Analysis of an Existing Moment Frame

[KHoff]

I am working on a project that includes structural modifications and a small addition to an existing retail building. Part of the structural modification includes moving a braced frame approximately 10 feet, and therefore increasing the lateral load on an existing interior moment frame. In my analysis of the existing moment frame, I found that the additional lateral load is not a concern but two beams appear to be failing from gravity loads (dead load + snow load). One beam is at approximately 200% capacity and another is at 150%. The design capacity is controlled by lateral torsional buckling due to the negative moment and the bottom flange being unbraced for the full length of the beam. I reached out to the engineer responsible for the original design, he indicated that these beams were assumed to be braced at the inflection points. This building was designed under the 2003 IBC (and AISC 335-89s1) so at the time that was acceptable practice per code. Obviously, current codes state that the inflection point should not be considered a brace point (per Appendix 6).

My question is whether or not this issue warrants making structural repairs to the existing moment frame. The frame was adequate based on the applicable code at the time of the original design. Am I required to bring the frame up to current code requirements? More importantly, does this condition actually present a safety concern for building occupants? To further complicate the matter, the moment frame extends into an adjacent bay in the building that is currently occupied by a tenant. Making repairs would require our client informs the tenant and has them close down until repairs are made.

I appreciate any opinions or thoughts on this issue. As always, public safety is the primary concern.

 

[WARose]

The code lets you make alterations (without modification) as long as you are not driving up the load by 5% (for gravity) or 10% (for lateral). (See Chap. 34 in IBC 2012.)

In your case, sounds like you are doing more than that so you may be stuck with doing some beefing up. And I'd definitely address the unbraced length issue while doing this.

 

[KHoff]

I am above the 10% increase allowed for lateral load, so a detailed analysis is required. The analysis showed that the increase in lateral load is not a problem, but the unbraced length at the bottom flange is a potential problem based on the current codes.

 

[WARose]

I think the increase puts you under current codes and ergo the unbraced length without considering the inflection point.

Putting aside codes for a moment, that isn't good practice anyway (i.e. considering a inflection point braced).

 

[jayrod12]

Does installing a bottom flange brace as close to the existing tenant space without inconveniencing the current tenant do enough for you? Technically if you were above to brace the beam at mid-length, would that improve the LTB capacity enough that you could get it to calc out?

 

[KHoff]

jayrod - One of the beams is entirely above the current tenant's space. Any repair would require disturbing their space. A single brace point at the midspan is enough to get the beam to work.

WARose - I don't necessarily disagree with your comments, but to play devil's advocate, I think there are some questions that need to be asked. There are a lot of buildings out there that were designed based on the assumption that an inflection point is a brace point as that was common practice years ago. Does this mean that any time one of those buildings is renovated/modified braces need to be added? This building has already been standing for over 10 years without any problems. And I think it is important to remember that the "failure" is based on gravity loads, which in this case are not changing. Again, I don't necessarily disagree, but these are things that I think warrant some consideration/discussion.

 

[WARose]

WARose - I don't necessarily disagree with your comments, but to play devil's advocate, I think there are some questions that need to be asked. There are a lot of buildings out there that were designed based on the assumption that an inflection point is a brace point as that was common practice years ago. Does this mean that any time one of those buildings is renovated/modified braces need to be added? This building has already been standing for over 10 years without any problems. And I think it is important to remember that the "failure" is based on gravity loads, which in this case are not changing. Again, I don't necessarily disagree, but these are things that I think warrant some consideration/discussion.

How many buildings really see the design loads (especially when it comes to live)?

There are also a lot of other safety factors that cover that too. For example, when I first came into this business, a lot of people used that as a braced point.....but they almost always (also) used Cb=1. (Every time.) So it kind of balances out.

No question that using a inflection point as being a brace point is bad engineering. Yura himself <insert the sound of thunder here> has said this. (Including at a seminar I attended years ago by him.)

 

[KHoff]

Absolutely agree that using the inflection point as a brace point is bad engineering nowadays. The fact that it was added to the AISC 360 spec makes that clear.

Maybe this is the more interesting question. If braces are not added and there were to be an issue with the frame in the future, who owns it? The gravity loads and unbraced bottom flange are the same as they were in the original design, my work is not changing that. My design is slightly increasing lateral load in the frame, but analysis shows that the failure mode is LTB from gravity loads.

 

[jittles]

I don't see how anybody but you would own it at that point.

Your modifications increased the loads to the frame by more than the 10% limit and if you don't analyze it and bring it to current code as the IBC dictates, you would be in violation of the IBC. The original engineer presumably designed correctly to meet the code at the time.

Original designer designed to code, your design purposefully violates the code. Regardless of the fact that the failure might have occurred under the original design without your modifications, you're still 100% on the hook for the thing.

 

[WARose]

Maybe this is the more interesting question. If braces are not added and there were to be an issue with the frame in the future, who owns it? The gravity loads and unbraced bottom flange are the same as they were in the original design, my work is not changing that. My design is slightly increasing lateral load in the frame, but analysis shows that the failure mode is LTB from gravity loads.

To me, that falls back on the 5/10% rule......once you've crossed that line, you are then under the current bracing requirements. Sounds like you aren't (for gravity loads).

However, in the interest of safety, it might be a good idea to advise the client of the situation (in writing) and recommend a fix. If you are already doing modifications to this building.....adding some bracing shouldn't be that big of a deal.

 

[KootK]

and AISC 335-89s1

Does that document explicitly say that inflection points may be used as bracing? I was not aware that was the case. I've always thought it was just a first principles thing that we'd been wrongheaded about.

The design capacity is controlled by lateral torsional buckling due to the negative moment and the bottom flange being unbraced for the full length of the beam.

Any chance these beams would work as simple span member with the top flange braced ~continuously?

 

[BAretired]

A beam with an applied negative moment at one end, bottom flange unbraced and zero gravity load can buckle even though it works as a simple span with top flange braced.

While an inflection point is not a braced point, there is a variation of moment along the span. There is zero moment at the inflection point, positive moment on one side and negative moment on the other side. An analysis of LTB should take into account moment variation throughout the unbraced span. Has that been done?

BA

 

[KootK]

A beam with an applied negative moment at one end, bottom flange unbraced and zero gravity load can buckle even though it works as a simple span with top flange braced.

It can indeed buckle but, in many cases, I would submit that it would be non-catastrophic as the beam could turn into a simple span member post-negative region buckling. Obviously, this strategy can't be used under lateral loads that would cause non-redundant moments in the beam/column joints. My interpretation of OP's problem is that "failure" is occurring under load combinations that are gravity only.

 

[JoshPlumSE]

This is the sort of thing that drove me nuts when I was tech support at RISA talking to engineers on the subject. The beam was designed under the 2003 IBC, that is long after AISC started beating the drum about inflection points not being a point of bracing. So, it was poor engineering at the time (IMHO). The engineer may have followed the normal "standard of care" for the industry. But, it still really bugs me because I had a lot of talks with engineers about the subject that were along the lines of "I've been doing it this way for years. I scoff in your general direction for suggesting that I should consider doing it another way."

That being said, the reason why so few of these beams developed problems is because they also usually assumed a Cb of 1.0 when it really should have been higher. Something like 2.3. Does that help the beam enough that it works?

What yield was used in design 36ksi, or 50? That's something else that helped with a lot of those beams. Where engineers thought they were getting 36 ksi steel, but were really getting a lot more.

 

[JoshPlumSE]

KootK's response is kind of like the concept of AISC appendix 1 (design by inelastic analysis). There are some rules in that appendix about when or how much of the negative moment can be re-distributed. Honestly, I've never used it for design. But, that could be a good code approved method of justifying the existing beam.

 

[KootK]

While I appreciate JP's support, I'm not sure that I agree. Plastic design is dependent on the formation of stable plastic hinges which would probably increase the demand for bracing. What I'm really proposing is yucky, unstable hinge formation.

It doesn't take much to brace a beam flange. Perhaps someone could enter the space at night and install some bolted kickers in field drilled holes. This could be like your own, super dull version of Citicorp. Sleep easy.

 

[JoshPlumSE]

Going through the appendix (which I've still never used for design), it looks like KootK's correct. The requirements that I see for re-distribution are merely that the member have sufficient ductility to withstand the hinging. I was only thinking of the b/t ratios to limit local buckling.

But, as KootK alluded to, there is also stuff in there for limiting unbraced lengths to ensure the ductility. Which kind of limits its usefulness for this situation. Sorry for introducing a red herring into the discussion!

 

[sandman21]

You could never use the inflection point as bracing While we are aware that some engineers assumed unbraced lengths based on inflection point locations, this practice was never specifically permitted by the AISC Specification. For example, note that the 1989 Specification states in Chapter F,“braced laterally… at intervals not exceeding…” and “the laterally unsupported length of the compression flange does not exceed…” An inflection point does not qualify as either. Similar language can be found as well in older versions of the AISC Specification.

 

[KHoff]

I don't believe the 1989 spec stated that you could assume an inflection point was a brace point, but I know that assumption wasn't prohibited by the code until AISC 360.

In my analysis I did calculate Cb based on moment variation, and I am using 50 ksi yield strength. At this point I believe I have exhausted everything in the code that could help prove these beams work. It's possible that the beam is simply getting by on factor of safety.

At this point I am leaning toward notifying the owner of the issue and providing a fix. If they proceed with moving the existing braced frame (therefore increasing lateral load on the moment frame) I will require they apply the fix. If they don't move the existing braced frame I will recommend they still apply the fix, but at that point the structure is unchanged in my project.

 

[Agent666]

While the inflection point is not a point of restraint, often the next purlin or roof joist past the inflection point where the top flange is in compression can qualify as a brace/restraint if connected to the top flange in a manner that prevents twist of the cross section or prevents lateral movement of the top flange (or both).

Hell if the cross member is stiff enough it can also brace the bottom flange in compression to some degree, see the following diagrams from NZS3404 for some guidance for example (critical flange is the compression flange):-

If going down the prevents movement path for lateral restraint to the top flange then something beyond the member needs to support the restraint of the movement (existing roof bracing, connection to something solid like a concrete wall, etc).