I am faced with a choice between 3 softwares to determine the soil liquefaction potential of earth dams and other geotechnical problems: Quake/W (Geo-Studio) - FLAC & PLAXIS. Although I am aware of some of the differences between each of them, for e.g. Quake/w uses the increase in PWP to determine te soil liquefaction potential from an estimated damping function, as FLAC uses a frequency analysis with rayleigh damping, etc.
The ease of use must also come into consideration as some are more complex than others (PLAXIS Vs FLAC). So, What do you suggest between the 3 softwares, and the reasons why?
Thanks for your input.
I would never go for Plaxis or FLAC in the case of liquefaction.
To my knowledge, the most advanced model (coupled u-p-U) is now available in OpenSees in both 2D and 3D. To get rid of OpenSees' scripting, you can use OpenSees Navigator or GiD. Some rather complicated phenomena such as cyclic mobility and bidirectional pore pressure generation/dissipation can be addressed in OpenSees in an explicit manner.
You can also put DEEPSOIL on the table as far as you are tackling 1D liquefaction.
Above software are all free to download and need only an academic email...!
Seyed:
thanks for your input, its a very interresting software, although it seems a little less user friendly. W/hen you say coupled u-P-U, what do you mean? And is it possible to design structural elements such as retaining walls, piles etc.
Also, I'm curious to know why you would NEVER go for PLAXIS or FLAC, they are pretty strong technical softwares (FLAC in particular). I can understand why not Quake/W ... but is it because of the calculation procedure?
regards,
OH
u-p-U is a quad or brick element (depending on dimension of the problem) in which each node has 4 DOF. Two for translational displacement of soil (u), one for pore pressure (p), and another for water displacement relative to soil (U).
And I would never use FLAC, Abaqus, etc. as much as I can, because they are commercial and didn't convey the most advances. OpenSees is an open source software which is fast growing. It is also well accepted among earthquake and structural engineers because of its powerful structural elements you would never find in FLAC, Plaxis, or even Abaqus.
BTW, you can use OpenSees Navigator which provide a simple yet efficient GUI so you can have FLAC-level ease..!
I kindly propose to google "OpenSees, effective stress" to have some interesting webpages or find the follwoing URL to have some geotechnical and structural examples
http://opensees.berkeley.edu/wiki/index.php/Examples
Seyed:
thanks a lot for your info, it is a lot of help and a great source of information, I will look into it.
Although this helps a lot, we still have the 3 softwares at my school and knowing that in Canada and the USA, these 3 softwares are used in the design firms, it would be interresting to learn how to use them. If anyone has any suggestions on the differences and preferences about the 3 softwares suggested in the explanation box, it would still be of interest.
Thanks,
The (full 7 dof) u-p-U formulation in OpenSEES (which I couldn't find in current sources) is dated and inefficient and known to have some bugs. A much better version is found in ESSI simulator (UC Davis).
Comparing these three softwares, I prefer FLAC. To soil liquefaction problem, the convergence would be a big issue to FEM-based program since lots element would yield during alanysis. It always takes time to let the calculation converge. However, sometimes the calculation may still diverges. To this, someone would enlarge the calculation's tolerance, but this is not a good idea, which may yield wrong results. Quasi-linear approach used by Quake/W (Geo-Studio) may be a relatively good way, but the yielding deformation is doulted. On the other hand, to FDM-based program like FLAC, as long as the time step is small enough, which would be self-checked by program, every time step could converge and yield satisfied results. Also, the deformation at every time step would be right.
Chih:
After weeks of work on the subject I have come to the same conclusion. The FDM will act just like the propagation of a wave at the roc foundation as it starts to calculate at the boundaries. FEM will tend to have a hard time to converge as it will try at each iteration to make the G value correspond to the distorsion (shear deformation). For state of the art application, non-linear is preferable, with FLAC. For the state of pratice, equialent-linear is acceptable, consiedering you use adequate functions for Gmax reduction and damping. Geo-Studio (Quake/W) would then be working just like Shake, in undrained condition, giving relatively conservative results.
here is a good article to summarize :
Olivier:
When conducting equialent-linear approach (Quake/W), it adopts a fixed G and damping value for a time-domain analysis. After finishing one round, it chooses 0.65 maximum shear strain to modify G and damping value for next iteration till meeting equilibrium condition. It should be noted that during the earthquake, the deformation would be underestimatd while LARGE acceleration amplitude and would be overestimated while small acceleration amplitude, which would let the resluts not be on safe side. By the way, the equialent-linear approach can not directly derive residural deformation after earthquake.
I never used Quake/W (Geo-Studio). Between FLAC and PLAXIS, I would choose FLAC undoubtedly.
My last experience with PLAXIS was terrible. The convergence is often not achieved even for simple simulations.
https://www.researchgate.net/project/ADONIS-A-Free-Nonlinear-Finite-Element-Program-for-Geo-Engineers
NovoLiq:
http://www.novotechsoftware.com/geotechnical-software/soil-liquefaction-software/