Crimped Joist Web Member Slenderness and SJI Table 4.3-1 1

[waytsh]

I am analyzing joists for some new snow drift loading and some of the web members are crimped angles. In the SJI Table 4.3-1 in Section IV for Web Members it says to use a k of 0.75 for kl/rx and a k of 0.9 for kl/ry. Since the equal leg angle is rotated 45 degrees for a crimpled condition which is the rx direction and which is the ry? Is it correct to use a k of 0.75 or 0.9 for this condition?

For what it is worth I am modeling the joist in RISA and do have these web members physically rotated in the correct orientation.

Thanks!

 

[JLNJ]

Maybe you need to use the principal axes rather than the geometric ones.

 

[waytsh]

Just posting to give this a bump and a bit of a follow-up. It appears that SJI has a spreadsheet for reviewing and reinforcing joist members. Based on the capacity shown on the spreadsheet for crimped web members it would appear that they are only checking the crimped web members with a k = 0.75 (kl/r_x)and not using the 0.9 value (kl/r_y). I am not sure how this is justified based on Table 4.3-1. Does anyone have some insight on this?

 

[waytsh]

I should have also mentioned that the spreadsheet is using the angle r_z with k = 0.75.

 

[bones206]

SJI has webinar on this topic next week (

https://steeljoist.org/professional-resources/webinars/).

You might try posing this question during the Q&A.

Here is a good research paper that delved into the K-factors for modern joist design:

https://www.aisc.org/globalassets/c...s-of-web-members-in-open-web-steel-joists.pdf

From the research paper results summary:

For buckling of web members in the plane of the joist, the computed effective length factors
ranged from K=0.50 to K=0.55. This indicates a significant amount of rotational restraint exists
at the ends of the compression web members, which can be shown to be the byproduct of
assuming fully restrained (rigid) connections and having top and bottom chords with in-plane
flexural stiffnesses that are much larger than that of the web members. The overall performance
of K-Series joists was similar to that of LH-Series joists, which indicates that the SJI may be
justified in allowing K-factors for both types of joists to be less than unity (see Table 1). Unless
the connections are specifically designed for the demands associated with using K-factors close
to the rigid-rigid case of K=0.5, the SJI’s value of K=0.75 for LH-Series joists seems reasonable.

A much wider range of out-of-plane effective length factors (K=0.52 to K=0.87) is observed,
with most values between K=0.5 and K=0.75. Given that web buckling out of the plane of the
joist is primarily resisted by the relatively smaller torsional stiffness of the top and bottom
chords, K-factors larger than the in-plane values would be expected. Because the ends of the
web members are almost always located (pinched) between the angles comprising the top and
bottom chords, the SJI’s assumption of K=1.0 (see Table 1) for out-of-plane buckling appears to
be conservative, and similar studies of additional joists may indicate that a value of K=0.85 or
K=0.90 could be used in routine design.

I recently undertook a joist evaluation project and used k_x = 0.75 and k_y =0.90, with the crimped angle physically rotated 45 degrees in the RISA model. The axial capacity was governed by the k_y, so I felt it was a reasonable and conservative result.

 

[waytsh]

I recently undertook a joist evaluation project and used kx = 0.75 and ky =0.90, with the crimped angle physically rotated 45 degrees in the RISA model. The axial capacity was governed by the ky, so I felt it was a reasonable and conservative result.

I agree that would be conservative and that is exactly how I had my model set up. I was getting failures under the original loading and started to dig deeper and came across the SJI spreadsheet. Using the same criteria as shown in the spreadsheet these members checked out.

Also found that SJI gives higher capacities for top chords than if you analyzed strictly under AISC. Is that what you found as well?

 

[theonlynamenottaken]

rx is for buckling in the plane of the truss, rz is for buckling in the direction of least member bending resistance which happens to also be in the plane of truss for single angle web members located between chord angles. ry is for buckling out of the plane of the truss. So, you get to use k = 0.75 for the minor principal axis buckling of your single crimped web member, and k = 0.90 for the major principal axis buckling, and geometric axes do not come into play (for your single crimped web).

 

[waytsh]

ahh, great explanation! That makes perfect sense. Thank you theonlynamenottaken.

 

[bones206]

But the SJI spreadsheet is not checking the major principal axis buckling with k = 0.90, correct?

 

[waytsh]

Correct.

 

[bones206]

I might be missing something here, but shouldn't that be checked as it could be the controlling buckling mode? (out-of-plane buckling)

 

[waytsh]

I probably should not have said that it is not checking it. The spreadsheet is locked and the equations are hidden so I can't see what is happening internally. Maybe it is doing some check to determine which condition is controlling.