[Tochnog-users] dependency diagram for druckprag

[t***@lists.sourceforge.net]

Hi all,

I'm testing the Drucker-Prager plasticity model. To enforce an isotropic
softening behavior, I use dependency_item and dependency_diagram. I want
to change `c` value (in group_materi_plasti_druckprag) according to
`kap` so as to decrease the yield strength `K`. In tochnog pro, there is
dependency_number option. But there is no such option in the free
version. So I write like this:

dependency_item 1 -group_materi_plasti_druckprag 1 -kap 2
dependency_diagram 1 0. 0.1 0.611 1.5e6 0.175 0.611 5246.89 0.175

Here `phi` and `phiflow` are kept unchanged and only `c` is decreasing.
But I find I cannot get a reasonable solution. I get a inhomogeneous
field which I expect not. Also if I delete this two lines (no
hardening/softening behavior), the result seems correct. Could anyone
tell me what causes this problem? Or how to enforce the softening
behavior for d-p model?

[t***@lists.sourceforge.net]

Message did not come well but I was able to read it and reply:
"...
Hi all, I'm testing the Drucker-Prager plasticity model. To enforce an isotropic softening behavior, I use dependency_item and dependency_diagram. I want to change `c` value (in group_materi_plasti_druckprag) according to `kap` so as to decrease the yield strength `K`. In tochnog pro, there is dependency_number option. But there is no such option in the free version. So I write like this: dependency_item 1 -group_materi_plasti_druckprag 1 -kap 2 dependency_diagram 1 0. 0.1 0.611 1.5e6 0.175 0.611 5246.89 0.175 Here `phi` and `phiflow` are kept unchanged and only `c` is decreasing. But I find I cannot get a reasonable solution. I get a inhomogeneous field which I expect not. Also if I delete this two lines (no hardening/softening behavior), the result seems correct. Could anyone tell me what causes this problem? Or how to enforce the softening behavior for d-p model? -- Geotechnical Group Department of Civil and Environmental Engineering Hong Kong University of Science and Technology Clear Water Bay, Kowloon, Hong Kong --------------000905000103040008020104 Content-Type: text/html; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Hi all,

I'm testing the Drucker-Prager plasticity model. To enforce an isotropic softening behavior, I use dependency_item and dependency_diagram. I want to change `c` value (in group_materi_plasti_druckprag) according to `kap` so as to decrease the yield strength `K`. In tochnog pro, there is dependency_number option. But there is no such option in the free version. So I write like this:

dependency_item 1 -group_materi_plasti_druckprag 1 -kap 2
dependency_diagram 1 0. 0.1 0.611 1.5e6 0.175 0.611 5246.89 0.175

Here `phi` and `phiflow` are kept unchanged and only `c` is decreasing. But I find I cannot get a reasonable solution. I get a inhomogeneous field which I expect not. Also if I delete this two lines (no hardening/softening behavior), the result seems correct. Could anyone tell me what causes this problem? Or how to enforce the softening behavior for d-p model?

--
Geotechnical Group
Department of Civil and Environmental Engineering
Hong Kong University of Science and Technology
Clear Water Bay, Kowloon, Hong Kong

"

It will work well but you have to give it the data in the proper waY: Let's say that your property -group_materi_druckprag depends on the equivalent plastic strain -kap and that the data points are as follows:

-group_materi_druckprag -kap
8.0 0.0
10.0 0.05
45.0 0.10
100.0 0.20
130.0 0.25

Then, the way to specify the diagram is as follows:

dependency_item 1 -group_materi_plasti_druckprag 1 -kap 5 (The number 5 indicates the number of data pairs to read)
dependency_diagram 1 0.0 0.05 0.10 0.20 0.25 8.0 10.0 45.0 100.0 130.0 (note that ALL the values for kap are entered first
then the values for group_materi_druckprag are entered)

I hope this helps!

Best Regards,

[t***@lists.sourceforge.net]

On Saturday, February 25, 2012 01:54 AM,
tochnog-***@lists.sourceforge.net wrote:

Hi Fernando,

Thanks for your reply. I know the use of dependency_diagram. But the
number of dependency_item for group_materi_plasti_druckprag is THREE:
phi, c and phiflow, while I only want to set c dependent on kap. In
tochnog pro, there is a dependency_number to set which item, but there
is no such option in tochnog free. So I write all three parameters for
each kap (I have 2 values for kap 0 and 0.1). There is no error message
for the input file. Just the result seems unreasonable.

Ning

"...
Hi all, I'm testing the Drucker-Prager plasticity model. To enforce an isotropic softening behavior, I use dependency_item and dependency_diagram. I want to change `c` value (in group_materi_plasti_druckprag) according to `kap` so as to decrease the yield strength `K`. In tochnog pro, there is dependency_number option. But there is no such option in the free version. So I write like this: dependency_item 1 -group_materi_plasti_druckprag 1 -kap 2 dependency_diagram 1 0. 0.1 0.611 1.5e6 0.175 0.611 5246.89 0.175 Here `phi` and `phiflow` are kept unchanged and only `c` is decreasing. But I find I cannot get a reasonable solution. I get a inhomogeneous field which I expect not. Also if I delete this two lines (no hardening/softening behavior), the result seems correct. Could anyone tell me what causes this problem? Or how to enforce the softening behavior for d-p model? -- Geotechnical Group Department of Civil and Environmental Engineering Hong Kong University of Science and Te!
chnology Clear Water Bay, Kowloon, Hong Kong --------------000905000103040008020104 Content-Type: text/html; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Hi all,
dependency_item 1 -group_materi_plasti_druckprag 1 -kap 2
dependency_diagram 1 0. 0.1 0.611 1.5e6 0.175 0.611 5246.89 0.175
Here `phi` and `phiflow` are kept unchanged and only `c` is decreasing. But I find I cannot get a reasonable solution. I get a inhomogeneous field which I expect not. Also if I delete this two lines (no hardening/softening behavior), the result seems correct. Could anyone tell me what causes this problem? Or how to enforce the softening behavior for d-p model?

[t***@lists.sourceforge.net]

Ning,

I read your comment and still believe that you are not entering the dependency diagram items the proper way.
I checked the code and the open source version allows to enter more than one property. See the attached note
from the manual on dependency_item:

"….
Temperature dependent Young's moduli in two maxwell chains of element_group 1 ( for the first chain the
Moduli 1.e10, 1.e9,... for the second chain the moduli 1.e12, 1.e11,... all relaxation times are 1.e-2 . ):

dependency_item 1 -group_materi_maxwell_chain 1 -temp 4
dependency_diagram 1
1. 2. 3. 4.
1.e10 1.e9 1.e8 3.e5
1.e-2 1.e-2 1.e-2 1.e-2
1.e12 1.e11 1.e10 3.e7
1.e-2 1.e-2 1.e-2 1.e-2
…"

As you can see, you enter the list of unknowns, in the example above, the values of -temp, then you enter the list
of values of the
first chain moduli, and then the list of the relaxation time values, then the second list of the second chain moduli,
and then the list of the second relaxation times for the secon chain.

for your case, upon reviewing your file, you have

"…
group_materi_plasti_druckprag 1 0.611 1.5e6 0.175
group_integration_points 1 -maximal

dependency_item 1 -group_materi_plasti_druckprag 1 -kap 2
dependency_diagram 1 0. 0.1 0.611 1.5e6 0.175 0.611 5246.89 0.175
…"

Therefore, I understand that you are using:
group_materi_plasti_druckprag index phi c phiflow

and that your inital your initial values are phi= 0.611, c = 1.5e6 and phiflow= 0.175

You are keeping phi and phiflow constant and are modifying c. Your dependency diagram should therefore be as follows:

++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

dependency_item 1 -group_materi_plasti_druckprag 1 -kap 2
dependency_diagram 1 0. 0.1 0.611 0.611 1.5e6 5246.89 0.175 0.175

++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

So I am giving two values of -kap, two values of phi, two values of c, and two values of phiflow.

I changed the values and it runs well and very fast. I am attaching a copy of the modified file for your review.

Just a couple of reminders:

1. The updated, that is current version of tochnog allows the use of the Lapack solver which is faster than the
default biconjugate solver. I built the version for Mac OSX but currently do not have access to Linux or Windows compilers.

2. Rather than using "group_integration_points 1 -maximal" you can use the modern version of tochnog: "number_of_integration_points N" ,
where N is any number of integration points that you want to use. If you read the manual you will learn that -maximal has a limit,
but "number_of_integration_points" does not. Just be reasonable and understand how many points you will need.
3. If you use "control_print_gid "number" -yes" and use the same number or index for n as for control_timestep, then you can print
and plot (at least in GiD) the results for any time step and not only at the end of the calculations.
4. Finally, I keep suggesting to use the automatic iterations option in tochnog, as it allows you to control the reslution or error of the
problem. I included that in the modified file attached.

Good luck and please let me know if this works.

Best regards

[t***@lists.sourceforge.net]

On Sunday, February 26, 2012 10:35 AM,
tochnog-***@lists.sourceforge.net wrote:

Hi Fernando,

Thank you very much for the detailed explanations and kind suggestions.
As you pointed out, I wrongly used the dependency_diagram for multiple
parameters. With your corrected script, the program works smoothly and fast.

For options_solver, I cannot test the Lapack solver in my Ubuntu desktop
(`I do not know : -matrix_lapack`). How to install extra solvers and
make them accessible by tochnog?

Best regards,
Ning

Post by t***@lists.sourceforge.net
Ning,
I read your comment and still believe that you are not entering the dependency diagram items the proper way.
I checked the code and the open source version allows to enter more than one property. See the attached note
"….
Temperature dependent Young's moduli in two maxwell chains of element_group 1 ( for the first chain the
dependency_item 1 -group_materi_maxwell_chain 1 -temp 4
dependency_diagram 1
1. 2. 3. 4.
1.e10 1.e9 1.e8 3.e5
1.e-2 1.e-2 1.e-2 1.e-2
1.e12 1.e11 1.e10 3.e7
1.e-2 1.e-2 1.e-2 1.e-2
…"
As you can see, you enter the list of unknowns, in the example above, the values of -temp, then you enter the list
of values of the
first chain moduli, and then the list of the relaxation time values, then the second list of the second chain moduli,
and then the list of the second relaxation times for the secon chain.
for your case, upon reviewing your file, you have
"…
group_materi_plasti_druckprag 1 0.611 1.5e6 0.175
group_integration_points 1 -maximal
dependency_item 1 -group_materi_plasti_druckprag 1 -kap 2
dependency_diagram 1 0. 0.1 0.611 1.5e6 0.175 0.611 5246.89 0.175
…"
group_materi_plasti_druckprag index phi c phiflow
and that your inital your initial values are phi= 0.611, c = 1.5e6 and phiflow= 0.175
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
dependency_item 1 -group_materi_plasti_druckprag 1 -kap 2
dependency_diagram 1 0. 0.1 0.611 0.611 1.5e6 5246.89 0.175 0.175
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
So I am giving two values of -kap, two values of phi, two values of c, and two values of phiflow.
I changed the values and it runs well and very fast. I am attaching a copy of the modified file for your review.
1. The updated, that is current version of tochnog allows the use of the Lapack solver which is faster than the
default biconjugate solver. I built the version for Mac OSX but currently do not have access to Linux or Windows compilers.
2. Rather than using "group_integration_points 1 -maximal" you can use the modern version of tochnog: "number_of_integration_points N" ,
where N is any number of integration points that you want to use. If you read the manual you will learn that -maximal has a limit,
but "number_of_integration_points" does not. Just be reasonable and understand how many points you will need.
3. If you use "control_print_gid "number" -yes" and use the same number or index for n as for control_timestep, then you can print
and plot (at least in GiD) the results for any time step and not only at the end of the calculations.
4. Finally, I keep suggesting to use the automatic iterations option in tochnog, as it allows you to control the reslution or error of the
problem. I included that in the modified file attached.
Good luck and please let me know if this works.
Best regards
Fernando Lorenzo
------------------------------------------------------------------------------
Virtualization& Cloud Management Using Capacity Planning
Cloud computing makes use of virtualization - but cloud computing
also focuses on allowing computing to be delivered as a service.
http://www.accelacomm.com/jaw/sfnl/114/51521223/
_______________________________________________
Tochnog-users mailing list
https://lists.sourceforge.net/lists/listinfo/tochnog-users