Limit state checks for a steel angle bearing connection design 1

[oengineer]

I am working on verify an angle connection that is being attached to a concrete wall in order to support a wide flange beam. The W-Beam is resting/beaning on top of the angle.

I have attached an example of the situation I have in the link ( the red clouded angle is the connection in question):

https://files.engineering.com/getfi...-96c5-28fbaed7ff03&file=angle_connection_.pdf

After checking the actual angle for moment stress, shear stress, and deflection I wanted to verify what the additional limit states that need to be satisfied for an angle connection.

Section J4 in the AISC Manual talks about limit states to be considered for connecting elements ( please see images below).

For the condition shown in the link, it appears that "Strength of Elements in Shear", "Strength of Elements in Compression", and "Block Shear Strength" need to be checked, per AISC Section J4. Please feel free to comment if there are any other special requirements for angle connections.

What about "flexural rupture" for the angle connection?

Would these be the only limit states that need to be satisfied or is there another section of the manual that I need to consider since the connecting member in questions is an angle?

Suggestions/comments are appreciated.

 

[KootK]

I wanted to verify what the additional limit states that need to be satisfied for an angle connection.

In many instances it will be the bolt anchorage to the concrete that will govern the design of a connection like this. Is that part of the connection within your scope?

What about "flexural rupture" for the angle connection?

I suppose that one might check flexural rupture at a section taken through the bolt lines.

 

[oengineer]

In many instances it will be the bolt anchorage to the concrete that will govern the design of a connection like this. Is that part of the connection within your scope?

Yes, the bolt anchorage to the concrete is apart of my scope. I have already determined that a Hilti 3/4" Kwik Bolt TZ - SS 316 will work using a min embedment of 4 inches.

 

[r13]

In general, I'll check the seating leg for P*e_h, and H*e_v about the center line of anchors for the vertical leg. A stiffener plate is required if the legs failed the check. I'll also treat the angle as a short beam simply supported at the anchors, and check resulting shear and bending. You don't have block shear problem, but it wouldn't hurt to check bearing strength at the holes.

 

[palk7 EIT]

retired13, when you mentioned you would check the seating angle capacity what would be that resistance property that seating angle should be checked for or simple terms how to find the moment resistance of that seating leg alone in L-angle for the load P*eh ?
Thank you !

 

[JLNJ]

I think Kootk was alluding to the prying action in/on the bolts. When the connecting bolt is located in the lower part of the connecting angle (as it would normally be based on typical angle gauges) the prying force can be very significant. To a large extent it depends on your assumptions on the location of the force application.

The prying often controls the bolt design and the flexure in the angle. The rest is likely just gilding the calculation lily.

 

[r13]

palk7,

 

[palk7 EIT]

Thank you retired13, I am so sorry I think I should have been specific in what I wanted to ask, so that moment due to P*e is to be checked against the bending moment resistance of that L-angle between the fastener spacing and when it fails in that scenario we need to provide stiffeners?

 

[r13]

palk7,

The check assumes the top flange as a cantilever to assure the stiffness of the angle flange. If failed, for any reason that the thickness can't be increased, a stiffener should be provided. When design a seated beam support, be thorough and conservative.

 

[oengineer]

I think Kootk was alluding to the prying action in/on the bolts. When the connecting bolt is located in the lower part of the connecting angle (as it would normally be based on typical angle gauges) the prying force can be very significant. To a large extent it depends on your assumptions on the location of the force application.

The prying often controls the bolt design and the flexure in the angle. The rest is likely just gilding the calculation lily.

Would prying action really be at play in this connection? I ask because the force being applied to this angle is not a tension force. It's more of a shear bearing force, in my humble opinion.

 

[jayrod12]

It is a shear force. Being applied at an eccentricity. So yes, there will be prying. Recheck your anchorage calcs including the eccentric moment. 4" embed may not suffice.

 

[PROYECTOR]

After checking the actual angle for moment stress, shear stress, and deflection I wanted to verify what the additional limit states that need to be satisfied for an angle connection.

You must check de following limit states:

1- Beam web local yielding (AISC 360, J10.2)

2- Beam web local crippling (AISC 360, J10.3)

3- Flexural yielding of the outstanding angle leg (OSL):
To determine the eccentricity, you must determine first the effective bearing length (lb). The lb-distance is determined from the limit states of web local
yielding and web local crippling (J10.2 & J10.3), but not less than k_design and lb + setback, not greater than OSL length. To accommodate beam length tolerance, it is recommended to use a 3/4in set back in calculations (see attached figure).

4- Seat angle shear yielding.

5- Bolt Bearing and Tearout on the Angle.

6- Check anchor bolts according to ACI 318 Chapter 17.

For the condition shown in the link, it appears that "Strength of Elements in Shear", "Strength of Elements in Compression", and "Block Shear Strength" need to be checked, per AISC Section J4. Please feel free to comment if there are any other special requirements for angle connections.

What about "flexural rupture" for the angle connection?

Block shear limit state does not control due to the presence of the OSL. I think flexural rupture doesn't control either, but you can assess this limit state by assuming a critical section on the OSL bolt line. Depending on the lb-distance, the eccentricity can be very small or even negative, so flexural rupture would not apply.

Finally, the end of the beam must be restrained against rotation about their longitudinal axes. In my opinion, it is not recommended to assume that the grating provides adequate lateral bracing to the top flange of the beam. A simple solution is to add an stiffener at each beam end welded to the beam web. Another common solution is to provide a stabilizing angle near the top flange, but I think this solution is not easily applicable in this case.

 

[JLNJ]

If the clip is fixed to the beam, then there is just shear at the clip/concrete interface.

It is more conservative (and likely more correct) to consider the clip pinned at the beam. In this case the clip experiences shear and flexure at the clip/concrete interface.

P*e = T*d. Often, e is about 2d so T is 2P. The anchors get designed for P in shear and 2P in tension.

[URL unfurl="true"]https://res.cloudinary.com/engineering-com/image/upload/v1593436453/tips/angle_connection_ribm72.pdf[/url]

 

[dik]

Due to the difference in beam stiffness and angle stiffness, I usually consider the lb dimension closer to the end of the beam. The free end of the angle deflects and the load occurs near the end. Any thoughts on that?


Dik

 

[Rabbit12]

dik, I think at some point in the past I read somewhere the bearing distribution of the wide flange to the angle is triangular. I'd think 1/3rd between the end of the wide flange and angle would be appropriate, but would heart burn if it was the end of the beam either.



PCI has a good design example for these types of connections. I'd try to make the angle thick enough to eliminate prying action.

 

[PROYECTOR]

Due to the difference in beam stiffness and angle stiffness, I usually consider the lb dimension closer to the end of the beam. The free end of the angle deflects and the load occurs near the end. Any thoughts on that?

dik, I think it would be a logical approach, but it can be unconservative depending on the angle thickness. Attached is a quite straightforward and reasonably conservative design model for unstiffened seated connections. The design values in the AISC Manual Table 10-5 are based on this design model.

This design example is for a welded seat angle, but the design model is the same for bolted seated connections.

Reference: Unified Design of Steel Structures, 3rd Ed.

 

[KootK]

In my opinion, the considerations are quite different when it is a bearing angle fastened to concrete vs a bearing angle fastened to steel owing to the, usually, highly brittle nature of the connection to the concrete:

1) Designing to the required beam bearing values will be conservative for the beam and the horizontal angle leg in flexure as that leg can just yield as need be.'

2) Designing to the required beam bearing values will not capture the tension, and possibly prying demand on the critical bolt anchorage. In this sense, the angle being too thick and too stiff can be part of the problem. Your max beam slope is usually at the ends so, if your angle leg doesn't yield, there's every chance that the reaction is really delivered out near the tip of the angle leg.

3) To design to the required beam bearing length properly, you'd need to also consider an over strength / plastic yielding condition in the angle which I've never actually seen any body do in real life.

I think Kootk was alluding to the prying action in/on the bolts.

Most definitely.

Due to the difference in beam stiffness and angle stiffness, I usually consider the lb dimension closer to the end of the beam. The free end of the angle deflects and the load occurs near the end. Any thoughts on that?

At the connection, I believe that the parameter of interest is not stiffness as compared section to section but, rather, slope. Since the beams' slope at the end will be a function of it's flexibility over it's entire span, that changes the interplay between the angle and the beam with respect to the location of the reaction.

Of course your real load path is probably something like the sketch below util the beam fastener gives way. And that's a good thing as far as the anchorage connection goes. Keeping things mostly pure sure in the anchorage for as long as possible is appealing.

 

[PROYECTOR]

dik, I think it would be a logical approach, but it can be unconservative depending on the angle thickness.

dik, this is what I meant:

In my opinion, the considerations are quite different when it is a bearing angle fastened to concrete vs a bearing angle fastened to steel owing to the, usually, highly brittle nature of the connection to the concrete:

KootK, Of course, that design model is only for the outstanding angle leg flexural yielding check. The seat angle to concrete connection is another matter. For that interface, a more conservative approach must be follow. Either of the following approaches would be appropriate in my opinion:

1) Assume a conservative condition for the eccentricity. For example, the stiff angle bearing stress distribution shown above.

2) Design the anchorage for the design flexural strength of the outstanding angle leg.

3) The lesser of 1 and 2.

In the above figure, "s" is the distance from the support to the critical section (not the setback distance).

 

[KootK]

Yeah, the rigid angle, reverse triangle model is usually what I work to for the anchorage unless I'm desperate or paying special attention to the angle rigidity although even that's got a component of BS to it since there's no reason in the world to expect the stress distribution to be linearly varying.

 

[oengineer]

Attached is a quite straightforward and reasonably conservative design model for unstiffened seated connections. The design values in the AISC Manual Table 10-5 are based on this design model.

This design example is for a welded seat angle, but the design model is the same for bolted seated connections.

Reference: Unified Design of Steel Structures, 3rd Ed.

Thank you for this reference. Since you posted these images, I was able to go back and obtain the design example from this book.


I have a question? Using the specific forces, angle section, and beam section for my situation, if I was to obtain a negative "N min" number when Determining the minimum required bearing length for web yielding would I use the absolute value of the "N min" that I calculated or consider that an "Nmin" equal to zero is acceptable?

I also have the same question regarding a negative "N min" when Determining the minimum required bearing length for web crippling, assuming that N/d < 0.2. Would I either use the absolute value of the negative number or consider my "N min" value to be equal to zero?

I am thinking negative "N min" value would demonstrate that the bearing length is negligible due to the applied load, but i would like to verify/confirm.