Latest Discussions - COMSOL Forums

Contact penetration

Tue, 21 May 2013 17:28:57 +0000

Hi

I have a contact model and there are two domains that one moves with a prescribed displacement and presses down the other one.

I have a contact pairs, but the results seems like one domain penetrates the other one.

I have played with penalty factor but got the same results.

Can anyone help me with this problem?

I attached the file.

Thanks,

Jay

COMSOL Conference Boston 2013 - One month away from Early Bird Abstract Submission Deadline

Tue, 21 May 2013 16:54:45 +0000

Hello,

I hope you are as excited about the upcoming conference season as I am! As a reminder, the Early Bird Abstract Submission Deadline for the COMSOL Conference Boston 2013 is June 21st, exactly one month away from today! You are encouraged to submit your abstract before the deadline to receive a reduced registration fee. Once your abstract is submitted, you can make edits through "My Conference" up until the Final Abstract Submission Deadline, which is August 2nd.

A detailed abstract is required for both oral and poster presentations. For required information and author instructions please visit: www.comsol.com/conference2013/usa/.

We look forward to receiving your abstract!

Best,

Jinlan Huang
Program Chair
COMSOL Conference Boston 2013
Jinlan.huang@comsol.com

thin layers

Tue, 21 May 2013 15:20:50 +0000

Dear all,

I am analysing the waves induced inside a thin layer using a shaker. I apply a sine wave on one of the edges and the analysis is in the time-domain. I am using Comsol v4.0.

Altough I am using the right time step and the right maximum element size (compared to the frequency of the applied vibration), I can't see the sine wave propagating inside the medium. For example, if I apply a burst of 5 cycles, I can't see this 5 cycles. It seems that it doesn't sample the load in the right manner, even if I use a time step that corresponds to 20 samples in each period.

Could it be a problem of solver? I am using a linear time.dependent solver.

If I increase the thickness of the layer, the simulation works well with thick layer. Does it depend on the different circular frequencies? Should I calculate the time step and the maximum element size considering these frequencies and not the frequency of the applied load?

Many thanks.
Ramona

Packed bed mean field model

Tue, 21 May 2013 14:25:20 +0000

Hi,

I am trying to solve the time-dependent concentration profile in a packed bed using the mean field theory rather than the whole CFD definition of the problem. The mass giving rise to the concentration in the bed comes from a source (diffusive flux coming out of spheres in the bed).

I have tried the convection-diffusion equation in the Mathematics library. My problem though is that I am not quite sure how to couple the diffusion out the spheres to the external flow.

Any help is greatly appreciated

higer order modes

Tue, 21 May 2013 13:41:30 +0000

How could i find the higher order modes using comsol 4.2 version.

Samiul

Birefringence

Tue, 21 May 2013 13:38:16 +0000

Can any one help me to provide the answers of the following questions.

1. Provide a detailed physics on the pitch dependence of birefringence.

2. Please provide a detailed physics behind the increasing trend of birefringence as a function of wavelength.

How to implement a recursive function (in time) in COMSOL 3.5a?

Tue, 21 May 2013 12:01:44 +0000

I have a multiphysics problem, where I have a coupled, PDEs, and and one of them needs the viscosity to be updated. The updating formula uses both the concentration of a substance (from the transport equation), and the velocity (this comes from the other equation. The problem is, the velocity needs the viscosity as well, hence the recursivity. Of course, one can set initial values, but then, I would like to do something similar as in the imperative languages (Fortran, C), where you can set

set initial viscosity

FOR time =1 to N, step deltat
solve flow equation (with a given viscosity for the fluid)
solve transport equation
update viscosity with formula
viscosity(t+1)= viscosity (t) + concentration/velocity
NEXT time step

Can it be done without resorting to MATLAB? if so, how?
I tried on MATLAB, but I couldn't figure out how to call the model, it kept telling me that it could not be found. I do not know if it was that I didn't set a proper path, or was it the loop or something more subtle about the way this recursive functionality must be set up.

Thank you!

Error on simulating Lithium-Ion tutorial

Tue, 21 May 2013 11:19:44 +0000

Hello,

i was trying to do the tutorial of simulating a 2D lithium-Ion battery (www.comsol.com/showroom/gallery/9981/), b ut when i try to execute my simulation i get the following errors:

Failed to evaluate variable Jacobian.
- Variable: material.tplus
- Geometry: 1
- Domain: 3

Failed to evaluate variable Jacobian.
- Operator: mean
- Geometry: 1
- Boundary: 5 8

Failed to evaluate variable Jacobian.
- Operator: mod1.liion.tplus
- Geometry: 1
- Boundary: 5 8

I dont find the error. i compared it with the done tutorial model and i find no difference and it executes perfectly. i Always get the error and on materials it shows me the error that it wants for the electrolyte the electrical conductiviy and for the electrode materials electroltye conductivity and transport number, which makes no sense.
If someone could give me a hint what the error is, it would be great.

Greetings

Heinz

Comparison of Comsol versus Ansys

Tue, 21 May 2013 10:31:23 +0000

Hi,

I am doing my master thesis and I need to make comparison of Comsol and Ansys. My master project is focusing on heat transfer interface.
Still I am using the Comsol, because for me it is more enviroment friendly.
However I have one problem.... I need to know some advantages versus Ansys, because I need to prove why I choose the Comsol for further reasearch. Of coure it would be also nice to compare it with regard to heat transfer interfaces
Could you help me with this problem,

Thanks in advance,
Zygimantas

Boundary Condition at PML

Tue, 21 May 2013 07:32:19 +0000

Hi

I am bit confused with the BC for PML. For example I have a resonating disc attached to a substrate. Now The substrate is assumed to be fixed. It's a 2D axisymmetric problem. Substrate is surrounded by PML on 3 sides. When PML is added to the model, do I need to define any BC for PML as it is an added domain? Is there any rule of thumb for the PML scaling factor? Please advise.

stability of slopes

Tue, 21 May 2013 07:20:46 +0000

sir i wish to start a model on slope stability. can you please help me.
i want to know how to proceed with the physics part and which properties among these i can show- plasticity, friction, elasticity in joints.

Error in modeling a fluid flow and heat transfer in a 3D box

Tue, 21 May 2013 07:07:31 +0000

Hello COMSOL Users and Sir Amir Fadil and CFD Experts

I am trying to simulate a 3D NITF Problem by heating up the air inside a box and observe the thermal transfer and fluid flow.
The material properties of the air is composed of piecewise and analytic functions.
However, an error occured as:

''Failed to evaluate variable Jacobian.''
- Variable: pA
- Geometry: 1
- Domain: 1 2 4

I tried the same set up and a different fluid (WATER) with material properties similarly composed of piecewise and interpolation functions and there was no error.

Could somebody help me out on this, please.

Attached is the file.

Thank you very much.

Jonathan

Cylindrical geometry

Tue, 21 May 2013 05:43:48 +0000

Hi there,

I'm trying to model RF coils for MRI with Comsol, and I'm not finding it terribly easy to create the actual geometry. As can be seen in your example (www.uk.comsol.com/showroom/gallery/4101/), MR coils often end up looking like a flat circuit printed on another geometry -- there the example is a symmetric birdcage coil, but sometimes we want to create something not symmetric wrapped around a cylinder (or another more complicated shape).

My question is this: I'm not so sure how to easily create geometries like that shown in the example. Is it possible to "morph" or "melt" one structure onto another, as you can often do in other 3D modelling programs? I've tried playing around with creating a local coordinate system, but it appears that these are only for the solver's use -- and you can't easily define a workplane at constant r on a cylinder, for example.

Do you have any tips for how to create these sorts of structures? I basically want to draw copper on the sides of cylinders/other shapes!

Thanks for your help.

Electrode- Electrolyte Interface / Double Layer Help

Mon, 20 May 2013 19:13:07 +0000

Hello Guys

I was wondering if anyone here has any experience modelling the electrode- electrolyte interface?

To start with Im attempting to create a 2d two electrode model measuring the potential difference between the two electrodes when stimulated with biphasic current waveforms (possibly monophasic to start with).

There are so many ways to model the electrode- electrolyte interface so it would be great to hear off someone who has attempted it and what method they found to be simplest and what problems they had with it.

any guidance would be great.

Thanks

Duncan

Wavelengths seen in 2D plots

Fri, 17 May 2013 13:57:08 +0000

I am simulating the intereaction of EM radiation with a plasmonic element. I have set up the study to parametize the wavelength of the incoming radiation from 4 um to 20 um. In looking at the fields in the 2D plots, I notice that when I switch from wavelength to wavelenth, the interaction of the field with the element does change. Nevertheless, the number of field maxima and minima between the source port and the element stays the same. For example, my port is fixed at 30 um from my element. When I view the 2D plot of a 5 um incident wave I see the same number of maxima and minima in the incoming field as when I view the plot of a 10 um incident wave.

Why would that occur? Shouldn't I see twice as many maxima and minima in the 5 um plot as I do in the 10 um plot?

v/r

John

Error in free convection

Fri, 17 May 2013 08:34:41 +0000

I am trying to model a natural convection, creating a volume of air around the object i want to simulate.
I didn't have acceptable results, so I'm trying to realize larger volumes or to use alternative shapes for the volume.

But when I do it, and try to simulate i have the following error message that interrupts my simulation:

'Errror 6199
Attempt to evaluate square root of a negative number.
Function sqrt'

how could an enlargement of the geometric dimensions of the volume cause such problem?

Shallow water waves model in coefficient PDE or general PDE

Tue, 14 May 2013 09:49:12 +0000

Dear All,
I want to solve the Shallow water waves equations in PDE module in 2D. h is the water elevation, U and V the velocity components along x and y.

How can I implement an incoming pulse in the times domain. Coming from the left on the structure.

If I use the initial conditions, I can specify the shape at given time but I don t see any way of defining for example a incoming wave from the left to the right with a shape like:
h(x,y,t)=exp((t-t0)^2)sin(2*pi*freqt)

Best regards
Muamer

cannot select an edge to become an outlet boundary

Tue, 04 Dec 2012 18:48:52 +0000

Hi!
After struggling for quite a while with 4.2a, I haven't been able to enable one edge to become an outlet. What I did was to construct two objects: the first one is a sequence of lines, which I then made to become a single curve, which in turn I converted into a solid. Then I built a rectangle, which is supposed to embed the first flat, 2D figure. The first figure has an inlet, which was easy to make, just selecting the proper boundary, and that was it. The outlet, no matter how many times I selected that boundary, won't be added to the right blank place, so that I can declare the outlet properties of it. I gave a look, and it happens that most edges (or flat boundaries) are walls. I tried to build the two objects in several ways, and I still haven't figured out how to make them, so that I can take this boundary as an outlet.

This is a matter of some urgency, so please, please, if you have a clear idea of what I need, do write it down. Or even if you only have a hint, but has some algorithmic clarity. I am including the geometry

Yours

Arturo

microwave heating

Mon, 27 Feb 2012 09:00:53 +0000

hi,
I'm using the module of the microwave heating and i was viewing the model with the potato in the oven from which you can see the increase in temperature.
my question is: if I turn off the oven, it would be possible to observe the subsequent cooling phase?
how can I do to simulate this condition?
I would definitely warm the potato, for example, for 5 seconds, then turn off the oven and watch how quickly they reach a certain temperature.

I tried to use a function to the input power ranging from 500W to 0W and simulate the turning off the oven, but I have not found a solution. i used piecewise or step function.

can someone help me please?

thanks

Acoustic Band Structure in 2d

Mon, 28 Sep 2009 16:51:18 +0000

Hi Folks - Ive posted a couple of questions to some related discussion but it looks like they haven't been visited in a while so I thought I would make a new discussion group.

I am interested in finding out whether or not it is possible to calculate acoustic band structures using a reciprocal lattice simulation and periodic boundary conditions. I understand this is typically done by solving for eigan solutions to the helmholtz equation as you vary the direction of the plane wave vector. My question is... while performing an eigan value analysis and say, computing 100 eigan values in the vicinity of the stop band, how does one back out the wave vector direction related to the given eigan values

Thanks

Sean