weld end returns

huangyhg

weld end returns
when you have a framed beam connection, there is a length of return (example: pg 4-23 of asd).
i have a gusset plate connection.  for discussion purposes, lets say the gusset is 12".  this 12" plate is welded to the column and beam/base plate with fillet welds.  the fabricator calls for 11 1/2" of weld.  i sent back the shop drawings telling them to continue the weld around the corner, under the same principles of that of a framed beam connection (pg 4-23).  they are questioning why they have to continue this weld.  also, on pg 16.1-55 of 3rd ed. lrfd, section 2 under fillet weld terminations, states that for cyclic forces (which i have at gusset plates/brace connections), fillet welds shall be returned around the corners.
am i right in having the fabricator return these welds, or am i just making extra work for them?
find a job or post a job opening
was 11 1/2" called for in the drawings submitted to the detailer?  if so, having them add 1/2" plus the return is asking them to do more than what was originally specified.
as for whether it is asking them to do a lot of work or not, it probably isn't.  it is less than 5% more weld than what they would have to provide anyways.
for cyclicly loaded joints, return weld will reduce the stress concentration; however, is 11 1/2" weld can carry the demand with enough reserve for additional loads, it probably doesn't matter anyways.
my two cents for whatever it's worth.
pylko...unless your fab shop is an experienced bridge fabricator, they are likely not familiar with fracture critical fabrication.  with that in mind, a "ragged" weld termination can be critical, and the end of the gusset is a bad place for that to occur.  yes, it's a bit more work, but should be done.  
if they are hung up on length of weld and their productivity reduction (weak!), and intend to use this as a means to submit a spurrious claim for extras, have them terminate the weld short on the inside of the gusset but continue it at the outside.  i don't know your load conditions, but the plate stresses are likely lower at the inside than the outside.  further, your 11-1/2 inches should be considered a minimum length of weld.  i don't know what your weld symbol shows but you might want to check and see if you had a "weld all around" inclusion in the symbol.
i had originally designed the worst case gusset plate connection and carried it through the building, with the all around symbol.  i figured it would be easier to cut 108 gusset plates all the same size at once, instead of 6 plates 18 different sizes.  they came back, asked for brace forces, and submitted their own gusset plate and weld design, which i approved, figuring they'd know whats easier for them to do.  in the shop drawings (using example above), if they had a 12" plate, they called 11 1/2" of weld.  thats when i sent them back with the "all around" symbol with the shop drawing mark "furnish as corrected".  they're questioning that as "we've never had to do that before".
the provision in 3rd ed. lrfd (pg 16.1-55) seems like something more for bridges, not the rare wind/seismic event.  but this also seems like a similar condition to providing a return at framed beam connections (illustration on pg 4-23 of asd.)
pylko....good for you!! hold those "delegate" engineer's feet to the fire!
looks like you did all the right stuff.  while buildings are generally designed as static structures, that's your call.  after all, pieces and parts can fail from any number of cyclic loads...from 1 to bezillions...depending on the level of stress relative to the material's s-n curve.
deciding whether or not you need a return usually involves the nature of loading on the fillet weld.  if your weld will experience bending, i.e. the ends of your weld are more stressed than at the middle of the length, then a return will drastically improve your performance, esp. for fatigue.  if, however, you have sized and layed out your joint such that the welds are subjected to uniform loads, as one does when carefully selecting gusset plate details, then, all you really care about is the weld length, and 11.5 vs 12 inches is not really that different.
that may be why your fabricator is puzzled.
that's my 2 cents ($canadian), so that's like 1.2 cents us...
ga
thanks all
aisc clarifies j2.2b as follows-
"section j2.2b(2) of the 1999 lrfd specification applies to the following
conditions:
1. connections/structural elements with outstanding legs.
2. cyclic forces of frequency and magnitude to cause a fatigue failure.
wind gusts and seismic excitation do not qualify as cyclic forces that cause
fatigue.  even though they are cyclic forces, they do not provide enough
cycles to cause a fatigue failure.  this provision is intended for
applications were vibration, such as that from machinery, creates a high
number of cycles to create a fatigue failure.  also, this provision applies
to outstanding legs (i.e. angle connection elements only.)"
pylko,
i agree with trainguy that how the loading is applied is important and that translates to 'how the structure was actually designed'. what end conditions were assumed by the designer are a must to consider before you say yes or no to the weld. if the moments are considered 'released' at the end, you may be satified with the 11.5" weld only. if not, the bending will have to be taken care of, as trainguy says and you may want to increase the weld to satisfy design conditions.
another 0.02 to the pile of $$.
regards,
flame