Seismic Load path (Static & Dynamic)

[alphaxy]

Sirs / Mam's,

Great day to all! I need your help..I need a brilliant idea with the attached file that Im dealing with right now. I have this confusion that you may have the idea of what is the correct approach

In figure A, the normal pressures of static soil pressure and water pressure were applied in the chambers (showing elevation view).

In figure B, the seismic of liquid based in ACI 440 accelerates in the X direction. Will it be the same direction with inertia force of wall? The diagram in ACI 440 seems in the same direction (with impulsive, convective & inertia forces) and in my model, I've used this approach.

In figure C, the seismic of liquid opposes the direction of earthquake..In relating to a fast car accelerating at a certain speed, the forces will be opposite and if stops, then it will be of the same direction. Do u agree with this? (The seismic of liquid will oppose the direction of earthquake)

As you may see, In figure B, I applied the dynamic pressure of soil in the same direction of earthquake by mononobe okabe approach but the basis of incorporating the seismicity of soil will be the node displacement of walls a value of 0.001 (active dense type) soil x height = which arrives to a value of 4.6mm allowable value. Now we compare the actual (based on staad model) vs allowable and if it exceeds the allowable, we will incorporate the seismic of soil but if not, it will be static perssure of soil only. (Do we really have the basis when to include seismic of soil?)

Another question is, the passive force due to soil is present in actual..Do we have to apply passive seismic of soil?

I hope you can give some idea. Thank you very much in advance

Thanks & Regards

Alphaxy

 

[ishvaaag]

For now just some related articles

http://www.dec.fct.unl.pt/seccoes/S_Estruturas/Dinamica/Util/SeismicAnalysisTanks.pdf

http://www.iitk.ac.in/nicee/IITK-GSDMA/EQ04.pdf

http://www.scribd.com/doc/22394223/Guidelines-for-Seismic-Design-of-Liquid-Storage-Tanks

http://www.csasc.org/2005/journal/ae03.pdf

http://db.nzsee.org.nz/PCEE/2003/View/Paper007s.pdf

http://openpdf.com/ebook/quasistatic-approach-liquid-filled-storage-tanks-pdf.html

http://www.osti.gov/bridge/purl.cov...CFA67D5FEB2500CC694AFE?purl=/10164109-OYLXN2/

Will try to be more specific later...

 

[ishvaaag]

Now, on the matter. I don't see that for a general case anything less that a FEM model of difficult assumptions and input can give areasonably accurate portrait of what happens in this kind of structure. So you would have to consider how to model the soil or abstract it by forcing functions, what damped spring to put at outer boundaries, then the same towards the interior with CFD and think of not overflow to keep the model within some constraints. So an almost impossible task for most except the most expert in CFD and mechanical FEM.

Make enough models like this (or measure behaviour in a significant number of cases) and you have something with what to start to portrait the general behaviour of these structures, and then one way or another reduce it to synthetic formulations.

This is what, within the scope of the respective codes, these must have within.

It is reasonable to expect that codes referring to some specific situation, for loading like these, where the direction of the force is unknown, are expecting you to apply the stated force in every direction, typically 4 or 8 and at 45 deg directions. As commonly, a code in this intent will be expecting you combine the stated equivalent forces in such a way as to produce the maximum solicitations. This will occurr for walls with a liquid dynamically pushing when the wall is accompanying the movement of the liquid ... yet remember you have the 8 directions to consider.

Also, for a liquid like water it is quite likely the maximum dynamical pressures will be well characterized and not much affected by the relative movement against or with the liquid; maybe for liquid mercury things would be different.

When the liquid pushes against the ground you will have less solicitation than when an above ground wall, since the ground will be restraining the outwards movement. So for these cases one if not making the sophisticated models first quoted quite likely would consider a wall in an above ground situation even if it is underground, for the ground (being the active transmissor of the excitation) may be receding from the wall at a higher rate of movement than one stiff structure itself.

Then for the inwards action of the earthquake on the walls you estimate it in whatever appropiate manner, and, conservatively, assume there's no liquid inside.

This resumes my view on the subject.

 

[ishvaaag]

Only to add (and partially correct what above) that for an inner wall the maximum equivalent static effect will always happen when the displacement of wall is at maximum and the push of the liquid in synthony with it, i.e., both assumed towards one side; when the wall is against the liquid movement the energy of the liquid first has to invert the contrary moment standing at the base from prior forces, and so the two effects are not adding to form a maximum absolute value of the moment.

 

[alphaxy]

Ishvaag,

what is CFD?

If the structure will head toward in "say" positive X direction..how will the soil react to this loads considering the static forces of soil only. (say let us ignore first the seismic effect of soil)

I am not sure if the passive force will be true at this point..as constant Kp is a huge value compared to active force, this really gives me a confusion. The forces from earthquake and sloshing of liquid will make the soil react (that is passive) but say this force is just 1kN, but then passive pressure is too large for 1kN. Do you have any idea with this?

I have spring supports in external side walls.

kindly refer to my sketch. Do I need to apply passive force in my model? or what will happen is the structure, soil in left and right will just move as one in seismic event? but what if the soil level in the other side is lower than the other side? or there is no soil on the other side..

If the soil in left and right will move as one, then, perhaps there is no active nor passive force to be applied in structure.

^.^

 

[dgillette]

I don't think I'm with you on this.

If, as in figure B, the foundation is accelerating to the right, it's the wall on the right that is subjected to the dynamic active load from the inertia of the soil. That's most easily modeled by the Mononobe-Okabe modification of Coulomb active pressure (which Nick Sitar at Berkeley says is a bit conservative based on his recent research). On the other side, the full passive pressure would almost certainly not be mobilized. There would be a reaction force, but less than full passive (probably much less). Some have modeled it by stiff springs, but I don't know whether that is a good approach, or if it even matters much for your case, where the splitter walls are apparently not structurally connected at their tops to form a frame that transfers P-active on the right side to the reaction on the left.

The hydrodynamic force on the walls is most easily modeled by the Westergaard model, which is based on the inertial force from the water, not the large surface waves you have shown on the sketch. (See link below. Remember that the sloshing waves would have a period of a second or more, whereas the earthquake's energy is, in most cases, dominated by much shorter periods. Come to think of it, the Westergaard model is based on an infinite reservoir against a spillway gate, so maybe it's conservative for this case? The answer probably depends on the p-wave velocity of water. Regardless, because of the high frequency, it's the inertia of the water that would create the unbalanced forces on the splitter walls, not the large-magnitude sloshing waves in the sketch.

http://www.pwri.go.jp/eng/ujnr/joint/39/paper/62iwashita.pdf

Figure C is more nearly correct, in that the net force on the splitter walls due to the inertia of the splitter walls and the water would occur in the same direction as the M-O active pressure, opposite to the direction that the foundation is accelerating.

Is this a buried tank with partitions or roof supports, or a 4-bay spillway, or what? Are the proportions shown in the sketch true scale?

Bon chance!

DRG

 

[alphaxy]

dgillete,

That is really great. Im doing figure C now. There is a soil extending at full height of wall in left most side and a lower level of soil (2.4 meters) offset from topmost height of wall in right side. I forgot to tell you that there is a beam that frames the walls and a superstructure that houses this filter beds (walls with 5 cells in a row). This cannot actually be considered as buried tank, because there is no soil in front side but a shearwall type of structure. On the right side, the soil is lower. Only the back and left side of the structure is flushed on top of wall height.

The sketch I provided is just a simplification of the entire structure. I attach the plan view of th frames on top for a more clear view.

In sides with soils, I put springs in my model and pinned at the bottom of the wall. I run earthquake in X direction and it catches the true response from earthquake. However, I am still confused with what to do with active and passive pressure forces in the event of seismic.

You said and I believe too that.., hydrodynamic, inertia forces and active pressure runs in one direction..but this time, Let us assume in my staad model, when you look at the plan view I attached..the positive direction of Z is heading towards the right side wall and negative is heading towards left side of wall.Then, we have a framing on top of walls that would transfer the forces..What do you think is the ideal soil presure force (that is opposing the forces hydrodynamic, inertia, forces and active pressure) in the right side of wall? (if the forces hydrodynamic, inertia, forces and active pressure is heading positive Z) Do I need to put active pressure in this side "Or" passive pressure?

Knowing I have spring in all sides which has soil..If I will put a counter pressure in this side, I am thinking that the spring will receive the net of this forces and therefore reducing the amount of force in the spring.

Your help is very much appreciated.

 

[alphaxy]

dgillete,

you mentioned, stiff springs. I am thinking if this idea might be correct. what I will do is obtain the value of passive pressure and load it as spring K value. So for example..The earthquake forces is going to the right side (Active static + dynamic soil pressure, hydrostatic pressure), then the spring will react compressively to these forces as passive static + dynamic pressure.

Thanks

 

[dgillette]

I would not try to make the force from the springs equal come out equal to the full Coulomb passive condition. It's just a reaction force, and will probably be less than passive. The Coulomb passive condition is an upper bound on the force that can be mobilized on the right wall. Just verify that the reaction force in the springs is less than or equal to the passive case.

Looking at the first picture again I might have told you wrong. I think the worst-case load is when the base of the model is accelerating toward the left, so you would get Mononobe-Okabe active pressure pushing on the left, and hydrodynamic pressure pushing on the left side of each partition. On the right wall, something equal to OR LESS THAN Coulomb passive pressure. I guess that's more like B. Sorry - too much hurry!

If I were the one doing the analysis, and I'm not, I would start very simply, using Mononobe-Okabe active on the left side, and some fraction of a-max*[weight of water] on each partition, and friction on the base. How much reaction force is needed to keep it from sliding rightward, and how does that compare with the passive condition? (Probably not much force is needed, given the size of the structure relative to depth.) Now I suppose the question is failure of the partitions as shear walls in one direction and bending plates in the other. Do you need much reaction force from soil on the right to prevent it?

Occurs to me that if you are doing a full-blown FEM analysis, it might be better to make the soil on the reaction side (right) part of the FEM model as a material*, rather than as springs, since it would be subjected to the ground motion also, making it want to pull away from the structure. My unconfirmed HUNCH is that the soil is stiff enough compared to a frame that the reaction force would not be reduced greatly, although your lattice of interior partitions would be pretty stiff. This is something we've talked about in our shop in relation to multi-bay spillways that act as frames, and I don't think we ever got to a completely satisfactory answer.

*Material could have a Mohr-Coulomb strength, so it could shear and create the passive wedge if the force gets high enough, but it would probably be simpler to make it a purely elastic material and verify that the pressure stays well below Coulomb passive.

Regards,
DRG

 

[ishvaaag]

Sorry, CFD is Computational Fluid Dynamics, still reading...