<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
 <channel>
  <atom:link href="http://static2.comsol.dk/community/forums/general/rss/feed.rss" rel="self" type="application/rss+xml"/>
  <title>Latest Discussions - COMSOL Forums</title>
  <link>http://static2.comsol.dk/community/forums/</link>
  <description>Most recent forum discussions</description>
  <pubDate>Fri, 18 May 2012 10:09:42 +0000</pubDate>
  <image>
   <title>Latest Discussions - COMSOL Forums</title>
   <url>http://static2.comsol.dk/shared/images/logos/comsol_logo.gif</url>
   <link>http://static2.comsol.dk/community/forums/</link>
  </image>
  <item>
   <title>Ideal mesh for a regular grid export</title>
   <link>http://static2.comsol.dk/community/forums/general/thread/28716/</link>
   <description>Dear Comsol community,&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
I have a simple 3D electrostatic configuration of two separated point charges in free space. What I want to do is solving the poison-equation and exporting the data in small cubic region between the point charges on a regular grid.&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
What is the optimal mesh for exporting on a regular grid with about 201^3 points? &lt;br /&gt;&#13;
&lt;br /&gt;&#13;
Since the error of the FE-method is minimal (vanishing) on the nodes of the mesh-elements I think it would be perfect to have a regular mesh with nodes exactly on the points of the regular grid. I found no way to achieve this yet and I would be thankful for any kind of help giving me a good exported result.</description>
   <pubDate>Fri, 18 May 2012 10:09:42 +0000</pubDate>
   <guid isPermaLink="false">.1337335782.28716</guid>
  </item>
  <item>
   <title>Designing Voice Coil Actuator with Flexures</title>
   <link>http://static2.comsol.dk/community/forums/general/thread/28708/</link>
   <description>Hi,&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
I am new to comsol and am trying to design a linear voice coil actuator by fixing each end to two diaphragm flexures, therefore keeping everything axisymmetric. I have designed the flexures with the structural mechanics and have placed loads and found eigenfrequencies.  I am wanting to know how to incorporate a voice coil actuator to drive the flexures?  I searched for some similar models without any luck.  The ideal setup would be applying some step function to the voice coil and get the response of the system, and tune it with a PID controller.  Any help whatsoever would be great as I am just getting used to the new software.&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
Thanks!</description>
   <pubDate>Fri, 18 May 2012 00:54:41 +0000</pubDate>
   <guid isPermaLink="false">.1337302481.28708</guid>
  </item>
  <item>
   <title>AMD vs. Intel (specs better for COMSOL?)</title>
   <link>http://static2.comsol.dk/community/forums/general/thread/28706/</link>
   <description>We run a muli-user lab and run COMSOL on 3-4 machines. All of them are PC's with 8 cores and 24GB of memory or there abouts. We are seeing increased use of COMSOL and our users are running into slow simulation times on some of their COMSOL projects, so we are budgeting a new server class PC. We are wondering AMD vs. Intel.&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
We are aware of this knowledge base thread but are looking for more detail: http://www.comsol.com/support/knowledgebase/866/.&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
PC #1: HP ProLiant DL585 G7 Server&lt;br /&gt;&#13;
(4x) AMD Opteron 6272 (16 core 2.1GHz) (total of 64 cores)&lt;br /&gt;&#13;
256GB PC3-12800R Memory&lt;br /&gt;&#13;
(4GB of mem per 2.1GHz core)&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
PC#2: HP ProLiant DL580 G7 Server&lt;br /&gt;&#13;
(4x) Xeon E7-4850 (10 core 2.0GHz) (total of 40 cores)&lt;br /&gt;&#13;
256GB PC3-10600R Memory&lt;br /&gt;&#13;
(6.4GB of mem per 2.0GHz core)&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
PC#3: just like PC#1, except with 512GB of PC3-12800R Memory&lt;br /&gt;&#13;
(8GB of mem per 2.1GHz core)&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
Either system would have an Nvidia Quadro 6000 card in it.&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
An interesting note is that the above AMD system is much more ecconomical than the Intel. The cost of the 512GB AMD server is about the same as the 256GB Intel. Are Intel chips just that much faster? Or is the faster bus speed of AMD (12800R) more important?&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
I'm not sure which modules in COMSOL are most important to our users although I could find out. However, we definitely need to be flexible because users/researchers come in and out of our labs so we may need to use many different modules from time to time.&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
Thanks for any tips!&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
Alex</description>
   <pubDate>Thu, 17 May 2012 16:15:33 +0000</pubDate>
   <guid isPermaLink="false">.1337271333.28706</guid>
  </item>
  <item>
   <title>How to view cross section temperiture distribution?</title>
   <link>http://static2.comsol.dk/community/forums/general/thread/28705/</link>
   <description>Dear all,&lt;br /&gt;&#13;
    I finished the calaulation of a very simple model of Heat Transfer.  How to view the cross sectional temperiture distribution?&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
Thanks&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
Baofeng&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
</description>
   <pubDate>Thu, 17 May 2012 15:46:59 +0000</pubDate>
   <guid isPermaLink="false">.1337269619.28705</guid>
  </item>
  <item>
   <title>More domains generated once elements exceed  32 by 32 </title>
   <link>http://static2.comsol.dk/community/forums/general/thread/28703/</link>
   <description>Dear all,&lt;br /&gt;&#13;
     I am modeling a panel with 40 by 40 holes. I use Block to get the 3D panel, and use Cylinder and Array to get 40 by 40 cylinders in the panel. But when I define the material properties of the solids, I found one more block solid was generated automatically. I try many times and find that once the numbers of the cylinders exceeds 32 by 32, the unexpected solid domain was then appeared. Therefore, I can not mesh the object once more. &lt;br /&gt;&#13;
    Are there somebody faced this problem before?&lt;br /&gt;&#13;
    I use COMSOL 4.2, this problem bother me for several days, I am looking forward to get help to this. &lt;br /&gt;&#13;
 &lt;br /&gt;&#13;
Thank you very much!&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
Baofeng</description>
   <pubDate>Thu, 17 May 2012 15:12:07 +0000</pubDate>
   <guid isPermaLink="false">.1337267527.28703</guid>
  </item>
  <item>
   <title>how to cut or trim a parametric curve ?</title>
   <link>http://static2.comsol.dk/community/forums/general/thread/28699/</link>
   <description>&lt;br /&gt;&#13;
Hi,&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
I'm trying to make a 3D coil using parametric curve and sweep, but I need to cut a part of my curve, I can't change it by equation, so i just need to cut a part of it. In solidworks we used to do this by trim. Any Idea how can i do the same  in COMSOL.&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
Indeed is there any changes or difference in results, when we do 2D modelling on MEF module for coil design with 3D modelling? &lt;br /&gt;&#13;
&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
Thanks in Advance</description>
   <pubDate>Thu, 17 May 2012 14:31:32 +0000</pubDate>
   <guid isPermaLink="false">.1337265092.28699</guid>
  </item>
  <item>
   <title>Modelling a moving, breathing animal</title>
   <link>http://static2.comsol.dk/community/forums/general/thread/28694/</link>
   <description>Hello,&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
I am rather new to COMSOL and so far I have not been able to pinpoint how to model an animal moving inside of a closed box. I have tried to use parametric sweeps but I don't think that is quite what I am looking. Basically, I need an animal (modelled as a solid ellipsoid) to move around the bottom of the box and then view the rather delicate pressure perturbation incurred by this movement.&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
The next step of the project would be to simulate the animal breathing and view the pressure &amp;amp; velocity changes in the box as the animal breaths in and out.&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
Any suggestions as to how to start? i.e. stick to the parametric sweeps and model the breaths as an interior wall with periodic flow?&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
Thanks,&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
Arman</description>
   <pubDate>Thu, 17 May 2012 14:14:34 +0000</pubDate>
   <guid isPermaLink="false">.1337264074.28694</guid>
  </item>
  <item>
   <title>Fluid Shear Stress in Thin Film Fluid Module</title>
   <link>http://static2.comsol.dk/community/forums/general/thread/28690/</link>
   <description>Hello,&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
I'm currently working with the journal bearing model (Comsol 4.2) from the model library and I'm trying to calculate the shear stress within the fluid film with is an essential for the torque calculation. Unfortunatly the shear stress is not one of the output values of this module.&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
I tried the surface integration and using the simple formula:&lt;br /&gt;&#13;
tau = eta*du/dh&lt;br /&gt;&#13;
with dh = tffs.dh &lt;br /&gt;&#13;
and eta = constant value&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
I used du = d(tffs.uy,y)+d(tffs.ux,x)&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
Unfortunately the results don't seem logical. I believe my expression for du or my usage of the integration (under results, derived values) is wrong.&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
Please help!&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
Thanks,&lt;br /&gt;&#13;
Toan&lt;br /&gt;&#13;
</description>
   <pubDate>Thu, 17 May 2012 13:49:04 +0000</pubDate>
   <guid isPermaLink="false">.1337262544.28690</guid>
  </item>
  <item>
   <title>3D coil inductance with frequency domain</title>
   <link>http://static2.comsol.dk/community/forums/general/thread/28688/</link>
   <description>I modeled a coil to calculate the inductance in both 2d symmetric axis and 3d with frequency domain. It gives a rise an issue that the value of calculated inductances are different. &lt;br /&gt;&#13;
&lt;br /&gt;&#13;
Since there is no directly option for inductance in 3D, I use the following fomula:&lt;br /&gt;&#13;
L=N*flux/I0&lt;br /&gt;&#13;
L: inductance&lt;br /&gt;&#13;
Flux: integration of magnetic flux density over a closed surface&lt;br /&gt;&#13;
N: number of turn&lt;br /&gt;&#13;
I0: the applied current&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
The value of inductance in 2D is equal to 5.0408e-5H, while in 3D is 4.226e-5H.&lt;br /&gt;&#13;
However, it gets closer to results from 2D when I changed frequency domain to stationary domain in 3D which is 5.0803e-5H.&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
Can anyone give me a favour?&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
Regards,&lt;br /&gt;&#13;
Hongyu Li</description>
   <pubDate>Thu, 17 May 2012 13:28:28 +0000</pubDate>
   <guid isPermaLink="false">.1337261308.28688</guid>
  </item>
  <item>
   <title>analyze movement of particles in ACDC</title>
   <link>http://static2.comsol.dk/community/forums/general/thread/28687/</link>
   <description>Hello everyone&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
I have a question and I wondered if anyone can give me some info.&lt;br /&gt;&#13;
I am using AC/DC module right now. However I want to add movement of A charge in my model and analyze the effect of its movement on electrical field of whole system. &lt;br /&gt;&#13;
Do you think it is possible to do these kind of simulations by using time dependent analysis? Because I have checked different models here that use time dependent analysis but their goal was totally different from mine.&lt;br /&gt;&#13;
I think in order to do this I should have access to particle tracing module. Am i right?&lt;br /&gt;&#13;
I would be really grateful if someone can give me some information or give me a link of example related to this.&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
Best&lt;br /&gt;&#13;
Bahar</description>
   <pubDate>Thu, 17 May 2012 12:22:26 +0000</pubDate>
   <guid isPermaLink="false">.1337257346.28687</guid>
  </item>
  <item>
   <title>Comsol 4.2 on big simulation machine</title>
   <link>http://static2.comsol.dk/community/forums/general/thread/28685/</link>
   <description>Hi to everybody,&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
I'm using Comsol 4.2 to run CFD and Conjugate Heat transfer complex models. First we used a machine called Machine1 (in the following text) which is a workstation. Than we bought  a new machine (Machine2) to have more power and  reduced analysis time. But this is not happened! The &amp;quot;expected to be faster&amp;quot; machine is actually slower.&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
Machine1 (Workstation):&lt;br /&gt;&#13;
Name: Dell T7500,&lt;br /&gt;&#13;
CPU: Intel Xeon E5620 2.4 Ghz,  2 processors (4 core each); RAM: 48 GB Ram Memory; OS: Windows 7 64 bit&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
Machine2  (Server):&lt;br /&gt;&#13;
Name: rack type Dell R715; CPU:  Amd Opteron 6282 SE 2.6 GHz  with 2 processors (16 cores each), RAM: 256 GB Ram Memory; Os: Windows Server 2008 R2 Datacenter Edition&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
Someone had experiences with this type of big machines. Me and my colleagues are mad about trying to configure the system. &lt;br /&gt;&#13;
We are thinking about some more optimized instruction to drive the calculations, to change the OS in a Linux Suse... and many others tests.&lt;br /&gt;&#13;
Any suggestion will be greatly appreciated!!!&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
Thanks!</description>
   <pubDate>Thu, 17 May 2012 11:21:20 +0000</pubDate>
   <guid isPermaLink="false">.1337253680.28685</guid>
  </item>
  <item>
   <title>Where there are two similar domains? in comsol 4.2</title>
   <link>http://static2.comsol.dk/community/forums/general/thread/28674/</link>
   <description>Dear all,&lt;br /&gt;&#13;
   I just created a solid use Block, and created 1600 smaller cylinders inside the block. When I mesh  the global model. COMSOL 4.2  gave me an information that the model have two similar block domains. And they cann't be meshed. &lt;br /&gt;&#13;
   Are there anyone can help me out of it?&lt;br /&gt;&#13;
Thank you very much!!&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
Baofeng&lt;br /&gt;&#13;
   </description>
   <pubDate>Wed, 16 May 2012 19:46:53 +0000</pubDate>
   <guid isPermaLink="false">.1337197613.28674</guid>
  </item>
  <item>
   <title>simulation of eddy current damper</title>
   <link>http://static2.comsol.dk/community/forums/general/thread/28638/</link>
   <description>hi all &lt;br /&gt;&#13;
i am working on eddy current damper....in which beam (cantilever type) vibrations (free) are to be damped...the material of beam is aluminium....i have simulated the two electromagnets..but i am not able to simulate the beam vibration with initial conditions @t=0 displacement of free end is 0.003m and velocity =0. i have used the eigen frequency study but the tip displacement is very large as per our initial condition....when i used the time dependent study it gives an error of singular matrix...and one thing more should i use the moving mesh for beam...&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
plz help&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
i've attached the model...&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
Thanks and regards&lt;br /&gt;&#13;
Mohit</description>
   <pubDate>Wed, 16 May 2012 03:59:16 +0000</pubDate>
   <guid isPermaLink="false">.1337140756.28638</guid>
  </item>
  <item>
   <title>defining a variable for each element </title>
   <link>http://static2.comsol.dk/community/forums/general/thread/28605/</link>
   <description>hi , &lt;br /&gt;&#13;
     i would like to know how can i define a variable (which i am using in calculating other properties of the material) such that it is solved for each element separately ?&lt;br /&gt;&#13;
     Right now what COMSOL does is solve for a time step and use that value to calculate material properties for all elements together , whereas i would like COMSOL to solve that variable for all elements differently (much like temperature in heat transfer) in each time step.&lt;br /&gt;&#13;
     &lt;br /&gt;&#13;
-  Hirendra</description>
   <pubDate>Tue, 15 May 2012 08:47:01 +0000</pubDate>
   <guid isPermaLink="false">.1337071621.28605</guid>
  </item>
  <item>
   <title>Temperature Sensor Model</title>
   <link>http://static2.comsol.dk/community/forums/general/thread/28467/</link>
   <description>Hello,&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
I am designing a temperature sensor using an interdigitated capacitor.  Using the electrostatics modules, I was successfully able to design the sensor with the starting capacitance value I wanted.  However, regarding changing capacitance with temperature is where I run into problems.  Thus far, I have been using the joule heating and thermal expansion module.  Can anyone tell me if this is the correct module for my goal?  &lt;br /&gt;&#13;
&lt;br /&gt;&#13;
I have had some measure of success using it.  However I find that when I remove the surrounding rectangle with air associated with it in the positive z direction and silicon in the negative z direction, the study completes, but with that rectangle, I get an error of singular matrix/solution does not converge and cannot complete the study.  Why is this? &lt;br /&gt;&#13;
&lt;br /&gt;&#13;
Also, without the surrounding rectangle, and without the proper dielectric in place, it seems that I cannot calculate capacitance correctly, and the values are all very small, approaching zero.   &lt;br /&gt;&#13;
&lt;br /&gt;&#13;
I have attached my model.&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
I would greatly appreciate help understanding this, thank you.</description>
   <pubDate>Tue, 08 May 2012 21:11:51 +0000</pubDate>
   <guid isPermaLink="false">.1336511511.28467</guid>
  </item>
  <item>
   <title>Circuit terminal question</title>
   <link>http://static2.comsol.dk/community/forums/general/thread/27666/</link>
   <description>Hi, All,&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
I'm trying to build a model four terminal nodes. Two of them are COMSOL voltage (terminal 1) and ground (terminal 0), another two is connected to 'external' resistor which is not defined as part of the multiphysics model. &lt;br /&gt;&#13;
&lt;br /&gt;&#13;
Should I name the two terminal as type 'circuit', and name them 2,3? Then choose the nodes in electric circuit (cir) to be 2 and 3? What else should I do? I'm have an error says The DAE is structurally inconsistent .&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
Thanks!</description>
   <pubDate>Mon, 09 Apr 2012 01:37:00 +0000</pubDate>
   <guid isPermaLink="false">.1333935420.27666</guid>
  </item>
  <item>
   <title>Need some help in meshing, inverted mesh elements warnings</title>
   <link>http://static2.comsol.dk/community/forums/general/thread/27295/</link>
   <description>I have an FSI problem, by moving a membrane I would like to see the fluid move. &lt;br /&gt;&#13;
&lt;br /&gt;&#13;
When I setup the simulation, I've played around with the settings, looked at how other models were setup, but I can't get rid of the 'inverted mesh elements' warnings.&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
Does anyone have any advice on this model?</description>
   <pubDate>Tue, 27 Mar 2012 07:23:37 +0000</pubDate>
   <guid isPermaLink="false">.1332833017.27295</guid>
  </item>
  <item>
   <title>parametric sweep (old issue, new problem)</title>
   <link>http://static2.comsol.dk/community/forums/general/thread/26952/</link>
   <description>Hi all,&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
I know this issue has been mentioned a few times on the forum, but each one refers to a slightly different aspect and none of them explains the problems I am facing.&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
The idea is simple - I have got a 2D axysymmetric model and want to run it with a parameter subject to parametric sweep. Problem is it seems that the only thing COMSOL lets me do is see the results for different values of the altered parameter only on a single 3D plot and NONE of the 1D plots I am interested in (only for 3D I get the drop down menu).&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
So can someone please tell me how to see 1D point graphs for different values of parameter which was altered via parametric sweep and secondly - how to plot them on a single graph? I know I can do it exporting the data, but this will be quite time-consuming and I would expect such functionality to be included in the COMSOL 4.2a.&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
&lt;br /&gt;&#13;
Thank you in advance,&lt;br /&gt;&#13;
Mike</description>
   <pubDate>Wed, 14 Mar 2012 03:20:59 +0000</pubDate>
   <guid isPermaLink="false">.1331695259.26952</guid>
  </item>
  <item>
   <title>Stress gradient</title>
   <link>http://static2.comsol.dk/community/forums/general/thread/2537/</link>
   <description>Hello,&lt;br /&gt;&#13;
I am trying to simulate the deflection of a cantilever due to stress gradient, how can I define an uniform stress gradient through the cantilever thickness??&lt;br /&gt;&#13;
Thanks for your help!&lt;br /&gt;&#13;
Andres</description>
   <pubDate>Mon, 11 Jan 2010 08:25:59 +0000</pubDate>
   <guid isPermaLink="false">.1263198359.2537</guid>
  </item>
  <item>
   <title>Batch mode of parametric sweep analysis</title>
   <link>http://static2.comsol.dk/community/forums/general/thread/2357/</link>
   <description>Is there a way to use the batch mode to simulate a parametric sweep analysis without using MATLAB connection? If there is, how?</description>
   <pubDate>Tue, 29 Dec 2009 14:21:43 +0000</pubDate>
   <guid isPermaLink="false">.1262096503.2357</guid>
  </item>
 </channel>
</rss>

