Discussion:
Pure thermal analysis with Abaqus
edwinnyirenda
2008-11-18 06:39:17 UTC
Permalink
Dear ABAQUS users,

I am trying to understand the usage of ABAQUS 6.4 by solving for
fluxes in a 3-D cuboid of size 15m-by-10m-by-10m under purely thermal
conditions. Temperatures are fixed at 8 on one side of the 15m-by-
10m face, and at 4 on the opposite side, the other side of
dimensions 15m-by-10m. All the other boundary faces where the
temperature is not given should have a boundary flux state of 0. In
this way I am trying to simulate the situation of one dimensional
flow from one side of the 15m-by-10m face to the other side, and
given that the material conductivity is unity, I expect the solution
flux to be (8-4)/10=0.4. After this basic problem is solved I hope
to move on to more complex boundary conditions.

Two problems arise.

1. I am unable to input the boundary fluxes of 0.0. Abaqus
outputs a message that "A uniformly distributed load must have a non-
zero magnitude".

2. To try and achieve a run, I changed the zero flux to 0.0001.
The program starts to run, but aborts with the message in the file
Job-1.dat

***ERROR: A HEAT TRANSFER ANALYSIS IS NOT MEANINGFUL SINCE THERE ARE
NO TEMPERATURE DEGREES OF FREEDOM IN THE MODEL

How can I input the zero fluxes on boundaries, and how do I define
temperature degrees of freedom, seeing that I cannot do it under Load-
BC?

Any assistance to get me started will be highly appreciated.
BenZ
2008-11-18 12:20:10 UTC
Permalink
Just make... nothing ! Like a zero load in mechanics...!

BenZ
Post by edwinnyirenda
Dear ABAQUS users,
I am trying to understand the usage of ABAQUS 6.4 by solving for
fluxes in a 3-D cuboid of size 15m-by-10m-by-10m under purely
thermal
Post by edwinnyirenda
conditions. Temperatures are fixed at 8 on one side of the 15m-by-
10m face, and at 4 on the opposite side, the other side of
dimensions 15m-by-10m. All the other boundary faces where the
temperature is not given should have a boundary flux state of 0.
In
Post by edwinnyirenda
this way I am trying to simulate the situation of one dimensional
flow from one side of the 15m-by-10m face to the other side, and
given that the material conductivity is unity, I expect the
solution
Post by edwinnyirenda
flux to be (8-4)/10=0.4. After this basic problem is solved I
hope
Post by edwinnyirenda
to move on to more complex boundary conditions.
Two problems arise.
1. I am unable to input the boundary fluxes of 0.0. Abaqus
outputs a message that "A uniformly distributed load must have a non-
zero magnitude".
2. To try and achieve a run, I changed the zero flux to
0.0001.
Post by edwinnyirenda
The program starts to run, but aborts with the message in the file
Job-1.dat
***ERROR: A HEAT TRANSFER ANALYSIS IS NOT MEANINGFUL SINCE THERE ARE
NO TEMPERATURE DEGREES OF FREEDOM IN THE MODEL
How can I input the zero fluxes on boundaries, and how do I define
temperature degrees of freedom, seeing that I cannot do it under Load-
BC?
Any assistance to get me started will be highly appreciated.
steve manifold
2008-11-18 13:33:11 UTC
Permalink
You are probably using the wrong element. You need to use an element that has temperature degrees of freedom, such as C3D8T or DC3D8, to list only two. Look in the Analysis User's Manual under continuum elements.

A surface with no flux defined on it is assumed to have zero flux, no need to define a zero flux for it

Steve Manifold
University of Texas


----- Original Message -----
From: edwinnyirenda
To: ***@yahoogroups.com
Sent: Tuesday, November 18, 2008 12:39 AM
Subject: [Abaqus] Pure thermal analysis with Abaqus


Dear ABAQUS users,

I am trying to understand the usage of ABAQUS 6.4 by solving for
fluxes in a 3-D cuboid of size 15m-by-10m-by-10m under purely thermal
conditions. Temperatures are fixed at 8 on one side of the 15m-by-
10m face, and at 4 on the opposite side, the other side of
dimensions 15m-by-10m. All the other boundary faces where the
temperature is not given should have a boundary flux state of 0. In
this way I am trying to simulate the situation of one dimensional
flow from one side of the 15m-by-10m face to the other side, and
given that the material conductivity is unity, I expect the solution
flux to be (8-4)/10=0.4. After this basic problem is solved I hope
to move on to more complex boundary conditions.

Two problems arise.

1. I am unable to input the boundary fluxes of 0.0. Abaqus
outputs a message that "A uniformly distributed load must have a non-
zero magnitude".

2. To try and achieve a run, I changed the zero flux to 0.0001.
The program starts to run, but aborts with the message in the file
Job-1.dat

***ERROR: A HEAT TRANSFER ANALYSIS IS NOT MEANINGFUL SINCE THERE ARE
NO TEMPERATURE DEGREES OF FREEDOM IN THE MODEL

How can I input the zero fluxes on boundaries, and how do I define
temperature degrees of freedom, seeing that I cannot do it under Load-
BC?

Any assistance to get me started will be highly appreciated.





[Non-text portions of this message have been removed]
edwinnyirenda
2008-11-18 17:08:50 UTC
Permalink
Thank you very much for the help. This was my first try at Abaqus,
and indeed I was using a wrong element. I have now achieved a run,
yielding teh result I expected. Thank you again.

Edwin
Post by steve manifold
You are probably using the wrong element. You need to use an
element that has temperature degrees of freedom, such as C3D8T or
DC3D8, to list only two. Look in the Analysis User's Manual under
continuum elements.
Post by steve manifold
A surface with no flux defined on it is assumed to have zero flux,
no need to define a zero flux for it
Post by steve manifold
Steve Manifold
University of Texas
----- Original Message -----
From: edwinnyirenda
Sent: Tuesday, November 18, 2008 12:39 AM
Subject: [Abaqus] Pure thermal analysis with Abaqus
Dear ABAQUS users,
I am trying to understand the usage of ABAQUS 6.4 by solving for
fluxes in a 3-D cuboid of size 15m-by-10m-by-10m under purely thermal
conditions. Temperatures are fixed at 8 on one side of the 15m-by-
10m face, and at 4 on the opposite side, the other side of
dimensions 15m-by-10m. All the other boundary faces where the
temperature is not given should have a boundary flux state of 0. In
this way I am trying to simulate the situation of one dimensional
flow from one side of the 15m-by-10m face to the other side, and
given that the material conductivity is unity, I expect the
solution
Post by steve manifold
flux to be (8-4)/10=0.4. After this basic problem is solved I hope
to move on to more complex boundary conditions.
Two problems arise.
1. I am unable to input the boundary fluxes of 0.0. Abaqus
outputs a message that "A uniformly distributed load must have a non-
zero magnitude".
2. To try and achieve a run, I changed the zero flux to 0.0001.
The program starts to run, but aborts with the message in the file
Job-1.dat
***ERROR: A HEAT TRANSFER ANALYSIS IS NOT MEANINGFUL SINCE THERE ARE
NO TEMPERATURE DEGREES OF FREEDOM IN THE MODEL
How can I input the zero fluxes on boundaries, and how do I
define
Post by steve manifold
temperature degrees of freedom, seeing that I cannot do it under Load-
BC?
Any assistance to get me started will be highly appreciated.
[Non-text portions of this message have been removed]
Loading...