Seism actions

disla

Hello everybody.
I'm trying to figure out which is the simplest and most efficient way to implement seismic actions in my model.
I have the spectral acceleration response graph derived from the applicable code.
Thanks in advance.

Victor

You can fix the base and apply the acceleration as a time dependent gravity to the whole structure. Paste the data into the table in the gravity load. There's no acceleration boundary condition input.

disla

Thanks Victor.
I can't find gravity loads. Do you mean to impose “forces”.
By other hand, Seismic actions are mainly horizontal effects and gravity is a vertical component. I suspect it is a complicated problem to solve with FEM.
In my case I should also simulate the fluid-structure interaction as the huge fluid content moves inside the tank.

Victor

Gravity is under Loads & Constraints for the Dynamic response 3D analysis type but it's not there for some other analysis types which might be why you can't see it. You can specify components in all 3 dimensions so you would add the constant real gravity to the vertical component of the acceleration data. Using gravity for acceleration puts the whole structure in the accelerating reference frame (see equivalence principle), which means you would fix the base with zero displacement and allow the rest of it to deform.

I don't know what to do about sloshing water, sorry. I suppose, ideally, you would use FSI which Mecway can't do. CCX has "Shallow water motion" which would model surface waves and looks like the same theory that I've used before for natural modes of liquid sloshing in a tank. But I don't know if that's useful here or if it can be coupled with a mechanical model or not.

Andrea

Maybe you are speacking about two different tecniques for seismic analysis.
1) Use response spectrum (acceleration Vs period) combined with modal analysis and CQC or SRSS combination)
2) Use the accelerogram (acceleration Vs time)

General method is the first above because standard code (FEMA - AISC - Eurocode) gives the spectrum of the seismic action for different locations (peak acceleration, ground type, periods to define spectrum shape) and different levels of ground peak acceleration.

Accelerograms are generally used to study the behaviour of a structure appliyng an historical registered ground motion. You can find many database of accelerograms of historical eathquakes

disla

Thanks Andrea, Victor,

I'm using method 1) Response spectrum (acceleration Vs period). Codes provide analytical formulas to compute equivalent static horizontal force. No dynamic analysis is done.
I'm now prety much convinced that running 3D dinamic simulation or FSI would not be feasable.
I was wandering if it is possible to obtain at least convective and impulsive periods of my tank filled with product to lately compute and apply equivalent forces provided by the code. At least I can contrast the periods provided by the code.

Victor, ¿seems you have some experience with natural modes of liquid sloshing in a tank?. ¿Is it done with CCX? ¿Do you have any reference?

Thanks

Victor

I misunderstood the type of data you have, sorry. What I said was for a time series. CCX has spectral response (*STEADY STATE DYNAMICS) but it's not supported by Mecway so you'd have to write the cards by hand and possibly use CGX for post-processing.

For the fluid sloshing, there's some detail in the "Shallow water motion" section of the CCX manual. Also, here's some pages from S. S. Rao which is what Mewcay's Acoustic resonance 2D analysis type is based on. You can probably use that for fluid by imagining the variables have different meanings. Here's how I think the acoustic variables correspond to fluid ones, but I'm not that confident:

Acoustic -> Fluid
speed of sound -> sqrt(depth * g)
pressure -> vertical displacement of free surface
frequency -> frequency

Here's an example of a 1m*1m*0.5m deep tank. The depth and gravity are specified together as the material's Speed of sound and the solution's modes would be the wave shapes on the surface of the liquid.

65mustang

Disla,

I good source for seismic performance of large vessels would be the following two treatise:
1) The Dynamic Behavior of Water Tanks
2) Dynamic Pressures on Accelerated Fluid Containers.

Both of these sources are by G.W. Housner.

disla

Thanks to all for the hints.

sofien_73

Hi everyone, I want to perform an spectrum analysis according to the EuroCode with MecWay/CCX. I can't use simplified method for this case. Is it possible to perform response spectrum analysis combined with modal analysis with SRSS combination?
Regards.
Sofien

disla

Hi,

¿Is there someone with experience computing elevation and oscillation period of enclosed water waves with the above proposal. I mean taking advanage of the paralelism with the available 2D acoustic vibration analisys :

Acoustic -> Fluid
Speed of sound -> sqrt(depth * g)
Pressure -> vertical displacement of free surface
Frequency -> frequency

I'm not obtaining a good agreement with the expected results.

@Victor. Seems you are considering g=1 in your example?¿?.
Is that correct ?

Any help is appreciated.Thanks in advance.
Disla

Victor

Yes, it looks like I used g=1 m/s to give c = sqrt(0.5m * 1m/s^2) = 0.707107 m/s in that example.

disla

Thank you, Victor ,

I have calculated the first expected oscillating period by hand. I have found the same formula in at least 3 different sources, and they all agree.
For a cylindrical shaped tank, Diameter=20m and depth H=25m, the expected first oscillating period of the water mass should be T=4.6s.

Without going into the detail of the maximum wave height, I do not get a good agreement in that first mode. Internal gives 2.18s

I would appreciate if someone could give me some advice or could reference me some verification examples using this approach. I'm pretty sure of the analytical solution but I can't replicate with FEA.

Thanks.

Victor

Not sure. I tried again using c=1 m/s and only scaling the frequency according to my old reference, but it ended up the same with T=2.18 s.