Not logged inAmerican Welding Society Forum
Forum AWS Website Help Search Login
Up Topic Welding Industry / Technical Discussions / Do all fillet welds always fail in shear?
- - By Kix (****) Date 02-11-2009 16:03
   I was wondering this when I saw this part that was subjected to destructive testing.  This picture is a rectangular piece of plate with a hole in it.  A 5" piece of pipe slides through the hole very tightly.  It gets welded all the way around the pipe.  You can see in the picture (hopefully) that the force is exerted downward while the pipe end stays in a fixed position.  I marked where the first base material and second base material failures were resulting in the arm falling to the floor..  Now for the question.  Did those fillet welds fail in shear the full length of the failure or just some parts of the fillet weld?

http://i11.photobucket.com/albums/a189/vdubin474/l-braket.jpg
Parent - - By swnorris (****) Date 02-11-2009 17:16
Am I understanding you correctly... there was no applied downward force, and the weld ans base metal fractured just by the pipe/plate assembly hitting the floor?
Parent - - By Kix (****) Date 02-11-2009 17:23
You will see by the arrow at the end of the arm that there is a downward force on the arm while the pipe end stays in a fixed position.
Parent - - By DaveBoyer (*****) Date 02-12-2009 04:31 Edited 02-12-2009 04:34
   My Guess: The first material failure failed in tension. The weld then failed in tension progressively from the point where the material first failed to the second material failure. Had the arm been made from wider stock that could withstand the load applied, then the weld would have been loaded in shear, in it's entirety, and the load at failure would have been much, much higher.

In My opinion this joint functions like a moment connection.
Parent - - By Kix (****) Date 02-12-2009 14:02
This part is made from 6061-T6 aluminum and welded using a 4643 filler with no post weld heat treat.  Why the 4643 was chosen in the first place is beyond me.  I went to the design engineer and asked if I could switch the filler to the much cheaper 4043 (4043 5.31 a # and 4643 21 dollars+ a #)  He said he based his decision off of some chart that had 4643 shear strength at 13.5ksi and 4043 shear strength was 11.5ksi.  He told me told me that this fillet weld failed in shear because fillets always fail in shear so we have to keep the same wire.  The part withstood 20,000 psi before it failed which is well above what it was designed for according to the engineer.  He is afraid that if we change wires that it would fail below 20ksi somewhere and he would loose his 2 to 1 safety factor.  I understand where he is coming from, but I don't think he will loose his 2 to 1 safety factor.  For one, the welds on the part that failed were absolute Sh!t.  No penetration to the root, lack of fusion everywhere and undersized welds.
     Now, here's my take.  I've never seen 4643 used in a chart for welds in the as welded condition because it is used for post weld heat treat applications.  I was able to get our procedure qualifications to pass in the as welded condition using the 4643, but tensiles went at just above 24ksi.  I also did a procedure qual with 4043 in the as welded condition and it's tensiles went at 25ksi almost 26ksi.  Now you look at any chart for 6061-T6 welded with 4043 in the as welded condition and tensiles are supposed to be at 27ksi.  Needless to say I don't go by charts that often.  I know I know, you think I must have did something wrong while welding up the sample, like it must of got to hot or something.  Well, we welded it really slow to allow interpass temps to get well below what they were supposed to be because I knew it was going to be a pain to get this thing to pass in the as welded condition.  So what do ya think about what I got cookin here?:-)
Parent - - By Metarinka (****) Date 02-12-2009 19:42
From my understanding. The reason they say "Fillet welds always fail in shear" is twofold

A) a true definition of a fillet weld is a weld placed on the exterior of 2 pieces of joined metal. Therefore in all common fillet joint configurations the forces between the two members will always cross the weld out of plane of the base material. This creates shear stress in the weld regardless of the type of loading present in the base material.

B) The allowable shear stress of a fillet weld is always much lower than tension or compression strength by a significant factor. Therefore even in the rare chance that the design is proven not to carry a shear force, allowable shear stress is generally considered the design limitation of fillet welds. This is because it is generally impossible to rule out that shear stress will be present in the weld in service. I have a picture here from an AWS book showing examples of Fillet welds in compression or Tension from a design standpoint. I can upload it if you're interested.

From a practical standpoint? No. Not every fillet weld fails in shear. In this particular case if I understood the picture the weld failed at the toe? Between the weld and base material?  In your diagram is the weld subject to a rotational moment about the pipe axis? or is it strictly a downward force it's kind of hard to tell. At either rate if the weld failed at the junction between the toe and base material. Then it means the allowable shear stress of the weld throat was not the controlling factor meaning there are metallurgical or design considerations between the weld filler and the base material.

Finally, and I hope someone will correct me if I'm wrong, but from my material between AWS and ASME I was always taught that fillet weld designs are based on the tensile strength of the weld NOT the shear stress. So while shear stress is considered the failure mechanism; from a design standpoint Tensile strength is used to determine allowable stress under ASD.  For example in ASD for non tubular connections allowable shear stress is considered to be .30x nominal tensile strength of the filler material.

Hopefully one of the more experienced engineers can correct me if I'm wrong, but I've never used the shear strength of filler material as the basis for designing a fillet weld connection.
Parent - - By Kix (****) Date 02-12-2009 20:18
Thank you Metarinka, you comment was very well understood and well put.  The weldment is subjected to a downward force out on the end of the arm.  The pipe end is fixed so  the arm would have to rotate, but it can't because it's welded together.  Your comment says what I was thinking all along, why would you base the strength of your part off of the shear strength of your filler metal in a fillet weld.  Especially on one that is best for post weld heat treating.  The weld failed because the base material area at the top of the pipe was to small.  That let go and then the weld ripped around the scribly side of the pipe.  The design engineer has increased that area on top of the pipe now and I want to change the filler.;-)

About the pictures in your AWS book, could you upload those for me?  I'd really appreciate it.

   Thanks, Kix
Parent - By Metarinka (****) Date 02-12-2009 20:29 Edited 02-12-2009 20:34
Here you go, both were taken from AWS Design for Welding, which is a useful book for engineers to get their feet wet in welding design theory. Again, I'm hoping a more experienced engineer will correct me if I'm wrong, but too my knowledge fillet welds can be loaded under tension or compresion  if there is even loading  such as  (C ) in the first picture.  An example would be a splice in an I Beam that's in compression.
I don't have any practical experience with welding design for tubular connections though.

Oh and while I have you attention. I remember you mentioned knowing Bill Figg from WCC in Michigan, I'm curious as to how you know him, have you ever visited WCC?
Attachment: typesofloading1.jpg (460k)
Attachment: typesofloading2.jpg (479k)
Parent - - By DaveBoyer (*****) Date 02-13-2009 04:55
   I agree with the design change to increase the strength of the top web of the plate. The original design did not transfer any load to the weld on the half of the diameter that didn't fail.
Parent - By jon20013 (*****) Date 02-13-2009 07:45
I think Metarinka said it about as well as anyone ever has, good job!
Parent - - By Kix (****) Date 02-13-2009 16:03
But do you agree with the change in filler.  That's my question.  We know the base material failed in tension at the top of the pipe.  Directly below the material failure the fillet weld failed in shear for an inch or two and the rest failed in tension.  Now that we know how the thing is going to fail in real life and that the base material area of the plate that failed has increased, can we change the filler?  Keep in mind that I had higher tensile pulls with the 4043 then I did with the 4643 in the as welded condition.
Parent - By DaveBoyer (*****) Date 02-14-2009 02:08
I agree with the 4043 filler for all the reasons You chose it.
Parent - - By Bob Garner (***) Date 02-13-2009 18:38
Metarinka,

I agree with everything you said about fillet weld design and theory - just some validation from an old fogey structural engineer.

BTW, I thought your explanation was very good! 

Bob Garner
Parent - - By 803056 (*****) Date 02-15-2009 16:13
One aspect of aluminum often overlooked is what happens to the strength of the 6061-T6 aluminum when it is subjected to thermal cycles while welding. The tensile strength of the base metal, 6061-T6 drops from about 42ksi UTS to about 24 ksi UTS if everything is done correctly.

The shear strength of the filler metal as well as the strength of the HAZ must be considered when designing the welds. My point is that even if the filler metal has a shear strength of 100 ksi, it provided little benefit if the weak link is the underlying HAZ with a tensile strength of 24 ksi. The numbers of 100 ksi and 24 ksi were selected to be widely different, not that it has anything to do with the 6061 aluminum alloy in question. The designer has to consider the shear strength through the throat of the fillet weld, but he also has to consider the shear strength of the HAZ.

I've been involved in cases where the designer only looked at the "as wrought" properties of the base metals and the shear strength of the fillet welds without due consideration to the loss of strength in the aluminum alloy from welding. In one case, two men were killed when a window washer unit collapsed while they were working on it.

Best regards - Al 
Parent - By Kix (****) Date 02-16-2009 14:14
Al,
     I agree that the as welded base metal strength is going to be much lower in the T6 after welding.  Somewhere my design engineer got these shear strength #'s for aluminum fillers.  They are as follows, 4043 = 11.5 Ksi and 4643 = 13.5 Ksi.  He really likes that 13.5 Ksi for the 4643.  These do not specify if they were calculated in the as welded condition or not. 
     It just bothers me that the information they are going off of was from a bunk test ( Crapy, small, no fusion fillet welds and a 1/2" tall lip on top of the plate).  They now have over an inch lip on top of the pipe providing more surface area in the base material and actual welding procedures that cook up excellent welds compared to what was on the test weldment.  That particular weldment still held well over what it was designed to hold (20,000 psi) so I'm sure it would go even higher now with the design changes, proper fillet welds and if it was welded with 4043.  Well, at least that's my assumption anyway and we just can't go off of my assumptions. ;-) 
Parent - By Metarinka (****) Date 02-17-2009 07:27
I'm glad my education is working, no one likes a young hot shot engineer who's quick to spout out information  and even quicker to be wrong about it.
Up Topic Welding Industry / Technical Discussions / Do all fillet welds always fail in shear?

Powered by mwForum 2.29.2 © 1999-2013 Markus Wichitill