Shell Thickner (Doubler) Stress Concentration Effects 4

[jeffhmc]

Hi everyone!

As a young engineer, I am turning to the experience of you older and wiser engineers. My situation is this: several tie rods that were used to negative pressure and thus buckling of a duct snapped during vibration tests. The tie rods were welded to opposing sides of the duct (a think aluminum shell). If you can imagine, it basically looks like sticking a pencil through the sides of a coffee can. Washers quite a few sizes bigger than the OD of the rods were welded to the outer face of the shell, then the rods were welded to the washer.

So, my question is, are there commonly seen problems with this sort of weld connection and strengthening the connection with a doubler? It seems the more common problems are fatigues and ruptures in the shell where the doubler's are connected, not in the tie rods. Also, where can I find standards in sizing up the doulber compared to the OD of the rod (or rivet for many other applications)?

I'd really appreciate any comments, suggestions of sources, or just intuitive experiences from anyone! Thanks!

~Jeff

 

[rb1957]

"sorry mate, don't get your banter" ...

i can't visualise "several tie rods that were used to negative pressure and thus buckling of a duct snapped during vibration tests. The tie rods were welded to opposing sides of the duct (a think aluminum shell). If you can imagine, it basically looks like sticking a pencil through the sides of a coffee can."

i can visualise a coffee can (albeit with a thick wall) with a pencil sticking through it (tho' i'd ask myself why?). do the ends of the pencil load the sides of the coffee can, and so cause the coffee can to buckle ? is the coffee can moving (being shaken, vibrated) so that the pencil ends are impacting a rigid foundation (loading the coffee can); or are the ends of the pencils attached to the foundation, and the coffee can is shaken (so that the pencils are nodes).

i think once this is described better, there may be some suggestions about fixing the problem.

 

[berkshire]

jeffhmc (Structural)
You may do well to post this question on forum 725, Welding Bonding and Fastener engineering.
I have seen the type of failure you are describing but not in an Aircraft application.
It seems to occur most, when insufficient tie rods are used, allowing the shell to continue to flex, resulting in a fatigue failure at the weld juncture.
B.E.

 

[Sparweb]

If I'm picturing it correctly, the tie rods are in tension because they are partially retaining the tube from hoop stresses, ie., the stress that makes the tube expand in diameter.
The term "tie-rod" often refers to 4 rods running along the outside of a cylinder that keep caps on the ends, like in a hydraulic cylinder. The case you describe puts rods in transversely, and that's why it's confusing. I've never heard of this, either.
This sounds like a "band-aid" for a tube that is underdesigned for the system pressure, but space limitations may have forced the situation. The tube will be "pin-cushioned" at each of the reinforcing rings at the end of each tie-rod.
Calculate hoop stress, find the load on each half-shell, and divide that load equally among the tie rods in the tube. Account for other stresses in the tube, temperature, and you ought to be in the ballpark.
BTW, hardware-store threaded rod is soft as butter, unless your tie-rods are made of something hardened.

A much more effective way of reinforcing a tube is to use rings around the circumference.



Steven Fahey, CET

 

[GregLocock]

The rods may have snapped at the end of the HAZ from the welding.

Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[jeffhmc]

HI Steven(Sparweb)..thanks much for the replies. Sorry about the confusing description. The duct is used in a heat exchanger for an aircraft with high velocity air flowing through it so it'll have negative pressure inside. The negative pressure is sucking in the sides of the duct and would cause the duct to buckle (i.e. implode), so the tie rods are placed to keep the duct from imploding. The failure occured during the vibration test, so the heat exchanger wasn't in operation. The failure of the connection actually occured at the ends of these rods rather than on the shell (which is thought to be much weaker than the rods).

You're right that these rods were put in to salvage a bad design. I would have gone with the reinforcing rings around the circumference, so its good to know my intuition was correct. However, I'm only asked to evaluate the failure, not design the unit.

The rods actually are aluminum, but the ends were not heat treated after welding so the ends where the connections are appear to be the softest/weakest.

Also, it was suggested that the failure resulted from large deformations caused by acoustic feedback within the duct. Does that seem possible (or plausible) to anyone?

 

[GregLocock]

It is conceivable - I know that the Shuttle had an acoustic fatigue test. Alternatively the rods mave gone into resonance themselevs, or the duct wall, etc etc.

Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[jeffhmc]

Thanks Greg...my team and I have already done a physical test of the system and determined the resonances and modes and so forth. Was just wondering if anyone has seen acoustic feedback contributing more to the dynamic response than pure configuration of the system and the resonance that results.

 

[aviat]

jeffhmc:

There is a failure mode called acoustic fatigue, which occurs in ducts. It has been known about for a long time, so you should be able to find some information.

First. Have you determined that this is a fatigue failure? Does it have the beach marks?

You say that your rod is a tie, which would put it in tension. Then you say that it is installed inside the duct and prevents the duct from collapsing, which would put it in compression and a strut, wouldn’t it?

Finally, what type of aluminium are you using? There are several types that are not weld able, and other that need heat-treating after welding. Welding and vibration can give lots of problems.

 

[Sparweb]

Positive pressure or negative pressure - it doesn't matter, rings reinforce tubes better than rods can, and if they're on the outside, they won't disturb the flow.

I still don't understand what's going on with your duct that made the rods necessary.

Which way to the rods go, BTW?
inside or outside?

_ \ /
(X) O
- / \



Steven Fahey, CET

 

[jeffhmc]

Hi Aviat:

- The materials engineer concluded it was fatigue, but there were no beach marks or necking where it occured.
- I appologize for using the term "tie rod" incorrectly. That's the term used in the design reports and such so have kinda gotten used to using it.
- The duct and the rods are aluminum 6061, standard stuff.
- Thanks for the hint to look into acoustic fatigue...has turned up some pretty good sources.

SparWeb:
- So what's going on is the air inside the duct is very high velocity (used with RAM air flow on a plane) while the air outside the duct is compressed by the cabin or inside air of the plane. So the pressure difference tends to compress the duct circumferentially. The shell of the duct is very thin and can't handle this compression.
- Nice sketches btw! Actually, the rods go like this:

(-)

Thanks for the comments and insight guys! Just discussing the problem is helping me see this problem in a differnet light.

After reading my original post...I'm a bit embarrassed...it's totally incomprehensible! You'll have to forgive me...was working late into the light

~Jeff

 

[jeffhmc]

Actually, I'm as confused as those of you keeping up with post on some of the things that were (or weren't done)...such as heat treating the connection after welding. But I'm just asked to figure out what caused the failure.

Also, I'm being asked to figure out why there was such low damping in the system (~ 0.1%, damping ratio of 0.001). Has anyone ever seen damping in this range? Just another question to add to the pile...

~Jeff

 

[GregLocock]

Damping ratios of 0.1% are common in metal structures that have no discrete damping, such as crankshafts, engine blocks, etc.

So, in a simple welded structure, 0.1% seems reasonable.





Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.