Hello:

in most of the Geotechnical Engineering software, you have three main types of BC: 1) fixed displacements, 2) Impervious boundaries, 3) Nodal / distributed forces / precribed displacements. The first two obviously depend on the geometry of your problem, as well as the nature of the ground flow, if it exists and you are trying to model it. The third one actually represents the input loading for your geotechnical design. In summary, to chose them you need to know how the geometry is, if you have got axisymmetry (if you do, usually all the horizontal displacements in the nodes located in the axis should not be allowed), where the external loadings are, and so on.

In addition, some geotechnical engineering software for dynamics problems allow the application of absorbent or silent boundaries. These boundaries are rather different from the regular ones, because in the later, the borders are rigid, and the stress waves, when reaching them, they reflect, producing unrealistic vibration within the soil mass which does not occur in open field domains. Absorbing boundaries allow these stress waves to pass through these boundaries. You just need to consider this option if you are dealing with dynamic problems in open field geometries.

I hope it helps.

Thank you Susana for the detailed answer,

In my case, I haven’t a dynamic problem; I generate a finite difference model, for a massif of a reinforced soil. So in the case of an embankment loading there is no problem, there are several methods to choose the boundary conditions as well as we can take a boundary condition three times of the large of a semi embankment. But, in my case, I have a loading of a circular oil tank, and I can’t find an expression or method for choosing a value of boundary condition.

Thank you so much.

You can choose some software to check boundary condition in this case. It is inconvinien but you can define exactly this result for your problem.

Hi again:

There are some methods for that, but the most effective manner to solve your problem, and making sure you are doing it well, is trying with different distances, to check their influence in the final result, which should be rather independent on the distance to the boundary. Try with 3 times, 3.5, 4... If there is not relevant difference in the results between 3 and 3.5, then you are doing well. Good luck!!

Hi,

For circular oil tank, if we assume diameter of the tank as "d", then influence zone you will be able to determine. Now the question is,which one you will choose axisymmetric condition or full scale? For axisymmetric condition you will find co-relation in literature. For full scale you may have a look the attached paper (Ground Improvement for Oil Storage Tanks in Panama). I met with them in GeoCongre2014, which might be very close to your problem. They used PLAXIS for their analyses. Good Luck !

Dear All,

Thanks for your detailed answers.

For Mr. Nguyen Le Du which kind of software can check boundary conditions? I don’t think that there is similar software.

Mr. Mohammad Islam, you’re right, but the co-relations existing in literature are just for the case of embankments, do you know some co relations in the case of oil tanks?

I think that the solution is similar to that proposed by Susana; I tried to make a parametric study with changing in distances of boundary conditions, and when I have a value of 0 in horizontal displacements I must fix boundary conditions in this axe.

Thanks’ all.

Hi,

assume height of the tank=h, dia of the tank=d.then find out which one is the maximum. divide the maximum dimension by 2 or more safety consider the maximum dimension as safe side. however, if the project area is improved by the stone column or prefabricated vertical drain or surcharge along with preloading, then you might need to change the assumption. is it instrumented like, inclinometer, profile gauge or settlement plate, if so then placement of those instrumentation is more crucial issue...on the other hand is there any CPT or CPTu data in the project area, what is their distribution along the surface and depth? is soil profile analogous?finally you need to implement your engineering judgement considering all of this factors...thumb rule may give you guide line but not accurate result..

One of the possible methods is to enlarge your boundary horizontally until you have no rotation in principle stresses at the edge of your boundary.

In order to avoid boundary effect in numerical analysis (FEM, FDM or DEM), outer boundary should be placed at a distance at least 3 diameter, height or width away from the boundary of structure that you will be placing in your model.

For static analysis, principally, first you have to determine the influence of the structure (loading) on the stress field (eg. below 0.2*Sigma_0) in your problem and position the boundary outside this perimeter.

For dynamic analysis, is much more difficult and it depends of what time of silent (absorbing) BC you use and what are the theoretical assumptions.

It depends on the problem you are dealing with. I can recommend a book by Prof. Potts & Zdravkovic

You can choose some software (PLAXIS 3D) to check boundary condition in this case.