Moment frame question 3

[WWTEng]

My question is in regards to the attached steel frame. It’s for a single story building with a lower roof & a higher roof. Since the wind load is not too large, I was hoping to fix the beam-column connection at only one level (high roof in this case) to get the frame to work against lateral loads. The reason I don’t want to fix the lower level is that the beam end moments at that level would too high the connection would be “large”, and I really don’t need the extra stiffness I get from fixing both levels. Frame drift with only roof fixed is acceptable (0.2”)

Could I do this or do I have to fix all the beams since I am using R value for ordinary steel moment frame in my seismic calc (wind still governs for lateral using this R value).

Thanks,

 

[ToadJones]

Interesting....

I have done exactly the opposite in the past but cant remember how I justified doing it.
I fixed the lower connections because the floor beams were quite large and the moment connections were as well + there was substantial concrete diaphragm at the first level b/c the owner wanted an extremely robust slab for document storage and heavy equipment.

The roof beams were cranked beams and I did not feel comfortable using moment connections on their ends.

I did not have two roof levels and there was no seismic.

 

[structSU10]

I would think you could do it as long as you can justify the load gets transfered into the frame at the lower level.

One thing to keep in mind is the K factor for the moment frame columns. I would think an additional lateral load midspan of the column could affect the K factor significantly, as it would basically be a destabilizing force from a buckling standpoint.

 

[DCBII]

Because a significant amount of seismic load is being transferred at the lower roof connection I would think you need to have a certain level of ductility at that level that is representative of the R factor you've chosen.

In moment frames with any amount of ductility the idea is that we want the beam to yeild rather than the column. Placing a point load midspan on your column without dumping it directly into a lateral force system seams to accomplish the opposite effect. It actually reduces the ability to dissapate seismic energy. I'd say this is closer to a cantilever column system in disguise than a moment frame system, since the column is likely to yield or buckle before the beam.

What seismic design category are you in? If you're in C or lower you can still use an R of 3 (which isn't far off from 3.25 for OMF's) and treat it as a "Steel Structure Not Specificallty Detailed for Seismic Resistance".

structSU10 brings up another good point. You will likely get significant P-delta effects in these columns. Pay close attention to your second order analysis methods.

 

[WWTEng]

@ DCB: If designed as ord. moment frame, wind is the governing lateral load and not seismic. SDC is B.

For P-delta, I can only trust the software to do that as I have turned on the option for P-D effects, so it should be accounted for.

 

[SteelPE]

Not to be critical but you need to make use that the software is properly accounting for the P-D effects.... not just I turned on the option in the software.

 

[WWTEng]

@ SteelPE: sorry if my comment sounded too layman (I think it did), I know for a fact that the software is taking into account any P-D effects.

 

[DCBII]

If you look closely at it I think you'll find wind does not "Govern". If you use R=3.25 ASCE still points you to AISC 341, which entails special seismic detailing requirements. Even though your wind load is larger than your seismic load, the only reason it is is because you've provided ductility to dissipate the seismic energy. To get that ductility, seismic detailing requirements will actually "Govern" your design loads.

Look at AISC 341-05 section 11.2a closely. You're still required to design the beam to column connection for 1.1RyMp, which I'm sure is much bigger than the wind load moment. There are a lot of other requirements relating to welding, beam and column limitations, beam to column moment ratios, continuity plates, and lateral bracing Use of R=3 for low seismic design will almost always be more cost effective, since it eliminates the need to comply with AISC 341.

With regards to P-delta effects, it looks from your original post like you're using RISA. RISA provides an abundance of clear concise documentation in their help menu and on their website on how to properly use the Direct Analysis Method for second-order analysis. Merely checking the P-Delta box in the LC spreadsheet may not be enough. Are you using the Direct Analysis Method or the Effective Length Method? This will affect your RISA model. Are P-little delta effects significant? You may need to add internal nodes on members.

 

[frv]

I strongly suggest you follow DCBII's advice. Your seismic base shear will increase by 8.3%, but you are no longer required to do any special detailing.

 

[WWTEng]

DCB, thank you for the detailed comment, its very helpful.

 

[JoshPlumSE]

One good point that DCBII makes is that the connection design will probably be governed by seismic even if the base shear from wind load is higher.

One nice caveat that some don't know about is that you can design the connection for R*Mu instead of 1.1*Ry*Mp for OMF frames. That can be a significantly lower force. There's some good discussion in the AISC 341 commentary about that.

I've done some structures that were reasonably similar to this in the past. Personally, I used to use R=3.0, but then detail it out as an OMF. A little conservative, I know. But, mine were always non-building structures during the post-Northridge days when you added in some extra conservatism just to get the plan check guys to allow you to use moment frames.

 

[WWTEng]

Josh: Did you design & detail the connections yourself or would you have the steel detailer do them for you?

 

[jeff91]

I believe this is an acceptable design approach. It is not required that the lower roof level beams have moment connections to the columns because you still have a stable structure with the moment connections at the high roof level. Due to the pinned connections the lower roof level is "just going along for the ride" with the moment frame formed by the moment connections of the higher roof level. You will induce bending in the columns at the lower roof level but this is ok as long as the columns are being analyzed and designed for it along with all gravity loads being considered.

 

[JoshPlumSE]

That was when I was working for a California engineering office. We almost always designed the moment connections ourselves. Even for projects in lower seismic regions. I know that some of our other offices let the detailers design more of the connections than we did.... At times, that difference in philosophy could be a source of conflict between offices.