EARTHQUAKE TIME HISTORY

[civil33]

Hi Guys:
I’m doing a model of a cylindrical storage tank (The wall of tank with SHELL63 and the fluid with FLUID80), subject to seismic loads, i.e El Centro earthquake accelerogram record (the accelerogram is in the X direction, the weight of tank is in the z direction). The record is in this form:

TIME, t1
ACEL, a1
TIME, t2
ACEL, a2
.
.
.
etc..

The accelerations data are between 0.02 seconds.
I need input the data in this form. (I did a macro with EXCEL to order all the accelerations with the time). I have this command. I would like to know if the commands are corrects. Please Someone can help me?

/CONFIG,NRES,40000
/SOLU
ANTYPE,TRANS
TRNOPT,FULL
NLGEOM,OFF
PSTRES,ON <============= Is correct?
ACEL,,,G <============= Is coorect (9.81 m/sg²)
DMPRAT,0
OUTPR,,ALL
OUTRES,BASIC,ALL
OUTRES,ESOL,ALL
TIMINT,ON <============= Is correct?
KBC,0 <============= Is correct?
AUTOTS,ON <============= Is correct?
DELTIM,0.005,0,0,0 <============= Is correct?
NSUBST,10,,,0 <============= Is correct?
/OUTPUT,TRANS,OUT
/INPUT,ELCENTRO,txt (this command call to the accelerations file)
FINISH
SAVE
FINISH
/OUTPUT

The accelerations file has this form:
ACEL, a1
TIME, t1
SOLVE
ACEL,a2
TIME, t2
SOLVE
.
.
.
etc

Are corrects this commands?????????

Other questions

I read a paper that say the following:

“ An initial deformations condition of the tank structure due to dead and hydrostatic loads was used as starting condition for the dynamic analysis”

How Can I include what say the paper???????


Thank you very much

 

[Drej]

You have brought up many, many issues here, way too many to deal with. This is a complicated subject area, and it sounds like you need to do some reading of both seismic/dynamics theory (see Anil Chopra - Dynamics of Structures to start with) and how this type of analysis is performed in ANSYS. I have a few comments to start with:

1) ANSYS will not directly accept acceleration time history data - you appear to be using the ACEL command to input your time history. This is incorrect (unless you are using the "large mass" method, which I doubt you are). Integrate your accelerations back to displacements and apply these displacements to the structure at the correct point(s).

2)

PSTRES,ON <============= Is correct? you don't need prestressing, it is irrelevant
ACEL,,,G <============= Is coorect (9.81 m/sg²) see my comment above
DMPRAT,0 this is irrelevant in transient analyses, you need ALPHAD and/or BETAD (Rayleigh) here
OUTPR,,ALL
OUTRES,BASIC,ALL
OUTRES,ESOL,ALL ! think about asking for accelerations and velocities as well, and also think about how big your results file will be
TIMINT,ON <============= Is correct? Yes
KBC,0 <============= Is correct? Yes
AUTOTS,ON <============= Is correct? Yes, but you need to ensure that it is not affecting the solution. See the help file.
DELTIM,0.005,0,0,0 <============= Is correct? No. The time step needs to be based on the frequency response of your structure, more precisely to capture the frequency of your structure that you are interested in. See the help file.
NSUBST,10,,,0 <============= Is correct? No. You can't have DELTIM and NSUBST: either one or the other.


3) “ An initial deformations condition of the tank structure due to dead and hydrostatic loads was used as starting condition for the dynamic analysis”

Yes, you need to account for this if the loads are of the same order (approximately) relative to the seismic loads. These ar the initial conditions, solve for these loads first transiently by turning off time integration (TIMINT,OFF) and then turn this back on and solve your displacement time history.

4) If all of this doesn't make sense, then read the help file and then some literature on dynamics. Then do a simple model that you can validate using hand calculations. Then think about doing an actual model using actual time history data. Otherwise, you're like a blind man driving a car.


------------
See faq569-1083 for details on how to make best use of Eng-Tips.com

 

[amini06]

Hi Sir

I am doing a model of a wall, subject to the seismic loads, too.
Considering to the above comments:
Do you confirm your method, yet?
Lets we have a simple structure, say a column. Gravity force, as you know is a "body force" and acting to the whole of the structure. But what about seismic loads?
In a seismic shaking, we have these steps:
1-Ground starts to shake.
2-The structure foundation starts to shake.(The base of the structure suffers some displacements vs time)
3-At last the structure, considering it damping properties, starts to shak(displacements) and it causes a new stress distribution vs time.
When we apply "g" as a variable vs time (ACEL,g1 TIME,T1 ,...)are you sure to consider the above steps? How you model the seismic load transferring steps?

Regards