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.

Coupling Stationary study -> Time dependent does not work

Please login with a confirmed email address before reporting spam

Dear all,

I have a very simple model (attached): a 2D model where I impose an initial strain, that generates stress. This stress will be relaxed by creep. The stress caused by imposed strain is simulated by a Stationary step, and then the resulting state of stress is used by the Time Dependent step that involves creep.

Now the problem is: running the simulation, the stress at t=0 obtained by time dependent step is completely different from the stress obtained by Stationary step...why? If I'm not wrong, they should be identical: I get a stress state from the stationary step, and then I use it for the time dependent step, but at t=0 creep has not run yet...

I have tried to change in "Dependent Variables 2" (in Solver 1) "Initial Values of Variables Solved For" using the "Store Solution". I also used, before running the Time Dependent Step, "Get Initial Value for Step" on the Time Dependent icon, but in this case again I obtain as initial value a stress state completely different from that one obtained by Stationary Step.

Thank you very much for any suggestion!

Regards
Fabio


7 Replies Last Post 2015年9月14日 GMT-4 06:45

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2013年7月24日 GMT-4 12:59
Hi,


I do not know why your approach does not work. I am not sure if it has something to do with disabling the "initial stress and strain" subnode during the time dependent step? I would be interested to know.

Anyway...an alternative suggestion - why not try the problem with two solid mechanics nodes - let us say the identifiers are solid_1 and solid_2. For solid_1 node, as you have done impose initial strain via the "initial stress and strain" subnode. For solid_2 node, impose initial stresses in the "initial stress and strain" subnode i.e. solid_1.sx, solid_1.sy as input variables. Of course, set up the solver sequence accordingly.


Suresh
Hi, I do not know why your approach does not work. I am not sure if it has something to do with disabling the "initial stress and strain" subnode during the time dependent step? I would be interested to know. Anyway...an alternative suggestion - why not try the problem with two solid mechanics nodes - let us say the identifiers are solid_1 and solid_2. For solid_1 node, as you have done impose initial strain via the "initial stress and strain" subnode. For solid_2 node, impose initial stresses in the "initial stress and strain" subnode i.e. solid_1.sx, solid_1.sy as input variables. Of course, set up the solver sequence accordingly. Suresh

Josh Thomas Certified Consultant

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2013年7月24日 GMT-4 15:41
I'm with Suresh. I don't know why this behaves like this.

Another workaround to consider if Suresh's doesn't work for you:

Just solve in 1 time dependent step where you ramp the initial strain from 0 to -0.0005 over a short step in the very beginning of the simulation. Use a Step function titled "init_strain" for example and then type init_strain(t[1/s]) in the edit fields on the "Initial Stress and Strain" node. Then, as time develops the creep phenomenon will take over.

Best regards,
Josh Thomas
AltaSim Technologies

I'm with Suresh. I don't know why this behaves like this. Another workaround to consider if Suresh's doesn't work for you: Just solve in 1 time dependent step where you ramp the initial strain from 0 to -0.0005 over a short step in the very beginning of the simulation. Use a Step function titled "init_strain" for example and then type init_strain(t[1/s]) in the edit fields on the "Initial Stress and Strain" node. Then, as time develops the creep phenomenon will take over. Best regards, Josh Thomas AltaSim Technologies

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2013年7月24日 GMT-4 17:55
Dear Suresh and Josh,

actually, I was using two Solid Mechanics on a more complex model but for a convergence problem I contacted the Comsol Support, and they replied saying that since initial stresses are implemented in Comsol, they are inelastic and thus unaffected by creep (they remain constant during the simulation). Therefore, the suggestion was to use one only Solid Mechanics and then for each study step decide the conditions specified in the Solid Mechanics.

Setting two Solid Mechanics, I don't reach the convergence of the time dependent study (creep study), neither on a more complex model, nor on an extremely simple model like that one attached. The issue, as I said before, is the imposed initial stress (using e.g. "Initial Stress and Strain") that remains constant, is not relaxed but creep. On the other hand, the creep strain rate is dependent of these stresses...

Josh, yes I ran the simulation with this workaround. But the problem comes up again when you need to set e.g. some material properties or boundary that change in Solid Mechanics.

Anyway, thank you all for your help!

Kind regards
Fabio
Dear Suresh and Josh, actually, I was using two Solid Mechanics on a more complex model but for a convergence problem I contacted the Comsol Support, and they replied saying that since initial stresses are implemented in Comsol, they are inelastic and thus unaffected by creep (they remain constant during the simulation). Therefore, the suggestion was to use one only Solid Mechanics and then for each study step decide the conditions specified in the Solid Mechanics. Setting two Solid Mechanics, I don't reach the convergence of the time dependent study (creep study), neither on a more complex model, nor on an extremely simple model like that one attached. The issue, as I said before, is the imposed initial stress (using e.g. "Initial Stress and Strain") that remains constant, is not relaxed but creep. On the other hand, the creep strain rate is dependent of these stresses... Josh, yes I ran the simulation with this workaround. But the problem comes up again when you need to set e.g. some material properties or boundary that change in Solid Mechanics. Anyway, thank you all for your help! Kind regards Fabio

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2013年7月25日 GMT-4 08:49
I am trying to solve two modules(drift diffusion and Joule heating) simultaneously. The model is of two electrodes with an insulator above them. In my simulation electrical and thermal conductivity is function of temperature. I am trying to solve this but still I can't couple them together.

Any advice on how to proceed would be greatly appreciated.

Regards,
Sanjoy
I am trying to solve two modules(drift diffusion and Joule heating) simultaneously. The model is of two electrodes with an insulator above them. In my simulation electrical and thermal conductivity is function of temperature. I am trying to solve this but still I can't couple them together. Any advice on how to proceed would be greatly appreciated. Regards, Sanjoy

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2013年7月25日 GMT-4 09:16
Hi Sanjoy,

I don't know what you mean with "drift diffusion", but in principle the coupling of two physics is carried out establishing which variable(-s) are common in the two (or more) physics. The first example that comes to my mind is Joule Heating coupled with Structural Mechanics: an electrical current generates heat due to resistance, thus a temperature field. In your structure (where the current is passing) you can have stresses due to thermal deformations, again caused by a temperature field. Therefore, in the setting of the physics, you establish this common point. How? in "Thermal expansion" of Solid Mechanics you specify, in the field "Model Inputs" the source of you temperature: instead of "User Defined", you should use temperature from the Joule Heating physics...Of course, you can have the inverse effect as well, for example the deformation of your structure will change the electrical resistance, and therefore the Joule Heating (like a circular dependence).
I know, sometimes it's not easy to find the common variables among the physics, but Comsol can help you showing the equation that the physics will use...
Sorry to give you a tip different than your model...if you have still problems, it's better that you upload your model here!

Regards
Fabio
Hi Sanjoy, I don't know what you mean with "drift diffusion", but in principle the coupling of two physics is carried out establishing which variable(-s) are common in the two (or more) physics. The first example that comes to my mind is Joule Heating coupled with Structural Mechanics: an electrical current generates heat due to resistance, thus a temperature field. In your structure (where the current is passing) you can have stresses due to thermal deformations, again caused by a temperature field. Therefore, in the setting of the physics, you establish this common point. How? in "Thermal expansion" of Solid Mechanics you specify, in the field "Model Inputs" the source of you temperature: instead of "User Defined", you should use temperature from the Joule Heating physics...Of course, you can have the inverse effect as well, for example the deformation of your structure will change the electrical resistance, and therefore the Joule Heating (like a circular dependence). I know, sometimes it's not easy to find the common variables among the physics, but Comsol can help you showing the equation that the physics will use... Sorry to give you a tip different than your model...if you have still problems, it's better that you upload your model here! Regards Fabio

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2013年7月31日 GMT-4 21:10
Hi Fabio,

I have same problem. I am trying to solve Drift diffusion(time dependent ) and Joule heating(time independent ) equations simultaneously. One important thing in my simulation is that electrical conductivity and thermal conductivity are function of Temperature(T) and concentration (cA). I have solved the drift diffusion equation first with a constant temperature(T0) and then coupled with joule heating equation by changing T0 to T.

The coupling was done by copying the solution and then changing " initial values of variables Solved for for" in the dependent variables option of the solver 1. After solving thing I don't get any electric field or potential in solution.
Could you please advise me ? I have attached the file herewith.

Thanks,
Sanjoy
Hi Fabio, I have same problem. I am trying to solve Drift diffusion(time dependent ) and Joule heating(time independent ) equations simultaneously. One important thing in my simulation is that electrical conductivity and thermal conductivity are function of Temperature(T) and concentration (cA). I have solved the drift diffusion equation first with a constant temperature(T0) and then coupled with joule heating equation by changing T0 to T. The coupling was done by copying the solution and then changing " initial values of variables Solved for for" in the dependent variables option of the solver 1. After solving thing I don't get any electric field or potential in solution. Could you please advise me ? I have attached the file herewith. Thanks, Sanjoy


Please login with a confirmed email address before reporting spam

Posted: 9 years ago 2015年9月14日 GMT-4 06:45
Hi,
I am using a Non-linear elastic material (Sensor: made of Polyimide) and Hyperelastic material gasket (NBR-Rubber), the sensor is placed inside the gasket and stress is applied from the top plate (Prescribed displacement: -1mm) to compress the gasket to check the tensile strength of the sensor (inside) and to record the material elongation after applying stress at the rubber gasket, keeping bottom plate stationary. (Simulation is attached)

My request is to know how to calculate the creep/ shrinkage behavior over time to observe the performance of gasket (O-ring) and to note the change of strain properties.
My geometry parameters are as follows.
Radius (Gasket/O-ring) = 2.5 mm
Width (Sensor) = 0.5 mm
Height (Sensor) = 0.005mm
Gasket Material= NBR Rubber
Sensor Material=Polyimide
Two steel plates (One at top and one at bottom)
In Solid Mechanics:
Gasket= Hyper-Elastic material
Sensor= Non-Linear Elastic Material
Bottom steel plate= Fixed
Top steel plate= Applying prescribed displacement 1mm downwards.

Help is requested.

Many Thanks,
Bilal
Hi, I am using a Non-linear elastic material (Sensor: made of Polyimide) and Hyperelastic material gasket (NBR-Rubber), the sensor is placed inside the gasket and stress is applied from the top plate (Prescribed displacement: -1mm) to compress the gasket to check the tensile strength of the sensor (inside) and to record the material elongation after applying stress at the rubber gasket, keeping bottom plate stationary. (Simulation is attached) My request is to know how to calculate the creep/ shrinkage behavior over time to observe the performance of gasket (O-ring) and to note the change of strain properties. My geometry parameters are as follows. Radius (Gasket/O-ring) = 2.5 mm Width (Sensor) = 0.5 mm Height (Sensor) = 0.005mm Gasket Material= NBR Rubber Sensor Material=Polyimide Two steel plates (One at top and one at bottom) In Solid Mechanics: Gasket= Hyper-Elastic material Sensor= Non-Linear Elastic Material Bottom steel plate= Fixed Top steel plate= Applying prescribed displacement 1mm downwards. Help is requested. Many Thanks, Bilal

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.