In dynamic analysis you can impose acceleration–time histories through boundary conditions by creating a tabular amplitude.

thank you for your answer but i want to know where can i impose acceleration–time histories on my model, what nodes exactly?

Ohh sorry, I interpreted it differently.

Tunnel is a wave propagation problem and not free vibraton like aboveground building. You should impose dispacements boundy conditions on your model

Presuming that you are running a 2D plane strain analysis of the soil-tunnel system subjected to vertically propagated shear waves, the acceleration is imposed along the base of the numerical grid as acceleration or displacement time history.

You suould first calculate the displacements profile in the ground using software like Shake.

Then you can apply the displacements history field to the bounderies of you model.

If your contact elements act as a suspension / elastic foundation of your tunnel in regards to the soil, then you might apply the displacements to the soil nodes. Be aware of damping in this elasto-dynamic foundation. The displacements can be evaluated once you work on the periodgram of the probable earthquake in the area of your tunnel.

With regards,

Alessandro

 thank you dear Grigorios

as the fact time history of acceleration recorded on the surface. how can i impose it on the base without considering this subject?

dear yori

I used time history dynamic analysis in my model. your method is quasi-static analysis. but thank you for your suggestion. 

dear Alessandro

can you explain more about periodgram ?

thank you

Dear Hediye,

with periodgram I intend a statistical distribution of the energy/acceleration vs oscillation periods in the earthquake, thus depending from studies in a specific region (e.g., for my country https://ingvterremoti.wordpress.com/la-pericolosita-sismica/ or http://esse1-gis.mi.ingv.it/s1_en.php). The periodgrams give amplitude, e.g. acceleration [g] vs period [s], but not phase relationship, also depending on the long time prediction base.

Hope this helps,

my best regards,

Alessandro

If you are analyzing an actual case study, where you have the acceleration time history at surface, then you should run a deconvolution analysis, using a simplified 1D code (e.g. CyberQuake) to estimate the acceleration time history at the bedrock (e.g. base of your numerical model). You may check this indicative link for further information regarding this procedure: http://geoserver.ing.puc.cl/info/IMG/pdf/UsersGuide2-01.pdf (search for deconvolution).

If you are running a generic analysis you may select real records from recent earthequakes that are closer to the needs of your study (e.g. in terms of frequency contain, amplitude etc). In this case, you should pick acceleration time histories recorded preferably in rock or in quite stiff soil.

In both cases, you may introduce the acceleration either directly on the nodes along the base of your numerical model (assuming a rigid bedrock) or through proper dashpots at the model base to account for the finite rigidity of the bedrock (e.g. Lysmer and Kuhlemeyer analog).

Hope this helps.

Dear Grigorios

it was so helpfull

best regards

hediye ashoori

dear Grigorios

I assumed my rigid bed rock by applying fixed boundary condition to the model base. and time history of earthquake was applied to the nodes of  model base. my database error because of to many boundary condition to mentioned nodes. this is reasonable  of course. do you have any suggestion to improve this problem?

If you want to simulate a rigid bedrock, you should fix the vertical degree of freedom of the nodes at the base of your numerical model. You will leave the horizontal degree of freedom free, and you will introduce either an acceleration or displacement time history as an acceleration or displacement boundary condition in the horizontal direction.  To do so, you should set the horizontal degree of freedom equal to unity (e.g. U1 or A1) and then select an amplitude function in the Amplitude tab. This function will actually be the acceleration or the displacement time history. In any case be careful with the units you use (e.g. displacement in meters, acceleration in m/s2).

Regards

Grigoris

thank you for your time. i will use your applicable guideline.