Quantcast
Channel: Latest Discussions - COMSOL Forums
Viewing all 26704 articles
Browse latest View live

Coaxial Probe in Concrete

$
0
0
Hi,

I am very new to COMSOL I was wondering how you would model a coaxial probe in concrete sample and see the wave propagation through the concrete?

Thanks a lot!

Implementing integral across time using analytic temperature-based function

$
0
0
Hello,

I have defined a global analytic function f(T) which is dependent on temperature. I need to integrate this function across time for a cut point in the model. The bounds of the integral are temperature values (where the temperature is itself a function of time, evaluated at a single spatial point in the model). Is there any way to have Comsol evaluate this?

Thank you!

How to insert hard differential equations (containing two variables) in COMSOL 5.2a

$
0
0
Hi Everyone,

I attempt to program new physics in my two-phase flow + phase field model. I search since several days how to inserts my partial equations in COMSOL 5.2a.
The concentrated form of my dimensionless equation is:

1- d(xi*F)/dt+Nabla(xi*F*u)=(1/Pe)*Nabla(xi*Nabla(F))-h*delta*J(F,f)
2- d(delta*F)/dt+Nabla(delta*F*u)=(1/pe)*Nabla(delta*Nabla(F))+h*J(F,f)
Where :
u is velocity so dependent variable of Laminar two phase flow
xi=1-c WITH c is an dependent variable of another physic (phase field)
delta=(3*sqrt(2)/a)*c^2*(1-c)^2
J(F,f)= Bi*(k*F*((1-l*c)/l)-f)
Bi is "0" when xi=1 and "1" when xi=0. Bi can take 0 and 1 only
k,h,pe,Pe and l are all constant.

when I try to do differentiation by part, I get:

1- xi*dF/dt + F*dxi/dt + F*u*Nabla(xi) + xi*Nabla(F*u) – (xi/Pe)*Nabla(Nabla F) - (1/Pe)*Nabla F * Nabla (xi) = – h*delta*J
2- delta*dF/dt + F*d(delta)/dt + F*u*Nabla(delta) + delta*Nabla(F*u) – (delta/pe)*Nabla(Nabla F) - (1/Pe)*Nabla F * Nabla(delta) = h*J

Dividing eq.1 by x1 and eq.2 by delta, I get:

1- dF/dt + Nabla(F*u) – (1/Pe)*Nabla(Nabla F) + F*a + beta* Nabla F = FRIGHT
where
a = (1/xi)*dxi/dt + (1/xi)*u*Nabla(xi)
beta = (1/xi)*(1/Pe)*Nabla xi
FRIGHT= – (1/xi)*h*delta*J

2- df/dt + Nabla(f*u) – (1/Pe)*Nabla(Nabla f) + f*b + gama* Nabla f = fRIGHT
where
b = (1/delta)*d(delta)/dt + (1/delta)*u*Nabla(delta)
gama = (1/delta)*(1/pe)*Nabla(delta)
fRIGHT= (1/delta)*h*J

Is it correct ?
how to translate in COMSOL 5.2 language ?
And finally, In what option, insert these equations which govern my additional physics ?

Thanks for your response.

Loudspeaker Dodecahedron

$
0
0
Good night.
I want to get a directivity of dodecahedron loudspeaker with Comsol simulation. My simulation model is set up using a combination of the Magnetic Fields interface and the Acoustic-Structure Interaction multiphysics interface. And, I import file from Solidworks ( for geometry). For an information, dodecahedron loudspeaker is 12 driver ( loudspeaker) in the dodecahedron shape. When I meshing this geometry, I have an error "out of memory in mesh memory ". Maybe, anyone can help me to discuss or show me how to simulate it ?

Inductances AC vs DC

$
0
0
Hello everybody!

I've discovered that coil with iron core has different inductance values depending on a study type.

Stationary study produces higher inductance value compared to frequency domain one. Field distribution is also different. DC has more homogeneous magnetic field within iron core while AC has all the field concentrated in a thin surface layer.

So now I have 2 questions:
1. Which of the two inductances is the correct one?
2. As far as I know, maximum achievable magnetic field in iron core lies around 2 T. In AC model I got magnetic flux density up to 5 T within iron core. Since inductance is derived from total magnetic energy and that energy is calculated by integrating all those flux density values, does it mean that I'm getting incorrect inductance values as well?

Parameter settings

$
0
0
Hey !! How could i set a parameter which is increasing with an exponential step??
Thank you

Acceleration missed in the transient modal analysis

$
0
0
Hi all,

I've got a little issue with comsol during the modellation of a relative simple problem.

I have a plate, impacted in the middle of its surface. I would like to make a simulation in two phases. The solid part of the plate, and the air around 20mm from the surface. In this way, I would like to simulate the acceleration of the plate, due to the impact, and the acoustic pressure in its prossimity.

What I particullary care is the damping of the material (5%), and the modal interaction due to materials, geometry and boundary conditions.

I tried the transient modal analysis, but for some reason that I miss, it gives me the displacemente and the velocity of the plate, but not the acceleration. Now, this is not a big issue, since I could derive the velocity plot in time to have the acceleration. The problem is about the acoustci pressure, that seems is not calculated as well as the acceleration.

I tried the only transient analysis too, but in this case the modal shapes neither the damping factor are considered. It returns acceleration and acoustic pressure, but they are strongly affected by the modal and damping errors.

Could someone help me and suggest the best way to simulate a problem relatively simple such this one?

Thank you very much in advance.

Magnetic Fields-No Current, for Permanent Magnets

$
0
0
I am doing magnetic analysis for 16 magnets arranged in linear halbach configuration. i am having difficulty in understanding if COMSOL calculates the flux density by averaging all the flux from magnets separately?? or some other method. Kindly some one guide me.

Hassan

Linearized Navier Stokes Module for Aeroacoustics Simulations

$
0
0
Hi all,

I'm trying to use the linearized NS to model the acoustics modes of a rectangular shape duct .

Could anyone help me on how to use it in COMSOL 5.2a?

I would really appreciate any kind of help

Best,
Mohammad Ghulam

COMSOL 5.2 gets stuck after some time

$
0
0
From almost two weeks I am facing this problem that whenever I run a simulation that requires long processing time, I find that the program stops responding. I just get a "waiting" cursor and a (Not Responding) message in the title bar.

I thought it might be due to the latest update for COMSOL 5.2 so I uninstalled the COMSOL program and re-installed v5.2 without the update. But, still it gave me the same problem. I also tried installing the update for it, and still I have the same issue.

I am running the program on Windows 10 Pro 64-bit with 32 GB RAM and i7-4790 processor with a 4GB NVidia GeForce GT730 graphics card.

It used to run fine before but now its frustrating me because its impossible for me to run anything. I waste hours and end up with nothing because the program just hangs up.

importing from solid works

$
0
0
first of all hi ,

i am new to this so i could use all the help i could get .

i watched a tutorial of how to design a wind turbine to solidworks .

I try to import said geometry to comsol . The problem i face is that it is imported as one item thus , i cannot mess around afterwards with the materials etc.

I have thought of taking each part of the turbine and importing it piece by piece , but if you have a quicker solution , much appreciated

undefined variable error (new to comsol)

$
0
0
undefined variable error

variable: comp1.emi.disp
geometry :geom1
boundary:12

failed to evaluate expression
expression:1000000*comp1.emi.disp
plot:surf1
Trying to get a displacement plot for simple CMUT membrane

extracting results from COMSOL using livelink for matlab

$
0
0
Hi everyone~
It is trying to do shape optimization involving stationary analysis. The code of optimization method was written by Matlab, and it needs to receive the simulation results from the COMSOL.
For example, in the case study of "Multistudy optimization of bracket", I want to extract the max stress value from the COMSOL using matlab, however I found the result was wrong. Does anyone know why?
Many thanks!


Matlab code :

model = mphload('multistudy_bracket_optimization');%load the model

x=[11.8 2.7 9.0 22.3 9.5 20];% parameters

model.param.set('rC', x(1,1), 'Radius of central hole');
model.param.set('zCo', x(1,2), 'Distance from bend to bottom of central hole');
model.param.set('rO', x(1,3), 'Radius of outer holes');
model.param.set('zOo', x(1,4), 'Distance from bend to bottom of outer hole');
model.param.set('yOo', x(1,5), 'Distance from edge to outer hole');
model.param.set('wInd', x(1,6), 'Width of indentation');
model.hist.disable;

model.sol('sol2').runAll;
Max_Stress = mphglobal(model,{'maxop1(solid.mises)'}); % get the max stress value

Use of coefficient form PDE with global equation (simple ODE)

$
0
0
Dear Sir or Madam,
I am trying to solve a coefficient form PDE "dC/dt+ u.dC/dz =D 1/r.d/dt (1/r.(dC/dr)" with a global ODE equation "dCt/dt=constant*(C(t,L) - Ct)" together. I am using 4.1 version so I know that I can not use C(t,L) with ODE so I replaced it with aveop1(C) which is the average operator for C at the location of L. But I am getting a zero answer across the geometry. Could you please someone help me?



How to plot vorticity

$
0
0
I can visualize only pressure and velocity.

Do you have any ideas about how to get the vorticity for a 3D problem?

Time Dependent Study for Loudspeaker Systems

$
0
0
Hi all,

I am interested in performing a Time Dependent study for loudspeaker systems. As a starting point here, I work with the tutorial model “Loudspeaker Driver in a Vented Enclosure”. Given that I have already studied successfully this model as it is i.e. in the Frequency Domain, I have tried a few things which fail.

(1) I added a second study (Step 1: Time Dependent) to the existing model. Given that this is available only for the Shell interface, I deactivated the Pressure Acoustics Frequency Domain Interface (Study 2 -> Step1: Time Dependent -> Physics and Variables Selection). Then, I ticked "initial values of variables solved for" (Study 2 -> Step1: Time Dependent -> Values of Dependent Variables) and I chose “initial expression” (Method), “Study 1, Frequency Domain” (Study) and a single frequency value from the list (Parameter value (freq(Hz))).

This study cannot be computed because the error “undefined variable, -Variable: shell.u_tz” appears. At this point, I wonder why this variable is also recognized as undefined in the Frequency Domain study of the tutorial model (Component1 -> Definitions -> variables -> av_apex(shell.u_tz)), but for this case, the study is computed successfully.

(2) When I changed further the expression for the electric driving force (Component 1 -> Definitions -> variables -> Fe) by replacing “Vo” with “Vo*sin(omega*t)”, I got the same error.

Is there a way to use the solution from the Frequency Domain study to perform a Time Dependent study?

Could be an alternative approach to work with the Pressure Acoustics Transient Interface? Given that PML layer is not applicable for this Interface, how can outgoing waves be absorbed for this case?

I would highly appreciate a response to these issues.

Fabry-Perot cavity with metal mirror on one side

$
0
0
I want to simulate a Fabry-Perot cavity, one mirror is dielectric mirror with multi-stack coatings, but another one is metal mirror. For metal mirror, I need to include Fresnel reflection losses and also attenuation due to imaginary refractive index.

I know there's a tutor example of Fabry-Perot cavity in COMSOL website, it's using geometric optics model, I already build a nice working Farby-Perot model with this model but it's with two dielectric mirror. Now I change one mirror to metal and include the complex refractive index of the metal, but the result is absolutely not right.

If geometric optics model could simulate my case, if yes, how can I correctly set metal mirror and include the losses. And if geometric optics model is not working, what model should I use? Thank you.

Attempt to evaluate real logarithm of negative number.

$
0
0
Hi, All

I got this error when I run the simulation:

" Attempt to evaluate real logarithm of negative number.
- Function: log
Failed to evaluate variable.
- Variable: comp1.semi.Efp
- Defined as: comp1.semi.Ev-(comp1.semi.Vth*log(comp1.semi.P/(comp1.semi.gamma_p*comp1.semi.Nv))) "

I tried to calculate the component in log (i.e., comp1.semi.P/(comp1.semi.gamma_p*comp1.semi.Nv) ) in Comsol Results and the result shows that this component indeed a positive number with a order about 10^-4~10^-9. If so, why I still face this error?

Can anybody help?

Best regards

Particle-Particle Interaction Cutoff

$
0
0
Hello!

I am trying to use the mathematical particle tracing module to accumulate some particles in a material.
Ultimately I would love to be able to spawn particles, have them travel in some direction then stick to a wall. However, I also do not want them to touch each other.

I figured I could make this happen by using the particle-particle interaction node and applying something like damped force when the particles get close to each other to reach an equilibrium distance.

Anyway, I am running into a problem with the cutoff length I think. I've attached a sample file where I release one particle, let it drift, then release another one some time later. I am using the cutoff length setting, so the first particle shouldn't "feel" any interaction force since the second particle hasn't been released yet. However, it does! I'm not too familiar with this physics set so I don't know exactly how it works, but something doesn't feel right.

Any help would be greatly appreciated.

Undefined value found

$
0
0
Hi,

i can solve my semiconductor model with a relatively coarse mesh.
However the found solution is not good enough. It shows some
differences to a reference (PC1D) semiconductor model.

That is why i want to refine the mesh.


But if i refine the mesh i get following error:

---------------------------------------------------------------------
Error: 7248
Undefined value found.

Details:
Exception:
com.femlab.jni.FlNativeException: Undefined value found Solvermodel
Messages:
Undefined value found Solvermodel

Stack trace:
at solvermodel.cpp, row 1837, (Undefined value found in the equation residual vector.
There are 85 degrees of freedom giving NaN/Inf in the vector for the variable phii
at coordinates: (0.000493076,-5.8675e-005), (0.00049406,-5.80051e-005), (0.000495044,-5.73351e-005), (0.000494026,-5.58557e-005), (0.000493041,-5.65256e-005), ...
)
at com.femlab.solver.FlSolver.femStatic(Native Method)
at com.femlab.solver.FemStatic.run(Unknown Source)
at com.femlab.server.FlRunner.run(Unknown Source)
at com.femlab.util.i.run(Unknown Source)
at com.femlab.util.aa.run(Unknown Source)
----------------------------------------------------------------------------

The model is quite complex and has 5 coupled application modes.
I use the Parametric segregated Solver.

I think if there would be a problem with my initialisation values or boundary contitions,
i would not be able to solve the model even with a coarse mesh?

I found out, that the problem especially occurs, if i refine the mesh in a region with large gradients.

The Comsol solution is represented by green circles and the reference solution is shown as line in the attached figure.

Has anyone an idea how the finer mesh causes "undefined value"-problems and how to overcome them?

Thank you,

Stefan







Viewing all 26704 articles
Browse latest View live