EPWP & Dafalias-Manzari model

aminrahmani · Post by **aminrahmani** » Sun Apr 19, 2009 9:45 am

Hi,

I have modeled a 23(m)x23(m)x10(m) saturated soil deposit in which the soil constitutive model is Dafalias-Manzari. My problem is that very little excess pore water pressure is generated during seismic excitation( e.g. I got 12 kPa which should be 50 kPa). When I used the following values for initial stresses, I got quit better results ( e.g. 25 kPa) :

set K0 0.2
set ph1 [expr $K0*(($rhos1-$rhof)*$gravZ*($H))]
set S1 "$ph1 0.0 0.0 0.0 $ph1 0.0 0.0 0.0 0.0"
(H is each element depth from ground surface)

What is your opinion about my problem? Which values should be used for initial stresses?

I used these parameters for the soil material(Nevada Sand):
set G0 150.0
set v 0.05
set Patm 100.0
set kc 1e-5
set Mc 1.14
set c 0.78
set lambda_c 0.027
set Xi 0.45
set e_cr 0.83
set m 0.02
set h0 9.7
set ch 1.02
set nb 2.56
set A0 0.81
set nd 1.05
set z_max 5.0
set cz 800.0

set K0 0.2
set ph1 [expr $K0*(($rhos1-$rhof)*$gravZ*(1))]
set ph2 [expr $K0*(($rhos1-$rhof)*$gravZ*(3))]
set ph3 [expr $K0*(($rhos1-$rhof)*$gravZ*(5))]
set ph4 [expr $K0*(($rhos1-$rhof)*$gravZ*(7))]
set ph5 [expr $K0*(($rhos1-$rhof)*$gravZ*(9))]

set S1 "$ph1 0.0 0.0 0.0 $ph1 0.0 0.0 0.0 0.0"
set S2 "$ph2 0.0 0.0 0.0 $ph2 0.0 0.0 0.0 0.0"
set S3 "$ph3 0.0 0.0 0.0 $ph3 0.0 0.0 0.0 0.0"
set S4 "$ph4 0.0 0.0 0.0 $ph4 0.0 0.0 0.0 0.0"
set S5 "$ph5 0.0 0.0 0.0 $ph5 0.0 0.0 0.0 0.0"

set MClist1 "$rhos1 $e0 $G0 $nou1 $Patm $kc $Mc $c $lambda_c $Xi $e_cr $m $h0 $ch $nb $A0 $nd $z_max $cz"
set alphaij0 "0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0"
set zij0 " 0. 0. 0. 0. 0. 0. 0. 0. 0. "
set ITlist "$alphaij0 $zij0"
set MP1 "MaterialConstant 19 $MClist1 InternalTensor 2 $ITlist"

set ES1 "Dafalias-Manzari 3 4 5 6 2 $S1"
set ES2 "Dafalias-Manzari 3 4 5 6 2 $S2"
set ES3 "Dafalias-Manzari 3 4 5 6 2 $S3"
set ES4 "Dafalias-Manzari 3 4 5 6 2 $S4"
set ES5 "Dafalias-Manzari 3 4 5 6 2 $S5"

set YF "Dafalias-Manzari 0 12 2 1"
set PF "Dafalias-Manzari 0 2 0 11 0 9 0 10 0 5 0 12 0 7 0 8 0 16 0 17 2 1\ 2 2"
set BackStressR "Dafalias-Manzari 2 11 9 10 5 12 7 8 15 13 14 3 1 2"
set FabricDilat "Dafalias-Manzari-fabric 12 19 18 1 2"
set TE "$BackStressR $FabricDilat"

nDMaterial NewTemplate3Dep 5 -MaterialParameter $MP1 -ElasticState $ES1 -YieldFunction $YF -PlasticFlow $PF -TensorEvolution $TE
nDMaterial NewTemplate3Dep 4 -MaterialParameter $MP1 -ElasticState $ES2 -YieldFunction $YF -PlasticFlow $PF -TensorEvolution $TE
nDMaterial NewTemplate3Dep 3 -MaterialParameter $MP1 -ElasticState $ES3 -YieldFunction $YF -PlasticFlow $PF -TensorEvolution $TE
nDMaterial NewTemplate3Dep 2 -MaterialParameter $MP1 -ElasticState $ES4 -YieldFunction $YF -PlasticFlow $PF -TensorEvolution $TE
nDMaterial NewTemplate3Dep 1 -MaterialParameter $MP1 -ElasticState $ES5 -YieldFunction $YF -PlasticFlow $PF -TensorEvolution $TE

avytin · Post by **avytin** » Mon Apr 20, 2009 9:15 am

make sure you use the latest edit of the code. it is posted but i think it is not included in the binary.

aminrahmani · Post by **aminrahmani** » Mon Apr 20, 2009 9:35 am

How can I get the latest edit of the code??

Thanks.

avytin · Post by **avytin** » Mon Apr 20, 2009 9:49 am

through CVS:
opensees.berkeley.edu./OpenSees/developer/cvs.php

aminrahmani · Post by **aminrahmani** » Mon Apr 20, 2009 10:14 am

I 'm using OpenSees ver. 2.0.0. Is the Dafalias-Manzari code changed in OpenSees ver. 2.0.1?

avytin · Post by **avytin** » Mon Apr 20, 2009 12:26 pm

Boris · Post by **Boris** » Wed Apr 29, 2009 1:24 pm

Hello There,

Just a repeat of an earlier note:

small and large deformation material models (von Mises, Drucker Prager, rounder Mohr-Coulomb, Cam Clay, Dafalias Manzari, SaniSand, various hardening rules, ...) in Newtemplate3Dep, Template3Dep and FiniteDeformation directories have not been maintained in OpenSees for long time and therefor should not be used.

This also applies to smal deformation up_ucdavis and u_p_U and large deformation u-p coupled finite elements, and also to 8nbrick, 20nbrick and 27nbrick single phase finite elements...

There were a number of changes to these files that cannot be found in OpenSees repository...

Bottom line is:

do not use them, they are outdated!

Boris Jeremic
Professor
Department of Civil and
Environmental Engineering
University of California, Davis

aminrahmani · Post by **aminrahmani** » Wed Apr 29, 2009 11:51 pm

Dear Professor Jeremic,

Thanks for the note, but I became confused. I have used small deformation up_ucdavis elements with both Drucker Prager and Dafalias-Manzari model. Does it mean that my results are wrong? and what should I do?

Thanks again.

rjaeger · Post by **rjaeger** » Sat May 02, 2009 9:41 am

I think it means that you need to verify that the materials and elements behave as they should. You should run single element and multi-element tests against known analytical solutions to do this.

sb1966 · Post by **sb1966** » Fri May 29, 2009 9:51 am

Dear Prof Boris,

Is there any version of OpenSees linked from your home page where these materials are maintained? For the present moment should we use only the PressureDepend/PressureIndepend Multiyield models? Hope these are well maintained? Pl. correct me if I m wrong

Regards

Boris · Post by **Boris** » Sat May 30, 2009 1:40 pm

Hello There:

>Is there any version of OpenSees linked >from your home page where these >materials are maintained?

No, OpenSees is a UCB product and you need to talk to UCB folks about maintenance of all code...

We have a different code, that is in public domain (there is/will be link from my web site as soon as I find some time to link it, later this summer most likely)

>For the present moment should we use >only the >PressureDepend/PressureIndepend >Multiyield models?

Do not know, see if they have verification and validation procedures/examples...

>Hope these are well maintained?

I do not know...

>Pl. correct me if I m wrong

I personally try not to use any program (libraries, models, algorithms....) that is not verified and validated...

Boris

sb1966 · Post by **sb1966** » Sun May 31, 2009 4:15 am

Dear Prof Boris

>I personally try not to use any program (libraries, models, >algorithms....) that is not verified and validated...

Thanks for your clarification and advice.

>We have a different code, that is in public domain (there is/will be link >from my web site as soon as I find some time to link it, later this >summer most likely)

Looking forward to get the link

Thanks and regards

(sukumar)