,
Do you think that would still be true if each panel has an optimizer?

[bob-n]

Expansion: I haven't measured it, but based on material property tables, aluminum will expand approximately 23.5 um for each meter of length and degree K (or degree C) of temperature rise. That means that a 2 meter long aluminum panel will expand approximately 0.0000235 * 100 * 2 = 0.0047 meters or 0.185 inches (~3/16") if heated from 0C to 100C. Based on that, 3/8" is very safe, even if installed on the coldest day of the year (which is worst case). If you install during warm weather, you could get by with much less gap. But I wouldn't ever go less than 1/8".

Disclaimer: I'm not a seasoned pro, just a rank amateur who knows material properties.

[Ampster]

I wouldn't even know where to start with a shade analysis. However if two MPPT are better then it may follow that multiple MPPT channels from using Tigo optimizers could be better.

The only argument to using one string with optimizers is that it might reduce the chance of low voltage drop out of the inverter if a lot of panels were shaded since the voltage will be higher on a single string. It all depends on the shading. I tried a few TS4-O on a patio cover that got some shade. Without panel level monitoring I didn't have enough data to reach a conclusion. This was on a west facing exposure that got 4 hours of full sun only in the afternoons in the summer. The further investment of 7 more optimizers and a monitoring device wasn't worth it for that part of my system. Your mileage might be different.

[scrambler]

Expansion: I haven't measured it, but based on material property tables, aluminum will expand approximately 23.5 um for each meter of length and degree K (or degree C) of temperature rise. That means that a 2 meter long aluminum panel will expand approximately 0.0000235 * 100 * 2 = 0.0047 meters or 0.185 inches (~3/16") if heated from 0C to 100C. Based on that, 3/8" is very safe, even if installed on the coldest day of the year (which is worst case). If you install during warm weather, you could get by with much less gap. But I wouldn't ever go less than 1/8".

Disclaimer: I'm not a seasoned pro, just a rank amateur who knows material properties.

Thanks. In theory quite small.
Still interested if anyone has a real life measurement, as contraction and expansion on the roof mounts also probably factor in.

[littleharbor]

Expansion: I haven't measured it, but based on material property tables, aluminum will expand approximately 23.5 um for each meter of length and degree K (or degree C) of temperature rise. That means that a 2 meter long aluminum panel will expand approximately 0.0000235 * 100 * 2 = 0.0047 meters or 0.185 inches (~3/16") if heated from 0C to 100C. Based on that, 3/8" is very safe, even if installed on the coldest day of the year (which is worst case). If you install during warm weather, you could get by with much less gap. But I wouldn't ever go less than 1/8".

Disclaimer: I'm not a seasoned pro, just a rank amateur who knows material properties.

I would think that the glass would have more expansion and contraction being that there's much more glass to expand than thin extruded aluminum frames.

That thought being stated I must confess I have zero expertise or education in this field, just a feeling being there's much more mass of glass to expand.

[bob-n]

Scrambler: You're right. I have not measured it. Roof mounts have flex, however. As the panel expands, they could bend slightly to take up the slack. It's a bit different if you are using a system with rails. In that case, aluminum rails will expand or contract roughly the same amount as the panels, and then again the roof mounts have to flex. I used Ironridge Flashfoot2 mounts. In addition to flex, these have some lateral play in them.

LittleHarbor: In panels I have seen, the glass is soft mounted in the frame, so that it can expand without moving the frame. Only the frame is rigidly mounted to the railing or to the roof.

[J.P.M.]

Most all engineering and construction materials change dimensions as f(temp.). If they are restrained, the resultant induced strees must be less than allowable. Usually both conditions of some flexure that induce some stress exist all over any design and are changing most all the time. The reality of good design is to keep both flexure and stress within safe and workable limits while meeting the design goals.

With regard to the stresses induced in components as a result of such dimensional changes, a good design accommodates the thermally induced dimensional changes by ensuring the induced stresses are maintained below allowable limits for all materials and all anticipated operating conditions while not having the dimensional changes create such things as interferences or other problems.

For panel attachment to racking and the racking itself, that's usually done by building some flexure into the design to limit stresses in components.

For differential expansion due to material contact such as with panel frames and the glass, the frame to glazing attachment design details and including the flexible seal material usually maintain flexibility and leak tightness - not an easy task and which, if you think about it, is pretty amazing. A good mechanical design takes all this and a lot more into account.

Most all the common support/attachment systems (racking) for roof top PV that I've seen should have no problem accommodating the thermal differential expansion they will encounter during operation when they are installed according to mfgs. recommendations, requirements and limitations. For such systems, problems most often originate with what the racking is attached to at the roof attachments and not the racking.

[bcroe]

If my aluminum panel frames are bolted to my aluminum mounting, will they expand together?
Bruce Roe

[scrambler]

Agreed, in a roof mounted system, I would assume that what will move the most are the attachment points on the rafters, as these are the one with the largest expansion possibility.
But how these move is quite unpredictable given the interconnection, the plywood underneath....
Some of these could actually compensate each other.

That is why I am curious if anyone has measured the outcome between panels

[scrambler]

The thing also is that even a tiny expansion could potentially cause significant warping if contained.
Using the nifty Arc calculator down below, fitting a 67" panel in a space that is just 1/16" shorter, would normally require it to bend and lift 1.25".
if 1/16" seems insignificant, 1.25" is not...

Arc calculator:
https://handymath.com/cgi-bin/arc18.cgi?submit=Entry

[J.P.M.]

If my aluminum panel frames are bolted to my aluminum mounting, will they expand together?
Bruce Roe

Assuming the "they" in your question refers to the panel frames and aluminum mounting and not just 2 or more panel frames, and discussing thermal expansion only without other stress/strain inducing considerations such as any external loadings, provided the fasteners/connections do not change dimension or fail, the short but probably incomplete answer to your very specific question is yes.

A bit more specifically, as long as the connections hold, the joined pieces of an assembly will be in the same location at the connections, but the connections will probably move as f(temp. of the components, restraints).

If the connections holding the pieces together do change dimension or fail, the most likely answer to your question is no.

Each component in an assembly will change dimension (and that means BTW, contract as well as expand) as f(its own temp.) as modified by (any) restraints imposed on it by other components, connections and anchorages. Those restraints will change the state of stress in each component and each state of stress will most likely be different for each component.

Both stress and deformation needs to be considered for each component and connection in the assembly as well as for all anchorages.

Read on if you want a bit more. If you do, I'd only respectfully request you not call be a condescending prick and suggest you be more careful in what you ask for.

Maybe a very over simplified example will help: Take two pieces of 6061 aluminum, say, two long sides of a PV frame, probably something like 1.62m (~ 67") long. Through bolt them together close to or at each end with a washer/spacer of a few mm thk. between the pieces and put some insulation of the same thk. as the washer between the pieces so the frame members are not in visual or physical contact except at the bolts. For simplicity (but realistically incorrect) assume the bolting is non conductive of both heat and electricity. Say the temp. of the pieces at assembly was 20 C. and the current ambient temp. is 20 C., the temp. is constant and it's midnight so no solar input.

Without energy exchange other than with the atmosphere, the entire assembly will be at 20 C. There will be no dimensional difference between this condition and the assembly condition and also between the pieces, and no strain in the pieces and no shear stress at the bolts. That also means there will be no axial stress in either piece. If the temp. of the entire assembly then changes in a uniform fashion, that will be a case of unrestrained expansion.

Now, add heat to one piece so it's temp. increases to 120 C while ensuring the other piece gets no heat input so that piece stays at 20 C giving a delta T of 100 C between the two pieces. Assume for now and for simplicity that the heated piece doesn't get long enough to buckle or get beyond its elastic limit and permanently deform. Also assume that the cold piece does not reach its elastic limit when stretched.

Q: With the pieces at two different temps., will the two components pieces expand together ?

For the simple example above, the strict and simple but incomplete answer is yes.

The rest of the answer for both your question and my example, at least to a pretty good 1st approx., is that the stresses in the pieces will increase as f(temp. difference between the pieces, attachment methods), the entire assembly will, most likely, and depending on the temperatures of the various pieces, change dimension, and those stresses and deformations need to be considered.

For your question and analysis, after the stresses and deformations are checked in the assembly pieces, the need will shift to a check of the stresses and deformations in the anchorages and connections (roof attachments, holes in the ground, etc.) both imposed and allowable in those components.

In general, if the stresses and deformations become large enough in any component or connection, bad things will happen. Commonly, members hold, connections weaken/fail first, usually or at least often from repeated occasional loadings or cyclic (including perhaps thermally induced) stresses.

There's lots more to it, but that's the main stuff.

If you're still interested, for the portion of my example and perhaps somewhat applicable to your question, when heat is applied to one piece and not the other (for your example perhaps when a panel frame that gets warmer than the racking), the hotter piece will expand - it'll get longer by ~ (1.62*(120-20)*2.34*EE-5)) = ~3.8mm/2 = ~0.075" from its assembly condition and temp. (and it's volume will change and so the section modulus which will be important if bending or buckling stresses will need to be considered, but let's keep it on topic).

Note: That 3.8 mm over a 100 deg. C delta T is divided by 2 due to the pieces being identical. Note also however, because the temp. of the pieces is now different, this can be considered a case of restrained expansion for each piece (but not necessarily for the entire assembly).

The strain will be the deformation of each piece/the length of the pieces : 0.075"/67" = 0.0011.

The induced (axial) stress in each piece will be something like (elastic modulus of 6061 aluminum * induced strain) = (10 EE7 psi)*(0.0011) = 11,200 PSI.

That stress or any other (more) severe thermal stress(es) will be combined in the appropriate way with the other stresses imposed by other external loading(s) and conditions as required by application of appropriate and required design codes.

For my example, because the pieces are identical, and assuming the bolting holds tight, the cooler piece will undergo an essentially identical dimensional change (it'll also get longer) but the resulting imposed stress will be tensile. Also assuming the pieces are essentially identical in all ways, the stresses will be approximately equal in magnitude but opposite in sign.
For your question, if the pieces are not identical in terms of dimensions including length, the axial stresses will probably be different.

Plug/chug/modify for your example as your needs dictate.

Respectfully,

[solarix]

1) I've never seen a problem due to expansion/contraction. Maybe in a commercial installation where the runs are much longer. There was a famous fire years ago on a big rooftop installation where they did not use expansion joints on the 100'+ runs of EMT conduit and and it pulled apart enough to cut into the wiring. Poof.
2) Definitely use both MPPT inputs. With equal strings, the gain might be small, but why not make use of the possible optimization? Will be a bit more wire I suppose. Also, its always nice to be able to observe both strings separately on the inverter monitoring. We try to use equal strings whenever possible so years later, you can compare the two input voltages in case there is any question of failed panels.

[bcroe]

Given the exposure to the elements, there could be quite a bit of dimensional change associated with a
solar array. The change occurring over and over for decades could be causing fatigue/cracks, but I
have never heard of this problem with PV solar so far. Clipped on panels might start working loose.

Given that the frame/support typically is much heavier than a panel frame, the panel might be taking
most of the fatigue. Or, perhaps the panel frames are already designed to relieve this stress without
damage?

With so many panels mounted with their frame perpendicular to the support, the frame might just
move in and out a bit. But with my landscape mount panel mount frames running the same direction
as the support, there might be less flexing, more potential for failure. Will report in 10 years.
Bruce Roe