Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.
2d Lam Flow Drag Calcs Time Dependent vs. Stationary
Posted Jan 17, 2011, 5:21 p.m. EST 5 Replies
Please login with a confirmed email address before reporting spam
I am a little confused in that I have integrated in the data set, and it seems from the desc. that reacf may integrate as well.
The "Drag on a Cylinder" model gives the drag coefficient as -reacf(u)*2/(spf.rho*U_mean^2*0.2)
From looking at a drag equation it seems that reacf(u) equals [ Drag_Force/Area]
In my model which is Stationary not Time Dependent like "Drag on a Cylinder" I get a 0 value for both reacf(u) and reacf(v) - Can I not calculate drag from a stationary model? Do I have some other problem?
Also it seems much easier to get a time dependent model to solve, I have been having lots of problems getting the stationary model to solve - should I switch to time dependency? I frequently get the error that it cannot be solver/ cant evaluate the expression.
Thanks for any assistance. I have attached my model to this post.
Alex
The "Drag on a Cylinder" model gives the drag coefficient as -reacf(u)*2/(spf.rho*U_mean^2*0.2)
From looking at a drag equation it seems that reacf(u) equals [ Drag_Force/Area]
In my model which is Stationary not Time Dependent like "Drag on a Cylinder" I get a 0 value for both reacf(u) and reacf(v) - Can I not calculate drag from a stationary model? Do I have some other problem?
Also it seems much easier to get a time dependent model to solve, I have been having lots of problems getting the stationary model to solve - should I switch to time dependency? I frequently get the error that it cannot be solver/ cant evaluate the expression.
Thanks for any assistance. I have attached my model to this post.
Alex
Attachments:
5 Replies Last Post Jan 28, 2011, 1:29 a.m. EST
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
I'm not sure I understand everything here, but I have a few comments on your model:
0) Have you considered upgrading your 4.1, there is a patch on the main COMSOL site
1) the solution looks rather turbulent, as do your "cell Reynolds" number, (for me at least) and you are using laminar flo w, but:
2) I find the mesh rather large, have you tried finer or the default meshing for fluidics ? Why not "calibrate" it for Fluidics too ?
3) your current inlet velocity is fixed at 5 [m/s]? but you also have a variable V0
4) your 1D plot is based on an integration, but you have not defined on which boundaries
5) you do not need the mesh of the inside of the airfin, so it's simpler to remove it with a Difference in the geometrical section
6) sometimes it helps to ramp up the velocity, or to define initial conditions closer to the reality i.e. some initial flow
7) is it an imported (from 3.5a) or a rebuild from scratch model. My experience is that on should rebuild the models fully in the new versions, at best import the geoemtry
What is the "*0.2" representing in your formula, are you sure there is no units missing ? furthermore, to check the units of a varaible nothing easier than to plot it, and check what COMSOL proposes
You have also the predefined Drag and Force variables
--
Good luck
Ivar
I'm not sure I understand everything here, but I have a few comments on your model:
0) Have you considered upgrading your 4.1, there is a patch on the main COMSOL site
1) the solution looks rather turbulent, as do your "cell Reynolds" number, (for me at least) and you are using laminar flo w, but:
2) I find the mesh rather large, have you tried finer or the default meshing for fluidics ? Why not "calibrate" it for Fluidics too ?
3) your current inlet velocity is fixed at 5 [m/s]? but you also have a variable V0
4) your 1D plot is based on an integration, but you have not defined on which boundaries
5) you do not need the mesh of the inside of the airfin, so it's simpler to remove it with a Difference in the geometrical section
6) sometimes it helps to ramp up the velocity, or to define initial conditions closer to the reality i.e. some initial flow
7) is it an imported (from 3.5a) or a rebuild from scratch model. My experience is that on should rebuild the models fully in the new versions, at best import the geoemtry
What is the "*0.2" representing in your formula, are you sure there is no units missing ? furthermore, to check the units of a varaible nothing easier than to plot it, and check what COMSOL proposes
You have also the predefined Drag and Force variables
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hello Ivar!
I have used all of your recommendations. I have built from the ground up a turbulent model. Before I was trying to adapt the tutorial models.
Mesh - Normal size calibrated for fluid dynamics. ( will make finer once things are working)
The .2 in the formula was from the drag on cylinder model, I also could not understand what it is since its not a standard drag force equation component.
I have a two week trial of all modules and want to try to get this done and export the solids to machine them before its up!
Output: This is where im stuck now. The turbulent model does not have a standard drag expression to select.
All the posts i have read about lift/drag involve using -reacf(u) I cannot get this to return a non 0 number.
I have used all of your recommendations. I have built from the ground up a turbulent model. Before I was trying to adapt the tutorial models.
Mesh - Normal size calibrated for fluid dynamics. ( will make finer once things are working)
The .2 in the formula was from the drag on cylinder model, I also could not understand what it is since its not a standard drag force equation component.
I have a two week trial of all modules and want to try to get this done and export the solids to machine them before its up!
Output: This is where im stuck now. The turbulent model does not have a standard drag expression to select.
All the posts i have read about lift/drag involve using -reacf(u) I cannot get this to return a non 0 number.
Attachments:
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
indeed reacf() is not working and there is no readily defined DRAG force I can find neither therein. And probably by deriving the fluid stress tensor "spf.K_stressx" around the edge is not precise enough as with turbulent mode you have more complex wall effects, than simple noslip, if I remember right
This is not the field where I have most experience, there are other CFD specialist out here that could give you better clues ;)
--
Good luck
Ivar
indeed reacf() is not working and there is no readily defined DRAG force I can find neither therein. And probably by deriving the fluid stress tensor "spf.K_stressx" around the edge is not precise enough as with turbulent mode you have more complex wall effects, than simple noslip, if I remember right
This is not the field where I have most experience, there are other CFD specialist out here that could give you better clues ;)
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Wanted to bump this thread up again....
I have a working model is anyone able to advise me on how to pull some information for lift or drag out of it so that I can compare a few shapes?
I have a working model is anyone able to advise me on how to pull some information for lift or drag out of it so that I can compare a few shapes?
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
that is good to hear, but it would be easier, if not delivering the model, at least summarise the physics used and the dependent (complete) variable list ofyour final model, this allows others to test or compare.
I always check the reaction forces (when solved for) with the integration of the stress tensors, too follow any discrepancy
--
Good luck
Ivar
that is good to hear, but it would be easier, if not delivering the model, at least summarise the physics used and the dependent (complete) variable list ofyour final model, this allows others to test or compare.
I always check the reaction forces (when solved for) with the integration of the stress tensors, too follow any discrepancy
--
Good luck
Ivar