Thermal expansion

[dkrpink]

I am having an issue on a job with thermal expansion.

The structure is an exterior canopy (open structure - no walls) that utilizes open framing (i.e. all the framing is open below and visible).

The purlins for the roof are 10" tall 18' long 12 ga. cold form C-channels. The client does not want animal issues (such as birds nesting) in the c-channels. As such a 20 ga. cap is being installed to enclose the section and make it appear to be a tube. The cap is a 20 ga plate that is being screwed to the lip of the c-channels at 12" o/c top and bottom with #12 ICC approved self drilling self tapping screws. The screws have an allowable load of over 650lb in shear and over 2,000 lb ultimate. The cap is strictly architectural in nature.

The issue is during temperature swings from day to night of 40 degrees the screws are shearing off at the joint between the cap and the purlin.

Originally we perceived this to be an install issue as the screws were installed with a screws gun with double the rpm's than allowed. However they have been re-installed meeting the requirements of the screw manufacture and failures are still occurring.

The purlin and cap are different gauges; purlin is 12 ga, cap is 20 ga; but are the same material spec. ASTM A653 cold form steel.

Based on our equations the length that a steel member will expand is not conditional on any other property other than the length, thermal expansion coefficient, and temperature swing; change in length = thermal coefficient * length * change in temp.

Based on this should the two members not being expanding the same length together?

Just to be safe after the original issues we calculated the force required to restraining both members from expanding. The equation for this being; force = cross sectional area * modulus of elasticity * coefficient of thermal expansion * temperature swing. Based on that we then took the difference in the forces and determined the number of screws needed and then placed them uniformly through out the connections. The issue is still occurring.

Any ideas on what we might be over looking or doing incorrectly?

 

[desertfox]

Hi dkrpink

A sketch of the joint would be helpful if you can provide one.
However the fact the materials are identical does not prevent failure due to thermal expansion or contraction, the assumption with the formula you are making is that the temperature rise or fall is uniform but in practice this is not always the case ie:- if once face of the channel is subjected say to the sun, then the channel will start to expand, but other parts of the channel not exposed directly to the sun will not expand as fast and hence that will generate thermal stresses.
You might be able to work this problem backwards, if you know the force required to shear the screw and consquently the shear stress, work backwards using the thermal expansion formula and solve for the temperature difference that would generate such a force as to shear the screw.
If the temperature difference is believeable then you know thats the problem, if not your looking for something else.

 

[JAE]

I can't see how thermal changes could affect this directly. My first thought had something to do with torsion as the channel is not doubly symmetrical and as it is loaded by its top flange (not at its shear center) you might get some torsional twist going on and with the plate trying to cover it, there would be longitudinal, torsional stresses developed across the screwed joint.

Is the cover plate one, long 18 ft. plate, or a series of smaller plates? Is there any bridging between the joists?

 

[dicksewerrat]

Stamp an elongated hole in the cap. Do not screw down tight.

Richard A. Cornelius, P.E.

http://WWW.amlinereast.com

 

[dicksewerrat]

Sorry I forgot to say, put the screws in the middle of the hole.

Richard A. Cornelius, P.E.

http://WWW.amlinereast.com

 

[Ron]

The coefficient of thermal expansion is a volumetric function, but the thickness is usually ignored as the delta L is usually so small in the thickness direction. In your case, the two materials are the same length but their volumes are quite different. The 20 gage material has only about 1/3 the thickness of the 12 gauge material, so its linear manifestation of volumetric growth will be relatively greater.

Further, if the "cap" is covering an airspace, over the purlin, then it is "shading" the web of the purlin, thus you have an even greater difference.

Paint the cap a light color and divide it into 3 or 4 linear sections, with lap joints. Spread your fasteners out a bit more.

 

[hokie66]

I would suspect defective screws. Suggest using a different manufacturer's screws, or the next larger size.

 

[hokie66]

Another thought...if the purlin is heavily restrained at one flange but not the other, bending due to the temperature change could mean this is a web to flange shear flow failure???

 

[Ron]

hokie66...I agree with your shear flow premise. That's probably the root of the stress concentration.

One problem with self-drilling/self-tapping screws is that they can become strain hardened during installation, thus susceptible to brittle failure.

Slotting holes in the cap would likely solve the problem as well.

 

[paddingtongreen]

Another variation, if the sun shines on the roof, it probably heats the top flange of the channel as well, but the cover doesn't have a flange to get heated. The channel would want to hog much more than the cover until temperatures equalize. I would use large holes and cover them with washers, or, if acceptable, cut the covers into shorter, overlapping lengths.

Michael.
Timing has a lot to do with the outcome of a rain dance.

 

[dkrpink]

Thanks everyone for your input.

I was at the site today. It turned out there were not many failures in our fix at all only 3 screws out of over 4,000 screws so less than 0.075% as such this is clearly a reasonable number to consider those screws defective.

However we have gone with over sized holes with neoprene washer hopefully to allow for some movement to occur.

If I don't post again the solution has worked.

I appreciate all of the thoughts they were all very similar to the thousands of scenarios I have run through my own head trying to figure it out; but I wanted to see how legitimate others though they might be as well.

 

[desertfox]

Hi dkrpink

Your welcome