How to calculate drift of response spectra analysis?

[rejoicer]

When perform reponse spectra load case in X direction, there is always some forces in Y direction. Sometimes the forces in Y direction is pretty large or even close to force in X direction. Therefore the drift in Y direction is close to the drift in X direction.

So how to check the story drift? Just check drift in X direction by neglecting drift in Y direction or use squre root of the sum of suqres of X and Y directions?

Thanks

 

[zeemas]

what I understand is that when we define Response spectrum cases we already define them the method for summing up the modes either by CQC or SRSS method therefore the drifts we get are already combination from some of the modes shapes.
What I do usually is that I define Response Spectrum function with the name SX+SY, SX+0.3SY and SY+0.3SX and check the drifts on this by checking Story drift option of ETABS. If i get problem at specific point then I just go to that point and try to check out in more detail.
Your comments are welcome.

Thanks
Zeeshan

 

[lisica]

Hi...
The earthquake waves are comming from one direction. You don't need to sum the influences from two ortognal directions.
Just analyse the influnce in two main directions (X i Y in most cases) separately.

 

[lisica]

However, it depends on your local code. Eurocode 8 demands taking 30% of influence from ortogonal direction.
So, it would be X+0,3Y or Y+0,3X.

 

[xnyzhangshulu]

hi, lisica.
you mentioned that we needn't to sum the influences from two orthogonal directions. However, the earthquake wave is 3D and can not accurately predicted. So many seismic records show that the ratio of PGA for the two orthogonal directions is approximately 1:0.85. By square root of the sum of suqres of X and Y directions with the above factors, we could get the orthogonal effects, Ex+0.3Ey or Ey+0.3Ex.
About the story drift check, I do agree with Zeeshan's viewpoints. Please refer to Chapter 4 Commentary of NEHRP 2003 for more details.
Waiting for your comments.
Shulu Zhang

 

[lisica]

OK, I agree with you.
My code requires using Ex+0.3Ey or Ey+0.3Ex, so I am using this relation.
Thank you for your post!