Tek-Tips is the largest IT community on the Internet today!
Members share and learn making Tek-Tips Forums the best source of peer-reviewed technical information on the Internet!
-
Congratulations cowski on being selected by the Eng-Tips community for having the most helpful posts in the forums last week. Way to Go!
Ansys Workbench 9---Thermal Expansion?
- Thread starter avic209
- Start date
As far as I know, Workbench 9 doesn not allow you to place any "gradiented" or spatially-variable load, so you have no chance to do it within the WB interface. You have to enter the Classical environment after writing the "Ansys Input File" (.imp) and finish your work there.
You just need to specify your coefficient of thermal expansion (alpha). From this you need to find the thermal strains 'Eth':
Eth = alpha(T-Tref)
alpha = linear coefficient of thermal expansion (a material property)
Tref = some reference temperature (e.g. ambient)
T = the unknown temperature to be found
Hence to find the expansion you can do a thermal stress analysis in workbench, and from this you can find the thermal strains and the expansion of the model.
Use the results from your thermal analysis to start a thermal-stress analysis (search for "filters" in the help file). Either that or do the thermal in WB and then transfer to ANSYS and issue the command LDREAD for the thermal results.
------------
See faq569-1083 for details on how to make best use of Eng-Tips.com
Eth = alpha(T-Tref)
alpha = linear coefficient of thermal expansion (a material property)
Tref = some reference temperature (e.g. ambient)
T = the unknown temperature to be found
Hence to find the expansion you can do a thermal stress analysis in workbench, and from this you can find the thermal strains and the expansion of the model.
Use the results from your thermal analysis to start a thermal-stress analysis (search for "filters" in the help file). Either that or do the thermal in WB and then transfer to ANSYS and issue the command LDREAD for the thermal results.
------------
See faq569-1083 for details on how to make best use of Eng-Tips.com
From above post from Drej:
<Eth=alpha(T-Tref)>
Yes, but my collegues who use Workbench told me that T must be constant over the application target, so if there is a variable T distribution over a surface, for example, either you slice the surface within the CAD and "discretize" the T gradient, or you must switch to ANSYS classical.
Is there something we are missing?
<Eth=alpha(T-Tref)>
Yes, but my collegues who use Workbench told me that T must be constant over the application target, so if there is a variable T distribution over a surface, for example, either you slice the surface within the CAD and "discretize" the T gradient, or you must switch to ANSYS classical.
Is there something we are missing?
I'm sorry but I don't see your point. What's an "application target"? How do you "discretise a T gradient"? Please try to use more conventional terms.
The problem appears to me to be very simple, but perhaps the OP could clear it up, or perhaps I am also missing something. You need a differential T over a surface (say) in order to get the thermal expansion (if the T was constant, there would be no thermal strain/expansion). You can apply this T either by transient or static means to obtain the thermal distribution. Once you have this, you can obtain the thermal stress, which implicitly contains the thermal expansion through the alpha constant.
------------
See faq569-1083 for details on how to make best use of Eng-Tips.com
The problem appears to me to be very simple, but perhaps the OP could clear it up, or perhaps I am also missing something. You need a differential T over a surface (say) in order to get the thermal expansion (if the T was constant, there would be no thermal strain/expansion). You can apply this T either by transient or static means to obtain the thermal distribution. Once you have this, you can obtain the thermal stress, which implicitly contains the thermal expansion through the alpha constant.
------------
See faq569-1083 for details on how to make best use of Eng-Tips.com
application target = one surface, or one body, or...
The fact is: OK, to have thermal expansion there must be a DELTA-Temperature. But this DELTA can be constant over the, say, surface, or not (=be dependent on the spatial coordinate).
I mean:
If the T is constant over the surface, it will be
T=f(x;y)=const
This can obviously be handled by WB. In this case, if T.ne.Tref, there is thermal expansion but not necessarily thermal stress (it depends on the restraints).
If the T is a gradient in the x direction, it will be:
T=f(x;y)=a*x+b
Let's generalize to an arbitrary T distrib (polynomial in this case):
T=f(x;y)=a*x^2+b*y^2+c*xy+d*x+e*y+f
For these cases, my collegues said that WB can not be used (that's why they shift to Clasical). Are we missing something?
You may note that this has nothing to do with time-dependency.
Discretization:
of course, when you have T=f(x;y), you can slice the surface in the x and y directions with sufficiently small pitch in order for you to consider T=const over the slice. The error with respect to the continuous distrib function can be very easily calculated.
The fact is: OK, to have thermal expansion there must be a DELTA-Temperature. But this DELTA can be constant over the, say, surface, or not (=be dependent on the spatial coordinate).
I mean:
If the T is constant over the surface, it will be
T=f(x;y)=const
This can obviously be handled by WB. In this case, if T.ne.Tref, there is thermal expansion but not necessarily thermal stress (it depends on the restraints).
If the T is a gradient in the x direction, it will be:
T=f(x;y)=a*x+b
Let's generalize to an arbitrary T distrib (polynomial in this case):
T=f(x;y)=a*x^2+b*y^2+c*xy+d*x+e*y+f
For these cases, my collegues said that WB can not be used (that's why they shift to Clasical). Are we missing something?
You may note that this has nothing to do with time-dependency.
Discretization:
of course, when you have T=f(x;y), you can slice the surface in the x and y directions with sufficiently small pitch in order for you to consider T=const over the slice. The error with respect to the continuous distrib function can be very easily calculated.
Sie auch muessen Seine Antwort zu erklaeren!!!
(You too have to make your answer clearer)
Refer to what? Must I understand that WB does handle gradiented BC?
Or, simply, did I get confused by reading "...for a thermal analysis, I got..." instead of "...from a thermal analysis, I got..."?
(You too have to make your answer clearer)
Refer to what? Must I understand that WB does handle gradiented BC?
Or, simply, did I get confused by reading "...for a thermal analysis, I got..." instead of "...from a thermal analysis, I got..."?
Ah, OK. This solves the OP question.
But, by the way, are gradiented BC possible in WB9, or not? I never used it until now except for some simple tasks, always used Classical, but as my company is pushing on CAD/FEM interoperation, WB would be great in conjunction with our CAD...
But, by the way, are gradiented BC possible in WB9, or not? I never used it until now except for some simple tasks, always used Classical, but as my company is pushing on CAD/FEM interoperation, WB would be great in conjunction with our CAD...
- Thread starter
- #11
Thanks you all, let make my point clear, first I do thermal analysis to find temperature distribution on model. secondly, in other, to find thermal expansion on model, I need to do stress analysis. unfortunately, in stress analysis enviroment, I can not apply "temperature distribution" as a load. Please show me how to apply thermal result as a load for stress. analysis? FYI: I use ansys workbench 9.0
Best regards,
a.v
Best regards,
a.v
- Status
Similar threads
- Locked
- Question
- Locked
- Question