Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.

error: Failed to Evaluate Expression

Efthymios Papadopoulos

Please login with a confirmed email address before reporting spam

Hi everyone,

I am a new COMSOL user so I apologize for any stupid questions.

I am working on soil mechanics [very simple verification footing problem for now] and I try to implement a material with stiffness-stress dependency, in the form of:
[stiffness]=a*[stress]^b.

I understand that I need some initial values for stress, so I also included them under the material model.

However, when I try to run the analysis, I get the message below as an error:


Failed to evaluate variable.
- Variable: p
- Geometry: 1
- Domain: 1
Failed to evaluate variable.
- Variable: mod1.solid.E
- Defined as: (unit_Pa_cf*mod1.mat1.def.Ecross((p/unit_Pa_cf)))
Failed to evaluate variable.
- Variable: mod1.solid.D33
- Defined as: (((1-mod1.solid.nu)*mod1.solid.E)/((mod1.solid.nu+1)*(1-(2*mod1.solid.nu))))
Failed to evaluate variable.
- Variable: mod1.solid.SlZ
- Defined as: (mod1.solid.siZ+((mod1.solid.eelZ*mod1.solid.D33)+((mod1.solid.D13*mod1.solid.eelR)+(mod1.solid.D23*mod1.solid.eelPHI))))
Failed to evaluate variable.
- Variable: mod1.solid.SZ
- Defined as: mod1.solid.SlZ
Failed to evaluate expression.
- Expression: 1.5*((mod1.solid.SR-(mod1.solid.SR+mod1.solid.SPHI+mod1.solid.SZ)/3)^2+2*mod1.solid.SRPHI^2+2*mod1.solid.SRZ^2+(mod1.solid.SPHI-(mod1.solid.SR+mod1.solid.SPHI+mod1.solid.SZ)/3)^2+2*mod1.solid.SPHIZ^2+(mod1.solid.SZ-(mod1.solid.SR+mod1.solid.SPHI+mod1.solid.SZ)/3)^2)


Does anyone happen to know what is going wrong?

Thanks everyone for your time!

Thymios


5 Replies Last Post 2 févr. 2012, 12:35 UTC−5
Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 31 janv. 2012, 13:36 UTC−5
Hi

its rather simple,:you define your Young modulus E w.r.t. a variable "p" but "p" is not defined in your model, hence a series of variables alld epending on E are unknown for COMSOL

You ave defined an analytical equation, but you could also define units for the argument and the results, then you do not need to use the [1/Pa] units trick to make your arguments without units while calling your function

Also the body load, normally one writes "-solid.rho*g_const" along the axis direction. In this way COMSOL looks after the gravity of Earth (g_const = 9.81...m/s^2, except if it was the step on the Moon pls correct accordingly, but I do not believe COMSOL as the gravity constant of the Moon stored by default ...) Like that you do not need to change all values, if you change the material

Also for the pressure parametric load, its easier to use a "solver continuation" sweep, rather than a "external" Parametric Sweep, the former launches one load, then continues to solve with the next etc, by using last parameter value as starting point for the next analysis. While the latter "external" Parametric sweep, runs through the geometry, and the mesh to update these, then starts each new solving process by applying the initial (mostly 0,0) conditions. The differences are subtile, and in many cases makes you gain quite some processing time.

--
Good luck
Ivar
Hi its rather simple,:you define your Young modulus E w.r.t. a variable "p" but "p" is not defined in your model, hence a series of variables alld epending on E are unknown for COMSOL You ave defined an analytical equation, but you could also define units for the argument and the results, then you do not need to use the [1/Pa] units trick to make your arguments without units while calling your function Also the body load, normally one writes "-solid.rho*g_const" along the axis direction. In this way COMSOL looks after the gravity of Earth (g_const = 9.81...m/s^2, except if it was the step on the Moon pls correct accordingly, but I do not believe COMSOL as the gravity constant of the Moon stored by default ...) Like that you do not need to change all values, if you change the material Also for the pressure parametric load, its easier to use a "solver continuation" sweep, rather than a "external" Parametric Sweep, the former launches one load, then continues to solve with the next etc, by using last parameter value as starting point for the next analysis. While the latter "external" Parametric sweep, runs through the geometry, and the mesh to update these, then starts each new solving process by applying the initial (mostly 0,0) conditions. The differences are subtile, and in many cases makes you gain quite some processing time. -- Good luck Ivar

Efthymios Papadopoulos

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 31 janv. 2012, 17:41 UTC−5
Hello Ivar and thank you for your flash reply,

I tried most of the things that you mentioned. They were quite helpful and the analysis ran. More specifically I related E to mean stress, as I initially wanted to.
However, I was not sure I was getting the correct results, so I included a parameter on the model and I ran a parameter sweep to check if the model takes into consideration the stress dependent stiffness.

As you can see from the model, for all parameters (namely initial stiffness of 10-100MPa), the results are the same, which means that the model take into consideration only the constant part of the equation. I double checked that by also running a parameter sweep for that value, and the results proved the same.

Could you possibly know why is that happening?

Thank you!

Thymios

ps: I cannot upload my new model for some reason. I get a message it's too large.
Hello Ivar and thank you for your flash reply, I tried most of the things that you mentioned. They were quite helpful and the analysis ran. More specifically I related E to mean stress, as I initially wanted to. However, I was not sure I was getting the correct results, so I included a parameter on the model and I ran a parameter sweep to check if the model takes into consideration the stress dependent stiffness. As you can see from the model, for all parameters (namely initial stiffness of 10-100MPa), the results are the same, which means that the model take into consideration only the constant part of the equation. I double checked that by also running a parameter sweep for that value, and the results proved the same. Could you possibly know why is that happening? Thank you! Thymios ps: I cannot upload my new model for some reason. I get a message it's too large.

Efthymios Papadopoulos

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 31 janv. 2012, 17:49 UTC−5
I managed to include it, but without the solution..
I managed to include it, but without the solution..


Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2 févr. 2012, 02:27 UTC−5
Hi

you are right the stress does not change, BUT the displacement does ;)

add a Plot More Max/Min Line to your strss plot, or add a separate surface defomation plot and you will notice it

--
Good luck
Ivar
Hi you are right the stress does not change, BUT the displacement does ;) add a Plot More Max/Min Line to your strss plot, or add a separate surface defomation plot and you will notice it -- Good luck Ivar

Efthymios Papadopoulos

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2 févr. 2012, 12:35 UTC−5
Hi again!

Well, If you notice, I changes only for change in the "constant" part of the stiffness k0.

If you do the same parametric study for the stress dependency factor k1, you will notice that all the displacements coincide.

I cannot understand how this works, since I get no error during the analysis, but still, half the equation is not taken into account :(.


Thank you again for your time!

Thymios
Hi again! Well, If you notice, I changes only for change in the "constant" part of the stiffness k0. If you do the same parametric study for the stress dependency factor k1, you will notice that all the displacements coincide. I cannot understand how this works, since I get no error during the analysis, but still, half the equation is not taken into account :(. Thank you again for your time! Thymios

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.