Error in Modeling in ETABS

[meistersd]

Hello,

I modelled an 8-storey building in ETABS with a Basement and Lift Core. After modeling and application of all loads, when I perform analysis I can see only the topmost storey deflecting. There doesn't seem to be any deflections in the storey below. As a result of this, the time period in the 1st mode of vibration is also only 0.283s. Not much mass is participating in the different vibration modes also, for both translation and rotation. What could be the probable reason for this erratic behaviour? Have I made any mistake during the process of modeling itself??

I have uploaded a screenshot of the Modal Information below for your reference.

 

[Agent666]

Are you sure you have the appropriate mass source defined?

https://engineervsheep.com

 

[meistersd]

Yes. I checked again and I think I have defined the mass source correctly.

For your reference, I will upload a screenshot of the deflected shape in 1st mode.

 

[meistersd]

I found the problem. Seems I didn't assign diaphragms properly. When I delete the diaphragms, the behaviour of the building looks appropriate. The problem now is how do I correctly assign diaphrams to each floor level? Should we take any special care if there are shear walls??

 

[Agent666]

Select the shell elements at all levels and apply the diaphragm. It should be that simple. But hard to say because you didn't go into how you applied them previously if that's what went wrong...

https://engineervsheep.com

 

[meistersd]

Sorry about that. I had previously assigned diaphragms by selecting all the joints at a particular floor level. I will try by only selecting the shell elements now!

 

[JoshPlumSE]

So, you didn't really have any REAL modes for your structure in that first image. You had very low mass participation, which suggests that these modes are LOCAL modes instead of GLOBAL modes for entire structure.

I'd suggest trying to use the Ritz Vector solution instead of the regular modal results. Ritz vectors are generally biased in a way that should discourage local modes.

 

[Agent666]

meistersd, there shouldn't be anything different with selecting the points at a level and assigning the points as diaphragm. But if you do the shell floor elements, there is a property of points that inherit the assignment of the diaphragm from a connected shell. That's the best way I think to make sure (if you mesh the floors further or add new beams for example) that any new points automatically have the diaphragm assignment without needing to reassign it.

The additional thing is, if you don't assign the shells to a diaphragm, the diaphragm assignment is used for determining the tributary mass for assignment of accidental eccentricity loads, so you're potentially not allowing for all the mass intended when automatically allowing for accidental eccentricities.

Joshh, I think there is more to it than that, as the modes have periods of 0.014 seconds (super stiff). So if the real translational modes are shorter than that for the structure shown I'll eat my hat.

meistersd, you should however increase the number of modes until you hit at least 90% mass participation. I'd second Josh's recommendation for using Ritz vectors instead of eigenvectors, generally using Ritz vectors also have the benefit of greatly increasing the analysis speed with respect to finding the modes which can be quite significant if you're dealing with a large model with many degrees of freedom. Also make sure your structure is not ill conditioned, review output analysis log for warnings to this effect. Etabs is good, in so far as it will solve with as many errors as you like and still give you a non-sensical answer in some cases.

https://engineervsheep.com

 

[JoshPlumSE]

I think there is more to it than that, as the modes have periods of 0.014 seconds (super stiff). So if the real translational modes are shorter than that for the structure shown I'll eat my hat.

Definitely.... I was only looking at the first couple of modes. Something like .287 seconds.