Conventional wisdom is not always the wisest counsel when it comes to fillet weld failures, advises welding consultant, Girish Kelkar, Ph.D.
While conventional wisdom about fillet weld failures is accurate when it comes to joints under stress from longitudinal load forces, it does not accurately explain failures for joints under stress from transverse loading, Kelkar, a presenter during CWB Group’s recent Welding Industry Day virtual conference, explains.
Acting as the bond between two components, fillet welds are one of the most common welds, comprising close to 80% of all the welding done, especially in automotive frame welding. What’s particularly challenging with fillet welds is that all the force is transferred through the weld fillet because there is no other direct connection between the two components the weld is holding together. So, obviously, integrity of the fillet weld is important.
There are two primary designs used in fillet welding: One is a Tee Joint in which, as the name implies, two structures form a tee junction, and the fillet weld joins the two. The other is a Lap Joint, which refers to a joint between two parts sitting on top of each other. The loading on the welds can be either transverse (one of the parts is being pulled upwards) or longitudinal (the two parts are shearing against each other).
“The conventional wisdom is that weld throat is the weakest link, and that all fillet welds will always fail at a 45 degree° angle along the throat, irrespective of the type of loading—longitudinal or transverse. Using the conventional wisdom you end up with field failures that don’t seem to make sense because they didn’t fail at 45°,” warns Kelkar, who consults on an international level and is the author of The Weld Nugget newsletter.
He explains that while this thinking holds true when under load in the longitudinal direction (parallel to the weld length), it doesn’t hold true for transverse loading (perpendicular to the weld length.)
There are three potential failures for transverse loading:
Shear at Leg 1
Shear in the fillet at Θ
Tensile at Leg 2
Since the weld is much stronger in tensile the two main failure likelihoods are the shear failures.
“What is interesting to see is that the maximum stress is not at the 45° angle, which is the conventional wisdom, but the maximum stress is at an angle of 22.5°, even though the size of the throat at 22.5° is larger. The stress at the plane is even higher and that can lead to failures at that angle,” Kelkar explains. “There are many examples of failures at close to 22.5° with transverse loading. This takes people by surprise, and they attribute it to some weakness or local defect, but it is supported by theory.”
Why is this important? Because if you’re assuming that the failure will be at the throat and don’t realize that in transverse loading the weak point could be somewhere else and the stress there could be 20% higher, your factor of safety calculations are not correct.
“The weld will fail at a 20% less applied force than what you had anticipated and that could lead to problems in performance of the weld. It is critical to understand the actual loading, not just thinking the throat is the weakest link,” Kelkar reinforces. He advises that when inspecting field failures to examine cross sections and look at the angle to see if the failure was in transverse loading rather than the shear loading.
Kelkar isn’t aware of any upcoming code revisions involving failures of fillet welds at 22.5° but said this is not a new concept or a new type of failure. In the meantime, should calculations for such weld loads increase the fillet by 20%?
“My analysis shows yes. If you are doing the calculations you have to assume you need the size to be 20% more if the loading is going to be in pure transverse. If you are seeing failures at angles that are not 45° I would definitely recommend using a larger fillet,” Kelkar says but adds there may be times when you might have a factor of safety already built in that covers that 20%.
There are also things to consider with respect to shear stress when fillet welds are completed as a multi pass, with at least two visible passes on top, compared to a same-sized single pass weld. With multi-pass welds, Kelkar explains, you are essentially placing two passes, one on top of the other. Sometimes they’re staggered, sometimes they’re right on top of each other. In filet welds typically they are going to be slightly offset.
“What happens in multi pass welds is the shrinkage from the second pass actually introduces a lot a stress in the first pass and you can actually get some work hardening, even with materials that are not work hardenable. For example, if you’re welding 304 to 304 with 309 filler wire you might actually get work hardening. You might get a weld strength that is stronger than what is specified by the manufacturer for that particular alloy. You are going to get a weld which is stronger but depending on the loading it could fail at a different location than at the 45° angle,” Kelkar says.
Another thing to consider is the fillet weld size. Should the material thickness be taken into consideration to ensure that the weld is large enough to maintain the tensile and shear forces or should the weld size be determined by calculations of force on the joint or structure itself?
Kelkar reminded viewers of his presentation that force is transferred in filet welding only through the weld itself. There is no transfer from one material to the other.
“The size of the weld will decide the strength of the weld. You can make it larger, shorter, smaller. You can use a weld wire, which is stronger or softer. The main constraint when you are welding large wall thickness pieces with fillet welding is the risk that fillet welding that is fairly small in volume can have very rapid cooling. That’s why you have to make sure that the fillet weld you design is not too small. Even though the theory might say you just need an 1/8 inch fillet, the cooling rate when the weld is small is very fast and you might find that it is better to make a larger weld in cases where you are not pre-heating the part. You don’t want to have cracking on very small welds. SMT