Geometry - Latest Discussions - COMSOL Forums

Problem in assigning boundary and initial conditions

Sat, 11 Feb 2012 17:14:18 +0000

Hello,

I am new to comsol & i want to simulate a very simple system; water in a pipe with a spherical ball kept as a disturbance.

I am using comsol multiphysics 4.0a. I opened comsol, clicked on 2D> fluid flow> single phase> laminar flow, then I chose stationary study and finished the initial settings by clicking on the checkered flag.

In geometry, i created a rectangle and a circle inside it. Under laminar flow, i clicked on wall properties to assign boundary conditions.

Here is the problem. By default, comsol has selected all lines & curves as boundary and the 'Selection:' has been disabled, so I cannot change it manually. I face same problem while assigning initial value conditions. I cannot give 'Initial Values' to my model. The problem remains same even when I switch between 'Form an assembly' to 'Form a union' under Geometry.

The problem remains same even when I try to simulate structural problems or any other physics.

I am sure this problem has a very simple answer but as a biginner I am unaware of the features and it'll take me life long to explore each feature all by myself. I appreciate anyone who would like to help me.

Thank you in advance.

Cylic Voltammetry

Fri, 10 Feb 2012 21:30:27 +0000

Hi

I want to model a 3 electrode electrochemical cell that I am using in cyclic voltammetry. I wanted to know if anyone had any suggestions on how to get start doing this using COMSOL as I am new.

I would like to be able to plot current vs potential and current vs time.

Thanks
J

Save comsol file as M.file, but without any parameter of geometry in script

Fri, 10 Feb 2012 20:46:47 +0000

Hi everyone,

I builded my model based on comsol 3.5a version, and then save it as to M.file. But when I save comsol file as to the .m file, the geometry file was produced as flbinary file .mphm (the example I list below). I do not want to get this script (flbilary), since the geometry parameter can not be changed. is there anybody know how to solve this problem?

Thanks.

% The name of this file is given by the variable 'flbinaryfile'.

flclear fem

% COMSOL version
clear vrsn
vrsn.name = 'COMSOL 3.5';
vrsn.ext = 'a';
vrsn.major = 0;
vrsn.build = 603;
vrsn.rcs = '$Name: $';
vrsn.date = '$Date: 2008/12/03 17:02:19 $';
fem.version = vrsn;

flbinaryfile='1.mphm';

% Geometry
clear draw
g247=flbinary('g247','draw',flbinaryfile);
draw.s.objs = {g247};
draw.s.name = {'CO1'};
draw.s.tags = {'g247'};
fem.draw = draw;
fem.geom = geomcsg(fem);
fem.mesh = flbinary('m1','mesh',flbinaryfile);

% (Default values are not included)

I want the form like this so that I can change the geometry parameter.

g1=ellip2(0.2,0.2,'base','center','pos',[0.4,-0.4]);
g2=ellip2(r,r,'base','center','pos',{'0','0'},'rot','0');
garr=geomarrayr(g2,pitch,pitch/sqrt(3),5,8);
[g40,g41,g42,g43,g44,g45,g46,g47,g48,g49,g50,g51,g52,g53,g54,g55,g56,g57,g58,g59,g60,g61,g62,g63,g64,g65,g66,g67,g68,g69,g70,g71,g72,g73,g74,g75,g76,g77,g78,g79]=deal(garr{:});
g80=ellip2(2.0E-4,2.0E-4,'base','center','pos',[-0.0010,6.0E-4]);
g81=ellip2(r,r,'base','center','pos',{pitch/2,pitch/(2*sqrt(3))},'rot','0');

Defining Geometry Domains

Wed, 08 Feb 2012 21:24:12 +0000

Hi,
I've made a closed loop with 4 different Bezier Polygon line segments. I'm able to select them as a group of boundaries but I can't find a way to define them as a domain to add later physics.
Thanks
LoongYi

import a bmp image into COMSOL geometry

Tue, 07 Feb 2012 14:10:48 +0000

Hey,

as stated in this webpage: http://www.comsol.com/products/formats/ Comsol can both read and write few types of images. However, when I tried to import a bmp image as my structure for flow simulation, it was not possible, because the options given include only mphtxt, dxf, etc. Does any one know how I can get a bmp or jpeg image imported? Thanks!

Chunyan

How to draw a block with rounded corner in geometry?

Mon, 06 Feb 2012 23:40:37 +0000

Hi All,

I am trying to draw a round corner cuboid in Comsol 4.2, it will be like a geometry between a cylinder and a block.

I saw somewhere saied "round the corner with fillet radius", but do not know how to actually make it. I tried making several cylinders and a block, and make boolean calculation, but it does not come out a satisfied result. Can anybody give me a way out, let me know the specific procedure of doing this?

Thanks!!

Droplet breakup in a T-Junction in a smaller channel

Sat, 04 Feb 2012 23:09:38 +0000

Hi all,

I have a specific set of parameters for wich I am trying to simulate the breakup of droplets in a T-junction. Entering these parameters for the geometry from the example MEMS module[1] gives fine results (droplets form).
However, if I slash the channel size in half (50x50x20 mum) I get no droplets but a coflow regime, where the dispersed phase occupies the upper part of the channel.
Now I have a real life system that has exactly the same dimensions (50x50x20 mum) and I can generate droplets for the same set of parameters in that.
So, I am somewhat stumped on how to proceed from here. I tried doubling the meshsize and increased the timeframe of the simulation and even fiddled around with the parameters, but it always gives me the cowflow regime.

Any help appreciated.

[1] The geometry in the example module is a 100x100x50 mum channel.

Time Derivation (or model in general) suddenly becomes unstable.

Tue, 31 Jan 2012 21:35:40 +0000

Hello everyone,

I am modelling electrical machines with the rotating machinery interface wthin the ACDC module.

This is my thread about the same model:
http://www.comsol.com/community/forums/general/thread/25702/

My problem is that the model won't solve when the permeability is non linear. Interesting is that after 2ms the simulation becomes very unstable and won't get to the 10ms mark.

In the attached pictures I want to show how unstable the time derivation of the flux linkage suddenly becomes.
You have an overview there and zoomed pictures.

Is it somehow possible to damp that derivation a litte? I'm afraid that this is causing my simulation to become unstable. Could it be that Comsol by default changes something at ca. 2ms that affects the derivation?

Here is a small overview about how I usually define my variables:

Psiges1 = 4*(intop2(mod1.Az))*...
e1 = -d(Psi1ges,TIME)
i2 = (u13+e1-e3)/4/Rsg/(-3)-...

Everything, including the derivations, is defined in the Definitions/Variables table.

Does anybody have an idea, why my model is so unstable? : (

Thanks in advance

avoided 2783 inverted elements by using linear geometry shape order

Tue, 31 Jan 2012 17:43:02 +0000

Hello all;
when i start running my i get this massage what it should mean and how i can avoid this error?
Best regard

Point3 boundary settings!

Tue, 31 Jan 2012 08:44:56 +0000

I am using Comsol 3.5a with matlab, I am creating in graphic interface a Point3 and when I want to set the boundary of that point is imposible to put any. The point is created in a 3D geometry and the 3D geometry is a single entity... for the points of this geometry there is no problem to set values but the extra point created does not give this option. Is it the license? I really appreciate any answer...

conversion from 2D to 3D

Mon, 30 Jan 2012 11:41:35 +0000

hello,
I am building a 'comb drive' model, in the 'electrostatics' division of the 'AC/DC module' in COMSOL 4.2. I need to 'extrude' a 2D model into a 3D one. How do i do that?

Thank you

Time dependent and non linear model takes (too?) long to solve...

Sun, 29 Jan 2012 21:52:09 +0000

Hello everyone,

I am simulating electrical machines within the rotating machinery interface in the ACDC module.

The model I calculate right now solves. But the way it solves is not satisfying in my oppinion. As soon as I choose the iron to be nonlinear, the solving takes days to calculate just 50ms with a time step of 100us.

I'm not sure if I'm allowed to upload the model in this forum, since the Geometry is not meant to be public.

The Geometry is a fourth of the motor and periodic boundary conditions are used for this. The current, stamped into the slots, is calculated by time dependent equations which involve the derivation of the flux field linkage. So comsol has to solve the equation for the current and vector potential simultaneously.

I already tried to improve the model this way:
http://www.comsol.com/support/knowledgebase/1127/
But it didn't really increase the speed. Actually with it's help the model could first be solved nonlinear...

Does anybody have general advices to get nonlinear time dependent models solved faster?

Thanks in advance.

Splitting up an imported geometry into different domains

Fri, 27 Jan 2012 13:08:06 +0000

Is it possible to import a geometry from a CAD file into Comsol in such a way that it contains more than 1 domain?

Till now I made all my geometries in comsol itself. But now I need to make a more complex geometry that contains different materials herefor the geometry needs to be more than 1 domain.

Small Copper Wire Coil Producing Magnetic Field

Thu, 26 Jan 2012 23:34:48 +0000

Hey all,

I'm pretty new to COMSOL, but I tried to build a very simple model which has a small, copper wire loop. The hope is that when a current is run through the loop there should be a larger magnetic field in the CENTER of the loop. Unfortunately this sort of "lensing" effect is not observed. I have consulted the model builder files both the current carrying wires model and the helmholtz coil model without being able to figure out whats wrong.

When I define the model I simply draw the parallel wires that bring current, and then a small loop and since I'm using the magnetic fields as my physics I then define the domains that make up the coil as having a current density in x,y,z under external current density. When I solve, I am able to produce a thermal map of the flux but it doesn't seem to make sense, as it implies the highest flux is along the wires and when I change it to magnetic field, it shows some nonsensical surface map that I'm having trouble believing. I would like to be able to produce data of the magnetic field gradient, the magnetic field, the heat produced by the sample when it's running, and the force produced by the wire, but I still haven't been able to make any reasonable results with the model I have. I fear that I may have somehow defined the current density incorrectly, but I'm not sure how. I also tried to build the model in 3D with similar results. I have attached the file, any help is always greatly appreciated.

Regards

Export Geometry File

Thu, 26 Jan 2012 22:20:26 +0000

I've used Comsol to optimize the geometry of an electrode. I was hoping to export this exact geometry so I can print a mask and then fabricate the electrode. When I've tried to export it, I can only seem to export it as a Comsol file. Is there any way to export it as a .tiff file?

effective refractive index RF module

Thu, 26 Jan 2012 20:37:35 +0000

Hi, quick question that may be easy for some of you. Say I have a fiber with a core of refractive index 1.45 and then I have an array of airholes in a background of refractive index 1.45. So the airholes lower the refractive index of the cladding region. How do I find out what this "average" effective refractive index of the cladding is. I know that I can solve the fiber and get modes with a wide range of effective refractive indices, but I just want to know what the cladding is effectively being "seen as". For example, I could then take this average cladding refractive index value and in a new model, place it as a cladding layer around my my core (think step index fiber, instead of a porous cladding) and compute the number of modes and then check this with the previous porous cladding fiber. Does anyone know how to do this?

New (local) coordinate system

Wed, 25 Jan 2012 08:47:44 +0000

Hi,
Can COMSOL define a new coordinate system or a local coordinate system to define analytical functions? Basically I need to offset an x y z coordinates for my function. In the case where the local coordinate is the same as the global coordinate, the offset coordinates are easy to determine. But in my case, the local coordinates system are rotated at an angle such that these offset coordinates cannot be determined. It will be helfpful if COMSOL has an option that allows us to define the analytical function through a local coordinate ssytem

Thanks.

BW,
EH

No-slip on interior boundary

Fri, 08 Apr 2011 12:07:58 +0000

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

Importing Boundary Conditions Using a .nas File

Mon, 17 Jan 2011 12:36:19 +0000

Is there any way to import boundary conditions into comsol using nastran file .nas mesh import.

Many thanks

Calculation of normal vectors on surface

Wed, 22 Dec 2010 14:10:11 +0000

Hi,

I need to calculate normal vectors on a 3D geometry surface to calculate boundary conditions from that (it's a rather complex matlab code). I took the postinterp help to get a starting point and tried to calculate the normal vectors in the center of each element. I used postinterp and tried to determine nx, ny and nz, but, whatever I do, I only get NaNs. In the postinterp help, a parameter s was used to do that, this also works fine in 3D, but I am not sure how I would transform certain points to that parameter... Here's my code:

% Create sphere surface geometry
clear fem
fem.geom=geomcoerce('face',sphere3(1,'pos',[0,0,0]));
geomplot(fem,'edgelabels','on'), axis equal
% Generate mesh
fem.mesh = meshinit(fem);
fem.shape = 2;
fem.xmesh = meshextend(fem);
% Get element triangulation
a=get(fem.mesh,'el');
tri=a{3}.elem;
p=fem.mesh.p;
% Calculate center point of each triangle
pn=(p(:,tri(1,:))+p(:,tri(2,:))+p(:,tri(3,:)))/3;
% Calculate normal vectors
nx=postinterp(fem,'nx',pn)
% this gives only NaNs, why?

Any ideas? Also the interpolation in the mesh points does not work,

nx=postinterp(fem,'nx',fem.mesh.p)

gives me the same NaNs...

Nils