A practical example of, and warning about, welding distortion

admin

During my day job, I regularly weld AWS test plates that are used to produce mechanical and metallurgical test specimens.  The first attached image is a picture of a typical plate(not my weld).  The plate is 2 pieces of 1" thick A514 low alloy steel.  Each half is 6" wide and 12" long.The second is a photo of a strongback.  2 of these are welded to the back of a test plate assembly for vertical up or down welding.  The notch is to allow room for a backing strip on the weld root.  The plates must be flat in order to pass visual inspection criteria.  The strongbacks keep the weld from pulling the plates into a V.Notice the crack that extends through the plate.  The plate failed catastrophiclly from as a result of weld distortion.  The plate cracked through the narrow cross-section.  It is not a fatigue crack!   Weld distortion from the high strength, low alloy weldment was strong enough to rupture the plate.  The last photo shows the fracture surface of the crack.  You can see that the chevron marks on the fracture surface point back to the outside corner where the fracture started.  We hammer these plates to break them off of the test plate.  Damage to the plate from hammer marks served as a crack initiation point.Take this for what you will.  I thought it was a striking example of just how much stress weld cooling and shrinkage can create.  It's something to think about if you use similar plates to restrain welds.  Thankfully in this case, the plate remained attached to the AWS test plate, and didn't break loose and fly around when it cracked.If you re-use your strongbacks, take a moment to grind them smooth and, clean them up once in a while.  Hammering on them creates weak spots that can cause them to fail.Also, don't underestimate how much reinforcement you're going to need to restrain a weld from pulling; especially when you're welding thick plate and/or high strength steel. Attached ImagesBenson's Mobile Welding - Dayton, OH metro area - AWS Certified Welding Inspector
Reply:Originally Posted by A_DAB_will_do...The second is a photo of a strongback. ...Notice the crack that extends through the plate.  The plate failed catastrophiclly from as a result of weld distortion.  The plate cracked through the narrow cross-section.  It is not a fatigue crack!   If you re-use your strongbacks, take a moment to grind them smooth and, clean them up once in a while.  Hammering on them creates weak spots that can cause them to fail.Also, don't underestimate how much reinforcement you're going to need to restrain a weld from pulling; especially when you're welding thick plate and/or high strength steel.
Reply:Thank you,  I like it.  Never saw run off tabs like that.So this part has been heated and cooled many times, stressed and beat on.  Fatigue?Thanks for the postDavidReal world weldin.  When I grow up I want to be a tig weldor.
Reply:It's not to hard to figure out how much the part is going to distort and prebend the plates so when all welding is complete it will be pretty straitl --Gol'
Reply:Originally Posted by Go1lumIt's not to hard to figure out how much the part is going to distort and prebend the plates so when all welding is complete it will be pretty straitl
Reply:Originally Posted by denrepNot to hard to figure? It is for me! Go1lum - Any formula or rule of thumb you could share?How about this one: 1" low-carbon plate, butt joint, single V, 12"joint length, unrestrained,  7018 stick; how much will it pull from flat, at 12" out from the joint?Thanks
Reply:There are some guidelines given in Lincoln's handbook procedure.  I get about 9/16th inch lift at 12 inches from the joint.  That makes for a 1 1/8th inch bow across a span of 24 inches.     How many passes are made has a great effect on the amount of bending.  As a rule of thumb I have found the most reliable method is to weld up one joint unrestrained and measure the movement.  On the remaining pieces preposition them that amount in the opposite direction.  Oh Yes!  Also waste an hour or more flame straightening joint number one.  Sometimes it is a small price to pay.
Reply:How much distortion you'd wind up with is dependent on how much pre-heat you use, how many passes are used to make the weld, and how much heat is input by each pass.  We typically preheat these plates to about 300F with a rosebud.  Interpass is maintained at 300F throughout the entire welding process.  Even so, they'd pull more than the 5 degrees that you're allowed by AWS standards.  Hence the Strongbacks.  I've welded a couple of these without restraint, and they do indeed move about one inch out of flat alignment.The runoff tabs are standard operating procedure.  It gives room to establish the puddle and get moving before you're in the region that will contain tensile and impact specimens.  It is a good indicator that for critical welds, using run off tabs in the field will help avoid defects.Denrep,  I've got to respectfully disagree with you.  The fracture initiation point is in the lower right corner of the picture.  the lines visible on the surface all point back to that point, but I don't believe they are indicative of fatigue.  almost the entire fracture surface appears to have cracked at the same time.  I might throw these under a microscope if I get a chance next week.  If I do, I'll post some photos.  It may be that the beach marks are there, and I just can't see them with my eyes.There were some secondary cracks on the left side of the fracture surface, recognizable by the shiny smooth appearance.  These appear to be fatigue cracks to me.  The shiny surface is caused by the sides of the crack rubbing against one another during cyclic loading.Benson's Mobile Welding - Dayton, OH metro area - AWS Certified Welding Inspector
Reply:The part itself doesn't see too much heat, especially not in the region of the failure.  The crack started on the outside edge, the point of maximum tension loading.  They do get the snot beat out of them with 4- 12 lb hammers when the welding on the plates is done and it's time to remove them.  I'd guess this broken plate was used 400-500 times over the last year. Originally Posted by David RThank you,  I like it.  Never saw run off tabs like that.So this part has been heated and cooled many times, stressed and beat on.  Fatigue?Thanks for the postDavid
Reply:Strait out of the Design of Weldments  book. ( it's a lincoln book) Formula for Transverse ShrinkDelta trans= .10 x the average width of weld.Longitudinal bending.005(Aw*D*L^2)/IAw= Weld areaD= Distance from center of gravity to nutreal axis. L= total length of all weld I= moment of inertia of memberAngular Distortion for T situationsDelta= 0.02*W*w^1.3/t^2W= total width of bottom plate on a t joint. ( not weld width)w= leg sizet= thickness of plateYou have formulas now just need to plug and chug. --Gol'
Reply:plug away and show me.  All I have is experience.  David Real world weldin.  When I grow up I want to be a tig weldor.