Static Seismic - Linear Model vs Non Linear Model

Posted by: Ivicom

Static Seismic - Linear Model vs Non Linear Model - 11/02/15 08:27 AM

Hi,

I have a question regarding load cases for static seismic analysis for linear model (no gaps on restraints and no friction) vs non linear model (restraints with gaps, and friction).

I know that there is a cookbook approach given in CAESARII User's Guide (Section 8) which I normally follow for doing static seismic analysis. This approach is based on the segregated earthquake load cases which user has to add in the end to the sustained case for code compliance.

But if we have a linear model, can we add uniform loads(earthquake) directly to the sustained case, like:

W+P1+U1
W+P1-U1
W+P1+U2
W+P1-U2

and then check code compliance or we still need to do it in the way that is defined in section 8 of the user's guide.

Why do I ask this. Recently I've received stress report with load cases for static seismic defined in a way like the ones above. Two uniform loads were set in the model, U1 and U2 and for occasional stress code compliance they were added directly to the SUS case.

Is this way of calculating OCC stress correct or not?
Posted by: Dave Diehl

Re: Static Seismic - Linear Model vs Non Linear Model - 11/05/15 10:44 AM

My opinion:
If the model is linear:
The maximum nodal stress from your two load cases "W+P1+U1" and "W+P1-U1" will be shown in L3 in the following load sequence -
L1: W+P1 (SUS)
L2: U1 (OCC)
L3: L1+L2 (OCC) [SCALAR]
While this will produce the maximum "sus+occ" stress, L3 will not necessarily provide the maximum structural response during teh event (e.g., deflection and load). That would come from "operating+U1" and "operating-U1".
Posted by: Ivicom

Re: Static Seismic - Linear Model vs Non Linear Model - 11/06/15 01:49 AM

Mr. Diehl, if I understood you correctly, we have to define seismic case separately (L2) in order to add to the sustained case for code compliance because we use scalar operation (L1+L2).
In case we have linear model we can't just add them directly (W+P1+U1) becasue here we have algebraic operation with defined loads.

That means that the following load cases aren't correctly set up, at least if we consider "sus+occ" (this is the load case set up that my customer gave me for checking):

1 (HYD) WW+HP
2 (OPE) W+D1+T1+P1
3 (OPE) W+D2+T2+P1
4 (OPE) W+D3+T3+P1
5 (OCC) W+P1+WIN1
6 (OCC) W+P1+WIN2
7 (OCC) W+P1+WIN3
8 (OCC) W+P1+WIN4
9 (OCC) W+P1+U1
10 (OCC) W+P1-U1
11 (OCC) W+P1+U2
12 (OCC) W+P1-U2
13 (SUS) W+P1
14 (EXP) L14=L2-L13
15 (EXP) L15=L3-L13
16 (EXP) L16=L4-L13
17 (EXP) L17=L3-L4
Posted by: Dave Diehl

Re: Static Seismic - Linear Model vs Non Linear Model - 11/06/15 03:38 PM

My guess is that, for a linear system, the maximum stresses from your L9 & L10 will match the (L1+L2) summation that I suggest.