Quantcast
Channel: Latest Discussions - COMSOL Forums
Viewing all 26373 articles
Browse latest View live
↧

the maximum number of time steps allowed in transient analysis

August 8, 2012, 11:27 am
$
0
0
Hi,
Comsol gave me a blank error message box when I had 5E6 time steps in my solver configuration. Everything was fine after I reduced the number of time steps from 5E6 to 5000. I am still very interested in solving the program with 5E6 time steps. What is the cause of this error? What is the factor that limits the maximum number of time steps I can define?
Thanks!
↧
Search

Inconsistent Units

December 21, 2010, 2:58 pm
$
0
0

Hi,


I´m trying to incorporate the following expression on Subdomain expressions:

D=0.926*T^2.5/(T*245)

the result should be in mm^2/s, but It shows me "Inconsistent Units"

even when I try the following expression:

T^2.5/T^2.5

It stills to show the same message (inconsistent units)


Best Regards,
Orlando

↧
↧

No-slip on interior boundary

April 8, 2011, 5:07 am
$
0
0
I like to solve (Comsol 4.1) Navier-Stokes (incompressible) in a 2D-domain that has an interior boundary (typically a line starting at some point on the exterior boundary and ending at some point inside the domain). On this interior boundary I like to impose a no-slip condition. However, Comsol does not allow me to do so. It says that this interior boundary is "not applicable". Does anybody know how to enable interior boundaries in order to impose boundary conditions there (I know that this was possible in 3.5)?

Leon
↧

Interpolation Function Help

July 18, 2011, 4:33 am
$
0
0
Good Morning,

I'm having trouble using COMSOL's interpolation function, and I was hoping someone might be able to help. I'm trying to load data from a tab-delimited text file. I've followed the instructions in the COMSOL user guide on interpolation functions, however I get the following error when I click "plot":

"Internal error in mesh initialization.
-Plot: previewplot1 (Surface)
-Function: int1"

Additionally, when I try to solve the model I get the following error message:

"-Feature: Compile Equations: Stationary (sol1/st1)
-Error: Internal error in mesh initialization."

When I disable the interpolation function, COMSOL successfully solves the model.

The input data used for the interpolation function was a velocity distribution exported from a previous solution surface. It has the following structure:

% x y z myfun1
0.005164328 0.094835672 0.2 0.098519135
0.015894103 0.087647192 0.2 0.511590803
0.015569986 0.095339275 0.2 0.22689919
0.006197193 0.083275657 0.2 0.312314232
0.005469283 0.016082383 0.2 0.2672031
0.004733427 0.004733427 0.2 0.081290304
0.012546053 0.01240331 0.2 0.43510989
0.015707096 0.004896326 0.2 0.232560416
0.027164381 0.094602393 0.2 0.368319369
etc.

A very simple flow model demonstrating the problem is attached, along with an excel version of the data I'm using. I'm a new COMSOL user, and any advice would be greatly appreciated.

Thank You!
↧

Rayleigh-Plesset bubble response

August 2, 2012, 5:52 pm
$
0
0
Hello everybody,

I try to model an micro air bubble in water under an incident ultrasonic pressure field. In near future I hope to be able to add walls to the system, therefore I need a model which doesn't use the spherical symmetry of the bubble.
To do that I suppose that the water flow is potential and that the bubble's gas flow is isentropic. So I solve Laplace equation for the velocity potential in the liquid domain and transient Bernoulli's equation at the water-air interface.
For that I've decided to use a domain pde for the laplace equation and a boundary pde for transient Bernoulli's equation. I use a weak form pde to compute surface tension. I follow the interface using moving mesh

For the moment I don't succeed to model the gas bubble correctly for high pressure amplitude. I think mostly because of the non-linear term in Bernoulli's equation. Weak contribution, separate weak equation or explicit adding of this term in a boundary coefficient form pde doesn't succeed. Therefore I don't know what to do. Normally weak form pde's should be able to solve non-linear equations. But for my problem it doesn't work. I'm sure that I've well implemented the model because I have the right response for low amplitudes, so when the non-linear term can be neglected.

Do someone know at least why my model doesn't work or what I should do to make it work with Comsol?
Thank you!

Karl
↧
↧

Gas flow into a reactor

August 9, 2012, 12:26 pm
$
0
0
Hi. I'm new to COMSOL and am puzzled on where to start. I would like to model the flow of a gas into a reactor. The gas would be at a temperature T1 and the contents of the reactor would be at T2 (T1 > T2).I can assume that the reactor is a cylinder as well as the pipes carrying gas in and out of the reactor. What I'm interested in learning is what is the heat exchange in the reactor, i.e. by how much does T2 increase as a function of time At what point does it reach steady state? How do I account for the liquid in the reactor being stirred?

Thank you!
↧

How to plot two or more independent vaiables in one plot?

August 13, 2012, 9:04 am
$
0
0
Hi,
I am solving some PDEs and I want to plot 2 or more independent variables in one plot. how can I do this?
↧

pzt deformed

August 13, 2012, 10:23 am
$
0
0

Dear Comsol,

I'd like to calculate the volume formed by pzt to be deformed, when applying a DC voltage.

How can I make this calculation in comsol ?

Thanks
↧
Search

complete elliptic integral

August 13, 2012, 5:25 pm
$
0
0
Hello everybody,

I need to compute the complete elliptic integral of first kind and of second kind in a complex Comsol model. I didn't find these functions in the mathematical functions page of the documentation. Is there another way to use them in Comsol 4.2?

Thanks a lot

Karl
↧
↧

deformed mesh

August 13, 2012, 9:26 pm
$
0
0
I'm using moving mesh analysis and piezo solid for simulating a piezoelectric cantilever .. I’m able to get the results only when I use the physics separately. If I combined both the physics and tried to run, the result is a wavery deformation.How to visualize the result after every run in v3.5.Is there any need to give the charge Q value forfloating potential. Please anybody help us.
↧

Convection between to domains

August 13, 2012, 10:16 pm
$
0
0
Dear reader,

I am trying to setup a convection between two domains. An issue I have is that the boundary condition "convective cooling", defines the external temperature (or environment temperature) as predefined parameter and fixed number. In my case the external temperature is subject to change and should be able to decrease. This decrease in temperature is a result of the convection between the two domains.

I have added an image explaining the situation.

Thanks for any help!
↧

calculating magnetization?

August 13, 2012, 11:31 pm
$
0
0
Hi,
In Comsol 3.5, I was able to calculate the magnetic moment for an object (in the model below, a sphere) with a constitutive relation B=mu0 mur H, when embedded in a bigger block with B=mu0 H, (ie mur = 1), with a boundary condition on the outside of the block of a magnetic flux density. In postprocessing, I can then select "subdomain integration", and integrate the "magnetization, norm" over the sphere. I then get a value that matches the textbook value for the magnetic dipole moment of a sphere in a magnetic field.

In Comsol 4.1, I cannot select "magnetization" as a variable to integrate over in this case from the "derived values". Is there a switch somewhere that I'm not seeing to do this calculation?

Thanks,

(Just tried to attach a model, but it's too big! Sorry!)

↧

reacf

August 13, 2012, 11:39 pm
$
0
0
I have not found the mathematical formula for the "reacf" operator. What is it?

Another issue: When I integrate "Current density norm" = ec.normJ or reacf(V) over an insulating surface, I get a nonzero value, while the integration of "Normal current density = ec.nJ gives zero, as expected.

Hence, I need a manual where all these quantities are explicitly described. I am using Comsol 4.3 and "Electric Currents" module.
↧
↧

Parameter Estimations by Inverse Analysis using Comsol 4.2.

August 14, 2012, 3:46 am
$
0
0
Dear Sirs,

I would like to have an opinion by more expert users than me. I am facing a problem of parameter estimation by inverse analysis for structural mechanics. More simply, I have to estimate the values for certain parameters of the constitutive law. The observation variable that I look at in order to understand if the current value of the parameter is correct is the displacement field in the domain.

The mathematical formulation of the problem involves a functional J that should be minimised and the so called primal and dual problems.

Have some of you ever tried to solve this type of problem? If yes, which package of Comsol did you use? I think the best would be to use the Sensitivity module that is in the General PDE module but I would like to have some advice.

Many thanks for the help,

I wish you a nice day,

Paolo Tricerri
↧

Number of degrees of freedom solved for: 96195

August 14, 2012, 5:00 am
$
0
0
Dear Sirs,

I am trying to understand from where the number 96195 comes from.

In particular, the mesh I am considering is a tetraedral mesh of a rectangular domain in 3D ( width = 5 cm, depth = 1.5 cm, height = 0.35 cm). The mesh, according to the mesh statistics the domain discretization is made of:

Prism elements: 6600
Triangular elements: 264
Quadrilateral elements: 1900
Edge elements: 276
Vertex elements: 8

I am using nearly-incompressible model for the material, there a mixed formulation is used to solve the problem and the pressure field is computed (which is not specified but since I use quadratic polynomial functions, it could be P1).
According to the formula presented in http://www.comsol.com/support/knowledgebase/875/ the number of nodes should be 6600*1.4 = 9240 and the DOFs (for the displacement) = 9240*3 = 27720. Adding the DOFf for the pressure (which is P1) we do not get 96195.

As someone an explanation for the number of DOF???

Many thanks,

Paolo
↧
Search

Modal Mass calculation

August 14, 2012, 5:06 am
$
0
0
Hi

Is it possible to calculate the modal mass for a particular mode shape? Can someone please explain how to do it.
↧

Time Integral in Frequency Domain

August 14, 2012, 6:17 am
$
0
0
Dear All,

I have got a simple geometry: a 1mm thick piston-disc is vibrating with a frequency of 1kHz at 1cm above the closed bottom of a (L=353mm long, Rc=38mm radius) cylinder. The maximum acceleration of the piston is a0=10m/s^2. The top of the cylinder is open and it is an acoustically matched layer, which means there is no wave reflection from this boundary. The fluid is air within the cylinder.

The study was performed in 2D Axisymmetric; Pressure-Acoustic, Frequency domain, and also in Acoustic-Solid Interaction, Frequency domain. The aim is to test the validity of the results with this simple setup, and also to compare the radiated acoustic power that leaves the cylinder with the mechanical input power required to move the piston.

It is simple to get the radiated power by multiplying the Rc^2*pi area with the Intensity magnitude (RMS), which gives about 2.37 uW.

It is a bit more complicated to calculate the mechanical power input, or at least I couldn't figure out a more simple way yet. In the variables used the Fp=F_int(2*pi*r*nz*(down(p)-up(p))) definition where F_int is a boundary integral applied on the piston's surface. After solving the study, the Fp acoustic pressure force on the piston is displayed on a Global 1D Plot using phase as the Parameter on the x axis. This plot is exported into Excel and now we have got all the values of the Force as the function of phase accessible for integration. The piston's displacement as the function of phase is calculated as Zph=-a0/(2*pi*freq)^2*cos(phase). By summing up all the F(phase)*dz we get the consumed mechanical energy during a full oscillation period which is -2.37E-9 J. Multiplying this with the frequency we obtain the consumed mechanical power of -2.37uW. As you can see it is not easy and simple to calculate the mechanical power this way.

My question is, whether there is any easier and simpler way to get the invested mechanical power in a frequency domain study?

I have not done this project in a transient study yet, but I suppose it should be straightforward there to get the mechanical power directly from COMSOL. Any suggestions how to do such "power-In vs. power-Out" analysis in the simplest way possible, without the need to export and externally calculate?

Thanks for suggestions,
Joe
↧
↧

what is COMSOL plotting when I surface plot a complex quantity in time harmonic analysis?

August 14, 2012, 8:57 am
$
0
0
I'm new to time harmonic analysis. I guess the answer to this question is obvious to many of you that are familiar with time harmonic analysis.

I am doing a 2D time harmonic simulation of a coil-over-plate setup with perpendicular currents. The AC current in the coil induces eddy currents in the plate.

When I plot the induced current density in the plate Jz_emqa (which should be complex valued) using "Surface Plot" in Postprocessing, what is COMSOL plotting by default? Is it the real part of the complex current? Should not be the norm of current since it has +- signs.

Thanks for any hint!
↧

Help Needed: Comsol solving Error

August 14, 2012, 9:14 am
$
0
0
Hi
I am new to Comsol Multiphysics 4.2a. I built a model & try to run it but after clicking compute and it gave me an error
specifying

the following feature has encountered a problem:
-Feature: Stationary Solver 1 (sol1/s1)
-Error: Unknown Function.

↧

Weakly compressible Navier-Stokes Equation

August 14, 2012, 12:06 pm
$
0
0
Can Acoustic module be used with Weakly compressible Navier-Stokes equation module to simulate the plate's natural frequency due to sound wave under fluid flowing over it?
↧
Viewing all 26373 articles
Browse latest View live
Search