Dual System Modeled as Sway/Non-Sway?

[Wajahat Latif]

Hi all, wanted your take on the following. 

Columns in a dual system (shear walls + IMRF) behave as non-sway since the shear walls provide bracing for the moment resisting frame (ACI 318-19, 6.2.5). When this system is modeled in ETabs, should we select 'Sway Intermediate' or 'Non-sway' framing type in 'concrete design overwrites'?

I believe the important IMRF checks of ACI chapter 18 would get ignored if we select 'non-sway'. 

An IMRF system on its own is designed to have an inherent ductility to resist seismic forces through sway action. 

Thanks.

[Badar (BAZ)]

1 hour ago, Wajahat Latif said:

I believe the important IMRF checks of ACI chapter 18 would get ignored if we select 'non-sway'. 

Yes, they will be ignored.

But why do you want to select the non sway option?  Your design moments will not be effected much by the moment magnification factors and P-delta analysis if l/r ratio of columns is small, and if lateral drifts are also with in the limits.

If you are enabling the P-delta analysis option, program will magnify your column moments anyway based on the results of P-delta iterations even if non-sway option has been selected (sway part of moment magnification equation of ACI 318 is still there because of lateral loads).

One should not rely on the seismic checks performed by ETABs. Unless you have specified the actual reinforcement ( the ones in structural drawings) to the program, these results are useless.

[Wajahat Latif]

Thanks. I agree with your points. I was running into a problem with a couple of columns in a dual system building. Based on the ACI 318-19 criteria of slenderness effects (Section 6.2.5), these columns are classified as slender in a sway frame but non-slender in a non-sway frame (based on the different slenderness limits for sway and non-sway frames in this section). When they're behaving slender, the ACI check mentioned in Section 6.2.5.3 (i.e, second order moments < 1.4 * first order moments, screenshot attached) should be applied. For my case, these columns are failing this check. However, I know from manual calculation that this is a non-sway frame and the columns are infact non-slender, therefore, this check can be ignored. 

In ETabs 19, this check is applied whether the framing type is sway or non-sway (columns fail this check in both cases). This makes me wonder, does ETabs perform all slender column checks and calculations regardless of whether the column is slender or not? Any idea how ETabs checks column slenderness? 

[Badar (BAZ)]

3 hours ago, Wajahat Latif said:

This makes me wonder, does ETabs perform all slender column checks and calculations regardless of whether the column is slender or not?

I have also pointed out this observation in this attached post.

3 hours ago, Wajahat Latif said:

Any idea how ETabs checks column slenderness? 

I  have looked at its manual. Here is what I found, and which also addresses why the ETABS magnifies moments even if you set column as non-sway.

ETABS split design forces in columns as M= Ms + (sway factor for amplification of forces)Mns; the component Ms is from lateral loads and Mns is from gravity loads only.

Sway factor is taken as 1, as the program "assumes" that Ms comes from the P-delta analysis. If P-delta analysis is performed, the moment magnification is already incorporated in that, and there is no need to amplify moments. So sway factor is taken as 1.

Once M has been calculated, the program amplifies it further by using a non-sway factor, see the attached.

So, if you have used non-sway option, then this non-sway factor will be 1. But, since you have performed the P-delta analysis for drift related checks, your forces have already been amplified. 

 

 

[Wajahat Latif]

I agree with your observations. The 'sway moment' is amplified by P-Delta considerations, whether the frame is sway or non-sway and whether the column is slender or not. Regarding your following statement:

1 hour ago, Badar (BAZ) said:

So, if you have used non-sway option, then this non-sway factor will be 1. But, since you have performed the P-delta analysis for drift related checks, your forces have already been amplified. 

I think the non-sway factor (catering for deformations between the member ends, i.e., P_smallDelta effects) would be computed in both non-sway and sway cases. No? In non-sway frames, only non-sway factor needs to be applied. However, since we're doing a P-Delta analysis, the sway factor is also inherently present. Similarly, in sway cases, both sway and non-sway factors need to be applied. So, in essence, ETabs is applying both sway and non-sway factors for all kinds of framing. 

As you've pointed out in the post you shared, as per code the moment magnification should not be applied for short columns, i.e., they can be designed on first order moments. However, given the above discussion, it seems these amplifications are being applied regardless if the column is short or slender. 

Now, I just noticed that the non-sway moment factor in "Concrete Frame Design Overwrites" is by default set as 1. We would need to set it to 0 for the program to compute this itself for each member for each load combination (screenshot attached). I suppose if we're confident that our columns are non-slender, then we can avoid the non-sway factor by keeping this as 1. The sway factor would still be applied via P-Delta. 

[Badar (BAZ)]

16 minutes ago, Wajahat Latif said:

I think the non-sway factor (catering for deformations between the member ends, i.e., P_smallDelta effects) would be computed in both non-sway and sway cases. No

Yes, I agree.

 

20 minutes ago, Wajahat Latif said:

I suppose if we're confident that our columns are non-slender, then we can avoid the non-sway factor by keeping this as 1. The sway factor would still be applied via P-Delta. 

Yes!