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

Internal error in SPOOLES linear solver.

March 27, 2017, 11:00 am
$
0
0
Hi everyone,

When solving the optimization problem, an error occurred as below,

Internal error in SPOOLES linear solver.
-Feture: Optimization Solver 1

But I do not know what happened and how to fix it.

Jiaqing
↧
Search

"Error: Duplicate variable name."

February 4, 2013, 5:40 am
$
0
0
I am using solid mechanics, electrostatics and moving mesh and I want to perform an Eigenfrequency Study. The solver crashes with an error message:

"Error: Duplicate variable name."

But I didn't find any variable names to duplicate.
Below is my file.
↧
↧

Failed to find a solution. There are 1 void equations

August 10, 2014, 6:38 pm
$
0
0
Hello,

I'm running a flow in a pipe and I'm using the Conjugate heat transfer physics. My model converges and gives me results when the fluid in the pipe is water. As soon as I change the fluid to Carbon Dioxide, I get the following error:

Failed to find a solution.
Singular matrix.

There are 1 void equations (empty rows in matrix) for the variable comp1.nitf.Pmfinl1.
at coordinates: (0,0,0), ...
There are 28 void equations (empty rows in matrix) for the variable comp1.p.
at coordinates: (0,-1.82291,0.0448905), (0,-1.8415,-6.245e-017), (0,-1.778,0.0635), (0,-1.82291,-0.0448896), (0,-1.73308,0.0448825), ...
There was an error message from the linear solver.
The relative residual (11) is greater than the relative tolerance.
Returned solution is not converged.
- Feature: Stationary Solver 1 (sol2/s1)


Can anyone tell me why when changing fluids my model wouldn't converge even though everything else is still the same? And what is the variable comp1.nitf.Pmfinl1?

Thank you
↧

Free Convection: Stationary or Time-Dependent Study

June 4, 2015, 3:14 pm
$
0
0
Dear all,

So far in all my free convection simulations I have used Time-Dependent Study but it takes a lot of time, weeks and even months to get the results. The point is I am not interested in transient study, I am looking for steady-state solution. I have seen some models from Application Gallery (Natural Convection Cooling of a Vacuum Flask, Free Convection in a Water Glass, etc.) and I have replicated them using my case, however, it is not converging. But if run the same model with time-dependent study Comsol is solving it but too slowly as I mentioned before. Any idea and/or help will be appreciated.

Regards,
Akim
↧

Time dependent pulse ultrasound wave

January 24, 2017, 4:16 pm
$
0
0
Hi,

I have a time-dependent model to simulate ultrasound wave distribution.
My pulse function is "sin(2*pi*10^6*t)-step1(t[1/s])*sin(2*pi*10^6*t)" from t=0 to 0.001 and step function is 0 to 1 at 0.0002. Output is 200 micro-second of sine waves with 1 MHz frequency and then it turns off for the 800 microseconds.

I used 8 time steps for each sine wave and this increased the computation time significantly. When the study was done and I saved the file, the file was 50 GB!!! Is there a way to decrease the computation time? Am I doing something wrong?

Another question: Is there any way to solve this model in the frequency domain with a pulse like above? The pulse in the frequency domain will have a frequency of 1KHz. It means that one cycle per milli-second. Each cycle composed of 200 microseconds of sine waves and 800 micro-seconds of zero pulse.

Please see the model using link below (I shared it in my google drive).
drive.google.com/file/d/0B1tLR...9ZZjM4VEVkTHc/view?usp=sharing

Thanks,
↧
↧

Optimization with Particle Tracing Module

January 27, 2017, 5:11 am
$
0
0
Hi everyone,

As mentioned, I would like to use COMSOL's optimization module to apply to Particle Tracing for Fluid Flow physics interface. The model refers to "Acoustic Levitator" of the model libraries. The model solves the Pressure Acoustics, Frequency Domain to calculate the pressure p, then apply this interface on Particle Tracing for Fluid Flow to trace the trajectories of the particles. Two studies are used included: Study 1 for Frequency Domain (calculate pressure in Acoustic Pressure, Frequency Domain interface), Study 2 for Time Dependent (trace the particle's position in Particle Tracing for Fluid Flow).

The optimization is maximizing Laplacian of acoustophoretic potential energy (in Particle Tracing Module) at a specific point. To do that, I've used the probe objective with the expression d(d(fpt.Urad),x),x)+d(d(fpt.Urad),y),y) with a random probe coordinates. I've also used the Global Control Variable to define the phase of the transducer (initial value= 60[deg]; bound range from -180[deg] to 180[deg]). However, when computing, it is said the "Only the BDF solver support analytic gradient". Therefore, I try "BDF" and "Initialization only" in the "Time-Dependent Solver" but none of them works.
I' ve tried to plot the particle trajectories with the above expression and the plot works well. I suspect the problem is because of the mathematical expression of potential energy fpt.Urad contains two terms that are in the form of real(a*conj(b)) but I'm not so sure. So the question is why I can't perform the optimization and do you have any suggestions?

I would very much appreciate if anyone could help me. The attached below are some settings in the model.

Sincerely,
Quoc
↧

Ultrasound perssure output

January 31, 2017, 7:33 am
$
0
0
Hello,

I am trying to reproduce a result from a paper (Please see figure 1).
The first thing that I tried was a model that include only transducer and water. As you can see in Figure 2, I got some waves that come out of transducer at an angle. I just want to have waves in the -z direction. How can I get rid of those angled waves?

Thanks,
↧

Duplicate parameter/variable name.

February 7, 2017, 11:35 pm
$
0
0
I have encountered an error which says that,
Duplicate parameter/variable name.
- Variable: r
- Variable: rTIME
- Variable: rxi1
- Variable: rxi1xi1
- Variable: rxi1xi2
- Variable: rxi2
- Variable: rxi2xi1
- Variable: rxi2xi2
- Variable: z
- Variable: zTIME
- Variable: zxi1
- Variable: zxi1xi1
- Variable: zxi1xi2
- Variable: zxi2
- Variable: zxi2xi1
- Variable: zxi2xi2
Error in multiphysics compilation.
- Feature: Compile Equations: Time Dependent (sol1/st1)



Kindly help me.

I am attaching the file.........
Please find the attachment and help....
↧
Search

Point plotting within a volume

March 12, 2017, 3:49 pm
$
0
0
Hello, thanks for looking at this,

I want to get tabled results at a certain point within my model. I create my multiphysics simulation by first importing a mesh from NASTRAN, and then simulating (so I normally can't create points from geometry this way).

Is it possible to take the geometry I get from a mesh import, select a point in 3D, and then get a output value (in my case: pressure)? I can view the volume and surface normally, but this is for some post processing.

Thanks for your help!
↧
↧

error

March 14, 2017, 10:41 am
$
0
0
Good morning everyone :)
I am a PhD student in engineering energy.
I work on the modelling of turbulent combustion of solid waste applied in rotary kiln using Comsol,
My problem is: when I click 'solve' it gives me error: incorrect property value. property: subjtech.
error in automatic sequence generation.
can any one help me please????
↧

Surface tensile stress vs. Radius of curvature in thin films

March 26, 2017, 10:26 am
$
0
0
Dear All

I'm trying to simulate the 3D-surface tensile stress distribution of thin films when it bent with a radius of curvature R, but I don't have any idea of how I can set the radius of curvature in the model.
Any help is greatly appreciated!

Regards,
Hiroyuki Kuwae
↧

Take the surface average on the outlet and inlet boundaries

March 27, 2017, 11:32 pm
$
0
0
Dear All,
I am trying to create a loop in MATLAB to control my COMSOL simulation. I am not good at using the COMSOL operators. Can any one tell me how to take the average of my variables at the boundary within the loop. For instance, I need to calculate the difference of my inlet and outlet concentrations within the loop.


.
Regards
↧

Natural convection

March 28, 2017, 3:03 am
$
0
0
Hi I'm beginner with comsol. I want to simulation the natural convection in the air. Heat transfer in fuid with laminar flow, It's works with short time (for 1 or 2 min) but it never give me results for more then 2 min. I have all the time this error message "Repeated error test failures. May have reached a singularity.".
Any idea why this is happening.
↧
↧

Electrochemical Machining

March 28, 2017, 3:14 am
$
0
0
Can anyone help me to do simulation of electro-chemical machining process in 3D. I am new to COMSOL, i am confused with how to assign material removal rate in 3D. I have done simulation in 2D.
↧

Piezoelectric stator eigen frequency study

March 28, 2017, 4:31 am
$
0
0
Hi,
I am modelling the resonance behavior of a piezoelectric stator which is a metallic structure bonded to a piezoelectric material. Under the Electrostatics physics module, I have a voltage terminal type boundary condition on the piezoelectric material top surface and a ground type BC on the opposite face.
The simulation run successfully and I get as result the set of resonance modes and there corresponding frequencies. However, I noticed that if I remove the terminal voltage BC, the frequencies of the resonance modes shift with non negligible values (500 Hz for a resonance frequency of 23 kHz).
From my point of view, the resonance frequencies depends only upon the structure dimensions and materials and have nothing to do with the presence or not of a terminal voltage.
If anyone can explain this phenomenon, I will be so thankful.
↧
Search

Sharing Knowledge HPC Linux Comsol with matlab

March 28, 2017, 5:01 am
$
0
0
Hallo everybody,
i had a lot of problems to start my Comsol Simulation via matlab with an accurate cpu performance. It took me a lot of time to figure out what I was doing wrong. Therefor I would like to share my experience:
1. the usage of the MUMPS solver was three times faster than the PARADISO solver
2. if i used -clustersimple as an shell option for comsol, the CPU ussage was very strange. It seemed that the one job ran on every cpu at once (every cpu 100%) --> no sharing. When i compared the performance with an normal -nn 1 -np 16 start the Simulation was solved three times quicker as well.

My job had to be started via matlab. Therefore i used the following Shell-Script

#!/bin/bash -l
#PBS -l nodes=1:ppn=16
#PBS -l walltime=24:00:00
#PBS -N Testjob

cd $dir
module load matlab
module load comsol

comsol -nn 1 -np 16 server &
sleep 10
matlab -nodisplay -nosplash -nn 1 -np 16 -r "run('$dir/Simufile.m')"

regards
Florian
↧

Metal Forming (rolling)

March 28, 2017, 6:31 am
$
0
0
Is there anyone can build a model about metal rolling in COMSOL, just like the FEM model in Ansys or Abaqus??

Thank you
↧
↧

How to insert a workplane into a object and define Neumann boundary on the workplane.

March 28, 2017, 6:46 am
$
0
0
Hi
I am studying diffusion by using COMSOL. I have a large object and I divided it into smaller parts by defining workplanes and using these workplanes to do partition of the object. After partition, I can define concentration on these planes, but can not define flux on these planes ( as you can see in the attachment, when define fluxes, the plane 6,9,12 are not applicable.). I know this is because these planes are interior, so fluxes can not be defined on them due to continuity considerations. I also know that when finalizing geometry, we can change the method from 'form union' to 'form assembly' to change the interface from interior to external, so we can define fluxes on it. But my question is that, at the first step, I have a large object to be divided instead of several small objects to be combined together. Does anybody know how to insert a workplane into a object and define Neumann boundary (fluxes) on the workplane? Or maybe divide a large object into smaller ones and therefore can define fluxes on the interface. Thanks a lot!
Jiangjin
↧

Optical Transition in Semiconductor

March 28, 2017, 7:00 am
$
0
0
Hi;
I am rather new to Comsol and I was hoping you guys could kindly help me get started!
I am trying to model a silicon p-n junction to study the transient drift-diffusion after I optically excite the junction by a Gaussian femtosecond laser pulse.
I have a preliminary model, but the study only shows the transient for the formation of the junction potential (like when you just bring the n-type and p-type together and then the junction forms) and does not show anything about the indirect optical transition and the following processes. I was wondering how I could force the study to start with the fully equilibrated p-n junction which is now being optically excited.
I introduce the pulse by using three Gaussian functions (gb1(x)*gb2(y)*gb3(t)) and then use the f(x,y,t) as my E0 which the only place I can use is E0 in the semiconductor model, which I know is not correct (but I couldn't find any other place to use it)

Thanks a lot,
Abe
↧

Comsol Cluster computing, run a model

March 28, 2017, 7:19 am
$
0
0
Hello!

I have a problem: I've installed Comsol on Linux supercomputer, but when I tried to run it like:

> comsol batch -inputfile 2_1_dz.mph -outputfile 2_dz_solved.mph

there is an error:

Output:
*******************************************
***COMSOL 5.2.0.220 progress output file***
*******************************************
Tue Mar 28 17:25:11 MSK 2017
COMSOL 5.2 (Build: 220) starting in batch mode
Opening file: /home/Artem.Samtsevich/COMSOL/tmp/2_1_dz.mph

/******************/
/*****Error********/
/******************/
com.comsol.design.migration.DeployedMigrationHelper.getMigrationHelpers(Unknown Source)
com.comsol.migration.h.e(Unknown Source)
com.comsol.migration.f.d(Unknown Source)
com.comsol.migration.f.f(Unknown Source)
com.comsol.model.util.ModelInternalUtil.load(Unknown Source)
com.comsol.model.util.ModelInternalUtil.load(Unknown Source)
com.comsol.model.util.ModelUtil$27.a(Unknown Source)
com.comsol.model.util.ModelUtil$27.run(Unknown Source)
java.security.AccessController.doPrivileged(Native Method)
com.comsol.model.util.ModelUtil.load(Unknown Source)
com.comsol.batch.BatchApplication.doStart(Unknown Source)
com.comsol.util.application.ComsolApplication.doStart(Unknown Source)
com.comsol.util.application.CsBaseApplication.start(Unknown Source)
org.eclipse.equinox.internal.app.EclipseAppHandle.run(EclipseAppHandle.java:196)
org.eclipse.core.runtime.internal.adaptor.EclipseAppLauncher.runApplication(EclipseAppLauncher.java:134)
org.eclipse.core.runtime.internal.adaptor.EclipseAppLauncher.start(EclipseAppLauncher.java:104)
org.eclipse.core.runtime.adaptor.EclipseStarter.run(EclipseStarter.java:380)
org.eclipse.core.runtime.adaptor.EclipseStarter.run(EclipseStarter.java:235)
sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
sun.reflect.NativeMethodAccessorImpl.invoke(Unknown Source)
sun.reflect.DelegatingMethodAccessorImpl.invoke(Unknown Source)
java.lang.reflect.Method.invoke(Unknown Source)
org.eclipse.equinox.launcher.Main.invokeFramework(Main.java:648)
org.eclipse.equinox.launcher.Main.basicRun(Main.java:603)
org.eclipse.equinox.launcher.Main.run(Main.java:1465)
Total time: 2 s.

What is it the problem can be?
↧
Viewing all 26704 articles
Browse latest View live
Search
<script src="https://jsc.adskeeper.com/r/s/rssing.com.1596347.js" async> </script>