Expert Reactions to the Directionality Increase for Fillet Welds to HSS

Peer Perspective: Dr. Jeffrey Packer, University of Toronto

We sat down with Dr. Jeffrey Packer, a professor at the University of Toronto, to hear about his most recent research and get his thoughts on the directionality increase for fillet welds to HSS.

Q: You have an impressive body of work, Dr. Packer. Can you tell us a little bit about your most relevant background experience and how you got to this point in your career?

A: Well, I got into this hollow structural section business during my graduate work in England. I’m originally from Australia, so after doing my bachelor’s degree there, I went to England for my graduate studies. After earning a master’s degree, I ended up doing a lot of plastic analysis and plastic design work and doing a Ph.D. in hollow structural sections, which was quite novel at the time in North America.

I came to the University of Toronto to do research and teach and naturally fell into continuing my work with hollow structural sections. I’ve been doing this for quite a long time now, and I do a lot of collaborative work with the American Institute of Steel Construction, Canadian Institute of Steel Construction, American Welding Society and of course the Steel Tube Institute. There’s a group of us now who are very involved in research and development work in hollow structural sections and furthering the use of it in practice and in codes.

Q: How would you describe the 1.5 increase in the simplest terms? Can you give us an overview of what this means for industry professionals?

A: Going back decades ago, the strength of fillet welds was just calculated as fracture along the throat area, so it would be the throat thickness multiplied by the length or the effective length of the weld. Then we had a body of work that showed the direction of loading on a fillet weld could produce an increase in strength but a decrease in ductility. The increase in strength was particularly of interest to code committees because the welds could be deemed stronger, but these tests were done on lap splice connections — all of them. Under those circumstances, they demonstrated that you could get this increase in strength depending on the direction of the load relative to the axis of the weld, but importantly in lap splice shear connections.

We eventually found that when it came to test tubes, pulled in tension perpendicular to another object and the fillet weld was on one side, on the outside of the tube, it would lever and apply a bending moment on the weld as well as an axial tension. So it would pull up and the weld would bend. That actually decreases the strength of a weld. And you can’t put a fillet weld on the other side because you’d be on the inside of the tube. That makes a real perpetual problem for what they call single-sided fillet welds.

We connected an HSS member being pulled in tension to a solid rigid plate, welded on one side of the HSS with a fillet weld. We would then pull on the HSS and the wall would be pulled up and we would fracture the weld. This allowed us to measure the loads, after carefully measuring all the weld sizes and properties, and we would know that the base on which it was sitting did not have any effect. It was rigid and all the perimeter of the weld would be effective. The direction of loading would have been at 90 degrees to the axis of the weld. Even though it was at 90 degrees, according to that directional strength increase factor, one plus 0.50, we should get a 1.5 times increase. A 50% increase in strength, and we didn’t get that.

Q: Interesting. So how did the team react to this finding?

A: This led us to look into this problem of the levering effect on the weld and the fact that it was causing a local moment, which was decreasing the strength of the weld. And so it would have been much easier if we didn’t have the directional strength increase factor in the first place, but it’s well embedded in all our codes now. Given that, should we use the 1.5 directional strength factor increase for when we’re pulling on a tube welded to a rigid plate? We could show that if we pulled a square or rectangular HSS, welded to a rigid plate, then we would get a slightly different strength increase compared to a round tube.

Q: So that calls into question the difference between square or rectangular HSS and round tubes?

A: We’ve basically knocked out the use of the directional strength increase for fillet welds to square and rectangular HSS for a wall that’s loaded in tension, and we’ve kept it for round HSS for a wall that’s loaded in tension because, in the latter case, we can justify this from a reliability point of view and for a square and rectangular HSS, we can’t.

Q: Was there a reason you decided to do this testing and that it came up that the 1.5 increase may not apply? How did you get to the point that you guys were looking at this?

A: Well, I suppose we’re one of the rare research groups that’s been testing welds to HSS members for decades. We’ve been fitting it to the sort of code criteria at the time and the expected safety indices and reliability indices. We started even in the 1980s with rectangular HSS gapped K-connections and then moved on to HSS T-, Y- and cross-connections. Then we did the rigid plate work and HSS T-connections under moment loading and so forth, and it’s been going on for decades. With this, we’ve been welding HSS to these rigid plates and pulling them in direct tension. And that’s when we found that the directionality factor didn’t apply to a square and rectangular HSS.

Q: All of this is fascinating. Can you tell us about some moments of surprise you had during your research?

A: Well, one of the issues is that with the design of fillet welds, they don’t of course fracture exactly along the effective throat. And yes, that’s the way we design them. Many of the problems we’ve had are kind of coping with the simplifications that we have in our codes here in North America. We had the simple rule that all fillet welds, regardless of how they’re loaded, will fracture along the effective throat. But they don’t. You pull them in tension and they fracture very close to the wall of the HSS that’s in tension. It’s about 22 degrees, not 45 degrees through the effective throat.

So, we’re always grappling with the fact that we’ve simplified the strength of a fillet weld to something that fails in a certain way, regardless of loading. And then we’ve bolted on this little enhancement factor for directional strength, which is kind of a complex formula that was derived for these sort of lap splice connections. Now we’ve boxed ourselves into a corner with this funny fillet weld strength rule in a way, which we’re now adapting to yet another quirk about it being loaded on one side and having this bending put on it. There are several simplifications so that designers can actually get around to designing a simple weld and not take days to do it, and that’s because a lot of engineers are doing concrete design, timber design and steel design but not necessarily HSS design as well. They can’t be expected to invest a lot of time in perfecting special techniques, really. So we’re continually grappling with simplifications and adjusting things so that the simplified rules work. That’s always a bit of a battle.

Q: Thank you, Dr. Packer. Where can we expect to see all of this work in black and white and what can we look for in the future of this space?

A: Virtually all the work that we’ve done for effective lengths and effective section sizes for welds has gone into Chapter K, section K5 of AISC 360. The latest work on fillet welds to rounds, whereby we now have an effective length factor for round-to-round connections, is going to be added to the next edition of Chapter K. That’ll be a new table in Chapter K in the AISC 360, 2022 edition.

One project we’re doing at the moment for AISC and STI is actually looking at what’s called the hidden weld in a rectangular HSS to HSS overlapped K-connection. So when you weld the two branch members in an overlapped K-connection together, down at the bottom you’ve got a weld which can’t be seen. That’s the hidden weld in an overlapped K-connection. We’ve designed some clever test specimens, and Atlas Tube is going to help by providing us with the HSS material. We’re very excited to do another project, which is a joint AISC and STI project. We won’t have findings in time for the AISC 360, 2022 edition because that’s coming to a close soon, but probably in the 2028 edition we hope to have some more helpful news about the hidden weld.

July 2020