ASCE 7-16 15.4-6 - Approx. Seismic Period Follow Up

[duckhawk]

This is a follow up question to the thread from 6/4 about ASCE 7-16 15.4-6

The commentary about the equation says to use lateral forces to determine deflection.

However, wouldn't you need to consider dead load in that load combo as well when deflection is calculated?

Asking because I've ran it with just lateral forces and lateral forces+dead and you get different deflections.

You can't act like dead isn't there and calc deflections based on lateral only, can you? It just doesn't sound right. Comparing this method with RISA's Eigensolution, it sounds like in RISA's general reference that mass is included.

In the example below I'm not quite sure if the seismic weight was considered in the same load combo when deflection are calculated.
https://res.cloudinary.com/engineering-com/image/upload/v1639177951/tips/Rayleigh_Method_m3ghfj.pdf

Appreciate it!

 

[driftLimiter]

It is odd that you are saying gravity loads cause lateral deflections. The only reasons I can think of that that would happen are if you have a lot of P-delta or If your lateral system supports dead load (thrust typically from non-vertical gravity load resisting elements).

First I think you need to investigate and understand WHY gravity loads are effecting your lateral displacement.

But yes lateral stiffness is simply F/d. Chopra's text for example has many instances where the mode shapes are calculated based on a simple determination of the stiffness of the MLFRS.

 

[HTURKAK]

....Asking because I've ran it with just lateral forces and lateral forces+dead and you get different deflections.

Could be two reasons ;

- P-Δ effect , ( as driftLimiter stated)
- If the structure and loading is not symmetrical .
Otherwise gravity loads will not cause lateral deflections.

 

[BulbTheBuilder]

You can't act like dead isn't there and calc deflections based on lateral only, can you?

Most calculations are based on theories and concepts and same applies to this. You want to determine the frequency of the structure - behavior of structure under lateral loads not the behavior of structure under dead load.

Based on this, the analysis is isolated solely for lateral behavior since we know there's a relationship between frequency and deflection. Same way you won't worry about lateral loads when checking for vibrations on a floor.

 

[bIockhead]

Probably getting too into the weeds, but delta would be a function of stiffness, which depending on the material would very much increase with the presence of more axial (dead) load. The example linked in the comment above looks at wind (not what I think OP is getting at - could be wrong) but I would argue that flexural and shear stiffness should be under the same axial loads used for the seismic weight (Dead + 0.25 x snow in Canada)

 

[duckhawk]

Appreciate everyone's responses. The structure is a trussed tower with equipment on it.

 

[JoshPlumSE]

You can't act like dead isn't there and calc deflections based on lateral only, can you? It just doesn't sound right. Comparing this method with RISA's Eigensolution, it sounds like in RISA's general reference that mass is included.

We're talking about using the Rayleigh method to calculate the natural frequency of the structure. Right? In this case, the only thing that matters is the lateral forces applied, the lateral deflection and the mass / weight at each level.... for the most part.

But, take a look at that example I wrote up for the RISA training manual that was shared on the previous thread. I demonstrated how the RISA eigen solution natural frequency is imperfect because the program (at the time) didn't account for the mild softening effects of P-Delta. In that example, I turned on P-Delta for the load case I used for the Rayleigh method. When I did that I got a frequency difference of about 3%, even though the stiffness difference was more like 6%.

So, it can make a difference. It's just not as significant as you might think.

 

[bIockhead]

We're talking about using the Rayleigh method to calculate the natural frequency of the structure. Right? In this case, the only thing that matters is the lateral forces applied, the lateral deflection and the mass / weight at each level.... for the most part.

But, take a look at that example I wrote up for the RISA training manual that was shared on the previous thread. I demonstrated how the RISA eigen solution natural frequency is imperfect because the program (at the time) didn't account for the mild softening effects of P-Delta. In that example, I turned on P-Delta for the load case I used for the Rayleigh method. When I did that I got a frequency difference of about 3%, even though the stiffness difference was more like 6%.

So, it can make a difference. It's just not as significant as you might think.

Thanks @JoshPlumSE, I printed it off the first time I saw it and went though to check my manual calc process. I spend a bit of time pleading with designers to check the period as it often makes a big difference - depending on the structure layout. Appreciate your contribution

 

[JoshPlumSE]

Whether or not your software accounts for the P-Delta softening effect in the eigen solution / natural period calculations can be tough to tell. When I was at RISA, the company exclusively used the "secondary shear" method of accounting for P-Delta. The problem is that this doesn't affect modal solutions....

Most other companies use a "geometric modification" to the stiffness matrix. This almost always is the best way to account for P-Delta effects in modal solutions.

Since I left RISA, I was told that they added a "geometric modification" option for P-Delta. But, I was also told that this method was ONLY applied to static solutions. That, for some perplexing reason, they chose not to include it in the modal solutions?!? That is the BIGGEST reason why anyone would want to use this method. Why even bother with it if you're not going to do it right?!

It's the kind of thing a company driven by the Marketing staff not the engineering staff does.... You know, "I need to be able to tell our users that we have this feature. But, the feature doesn't actually need to work for me to get the sale."

Caveat:
I worked for RISA for something like 16 years (early 2002 to late 2017), that was back when it was owned by the founder. I was essentially the VP of Engineering by the end of my time there (not my official title, but it's what my role really was). As soon as the company was sold to the hard to spell investment company from Germany (I think?), I was forced out. I no longer have hard feelings about leaving the company as I view it as a blessing in retrospect. However, this experience very much affects my view of how competent and ethical the management there is. I also work for their largest competitor (Computers and Structures Inc, which writes SAP2000 and ETABS). Therefore, any criticisms of the company that I make should be viewed with a grain of salt.... I'm not exactly an impartial observer.

 

[duckhawk]

Appreciate you sharing those experiences. It brings a very polished software down to what it really is - something made by people.