ANSYS Tutorial | Grid Independence Test In ANSYS Fluent Using Parametric Analysis
HTML-код
- Опубликовано: 20 авг 2024
- In this tutorial, it has been shown, how easily and with less time you can do the grid independence test using the parametric analysis method. A simple cylindrical fluid domain has been taken in this Tutorial, for simplicity and easy understanding.
What will you learn from this Tutorial?
1) Creating input parameters in meshing.
2) Output parameters set up in Fluent.
3) Comparing the result of various mesh size starting from coarse to fine,with the solution output.
A CFD Solution can never be trusted unless we check whether the result depends on the grid or not. or else we have any experimental data. The output result of a coarser mesh and finer mesh can neither be the same. So up to what extent, we need to vary the mesh to get the accepted level of tolerance, can be found out from Grid Independence test.
How it is Done?
This is done by varying the mesh size from coarse to fine and checking the output result for each mesh. When varying the mesh does not affect the result much then we can stop and select that minimum mesh size for our final solution output.
![[CFD] How Fine should my CFD mesh be?](http://i.ytimg.com/vi/60fDz2cVdy8/mqdefault.jpg)
![[CFD] How Fine should my CFD mesh be?](/img/tr.png)
you sir deserve a medal for all the help you upload. Thank you
Thank you for your appreciation. Please share our channel videos with your friends.
great! i am student since i dont really understood wht hv been thought in the class and then this video came, now im really understand
A fantastic idea using parametric for analysing mesh sensitivity and good informational video, thanks
Thank You:)))
1:40 - edge number of division mesh, increase to 60 at 2:15, increase to 100 at 2:22
2:42 check nodes and elements as output parameters, check number of divisions at 2:45
3:13 name the boundary layers, 3:40 named all the layers, hit update the mesh when done
3:55 check if all parameters are chosen
5:32 fill up boundary conditions and decrease residual for better accuracy
5:51 initilisation, compute from inlet
6:07 run calculations, check if its converged
6:35 go to “Reports” check the output results, change field variable to velocity, pressure, anything
7:50 change report type to facet average then facet maximum, velocity
Create the following output parameter : Facet average {out-velocity, pressure-out,} , Facet Maximum { max velocity} save project.
8:50 change the number of devision on the parameter window
my mesh is not linking to fluent.. why??
Great Video, thanks for having people like you
Thank you for your appreciation.
Althought i would have prefer a voice over rather than pausing the video, despite all that it was beautifully explained!! Great video! Learned a lot.
Thanks
Thank you. Our latest tutorials are like that.
@@AnsysTutor ahh okay 👍🏾👍🏾
Thankyou boss
i didnt know to add the result as a parameter out.
thankyouu
You are welcome
very gud tutorial...really delivered and elaborated the concept of mesh independency test.....
Thank you 🙂🙂🙂
Very useful video. Saved me a hell lot of time.
I am simulating a transient solution and want my output parameters to be after a specific timestep. I Will be very thankful if you can help me out.
I think you need to setup,after how many iterations you want the solver to save the results.
@@AnsysTutor Actually i want the mass flow rate of a phase (sand in my case) to be saved automatically after every 1000 timesteps or so (30mins in terms of flow time)
Thank you so much!
You're welcome!
10.12 , fourth column right hand side : you have not refined your mesh , but I see that the software predicts results ( by inventing a such number of cells and nodes) ( I have seen in other videos that persons refine their mesh and redo simulation)
Can we so trust in this results ?
Thanks you
Thank you nice explanation
You are most welcome.
Thank You, it was beautifully explained
You are most welcome.
Thanks for the tutorial! Very helpful
Thank you :)
Great video!
Thank you.
Nice work
Thank you.
Thank you so much. This is so helpful.
You are most welcome.
Excellent Video!!
Thank you very much!
thank u so much teacher, best information, 👏👏👏👏👏👍👍👍👍👍👍😀😀😀😀
teacher is that possible to explain in a tutorial how we much contact mesh between two parts also including inflation in the mesh? if its possible please explain that in a quick tutorial, thank u so much for the hard work my best teacher 😊
Thank You, alright will try to make a tutorial on that.
superb..
Thanks a lot
Great video! Really helpful. How to do it for transient simulations where Delta_T is also involved?
In that case, you can compare any variables with respect to flow time in each of the grid case.
Amazing, SO HELPFUL
Thank you.
Thank you so much... It's very helpful
..thank you
You are most welcome.
I enjoyed the background music
Thanks Joel
Thank you very much
You are welcome.
Very useful
Thank you:)
A great tutorial
thank you🙂🙂🙂
great tutorial thank you
Most welcome🙂🙂🙂
Great video. Do you know how to bring out the convergence point to the parameter. In case of your video the iter 65. I want to record for each of the sizing.
As per my knowledge its not possible.
Thank you so much
Most welcome.🙂🙂🙂
The residual curves are fluctuating vigorously in every iteration of my cfd model. What is the problem ? My model is a vortex tube where high pressure inlet air is coming through 6 nozzles and at cold exit and hot exit gets cold and hot temperature respectively.
Are you getting any reversed flow warning during iterations.
@@AnsysTutor No
thanks, very helpful
most welcome:)
Hello, can I do Grid Independence Test In ANSYS Fluent Using Parametric Analysis with an Pointwise mesh?
Also please what do you advise, I’m trying to simulate turbulent flow in a pipe with hydraulic diameter of 11.7 mm and a wall thickness of 1.5mm, do i need to put the pipe? Or i can just simulate with the fluid region
Mesh elements or mesh nodes column is not displaying in the table of design points? What to do?
You need to select the box near the side of elements and node in mesh statistics,as shown in the tutorial.
Very helpful video. Very few video of mesh independent test is in youtube. Do you have patreon or something to donate? I am really happy with what you shared.
Thank you for your appreciations. 🙏🙏🙏
You can donate via paypal or buymeacoffe, all the links given in the about tab of our channel page.
Great video sir. Really helpful. I’m new in ansys. Can you help me sir about my case? my calculation for iteration is not converge. My model is same as yours but different in size. I also used k-epsilon for my method. Do you have any idea to solve convergence problem? Thanks
Not all iterations need to be converged, if the residuals vs iterations plot is not fluctuate much then there is no requirement of any solution converge. You can stop after certain iterations.
Do we need to perform a time independence test as well, for transient cases? Or is it not required if we already have convergence for each off the mesh sizes!
Time independence test not required as in each case you need to put the same simulation time.
Ansys-Tutor what if it's a transient case? Shall we not plot different variables vs time plot, for different mesh sizes? ( Because you've done this simulation for a steady state case only )
Great video, thank you. I noticed you didn't specify under "calculation activities" after what time steps you want to be saving your results, I mean, do I need to be saving my result to be able to run a parametric study? I would really appreciate your feedback as soon as possible, thank you.
If you want to use the various parametes files in future then you can use the save option. Otherwise if you want to get only output data then you can skip that part.
I don't have have number of nodes or elements in my parametric table (Columns C-E). Any help would be appreciated
Have you tick ✅ the nodes and element box in mesh statistics.?
Thanks for the tutorial. Why you didn’t calculate the deviation percentage in results for every mesh size ? You compared the results directly and selected the appropriate mesh.
Yes you can compare that. This is only a tutorial demonstration ,due to some time constraints, we have skipped that.
@@AnsysTutor how to caclculate devaition percentage
i have a condition , i gave input BC to be 2.5m/s and outlet pressure of 0 pa on a pipeflow consisting k-w model but it doesnt converge even giving 1000 iterations, i already checked the mesh setting but all seems fine.
The calculation not always require to be converged, if the iteration vs residual graph is not fluctuating much & it is a horizontal line, then you can ensure that no need to iterate much. &It's OK if there is no convergence.
@@AnsysTutor thank you so much
I have a question, is GIT can be done in setup tab? suppose if I had only mesh file/case file, so it means mesh edit option won't be available so how to do GIT in that case?
In that case you need to run the setup manually each time with the modified mesh.
@@AnsysTutor Thank You
Hello, thanks for the video, it is helpful. I simulate a transient multiphase flow - will it be justified to make a conclusion about sufficiency of mesh refinement according to calculations of stationary flow based on the same conditions?
Yes, grid Independence test can be done in steady state for a transient flow problem.
Thanks for the video. In my case, the data is exported to Excel all in the same cell separated by commas. How could I make them all separated in cells as well as you?
Try to save the file in .CSV format and while closing the file never save it.
sir, I am doing transient simulation. I can't find the message "the application is busy" what should I do?
thats good if you dont receive this message . dont worry.
Sir, so should I set the monitors before calculation? Then, after completing the calculation, should check the parameters?
Great tutorial, but not sure why my Parameter set page is appearing blank although I have putted all p's correctly
You need to Reset the workspace window. You can join our telegram group or on facebook to get a personalized solution to the problem which you are experiencing.
@@AnsysTutor May I get the url for FB community?
facebook.com/ansystutor7
Telegram Group: t.me/ansystutor7
how to enter grid size of 24*36*120
Apply edge sizing.
In the parameter set, there is no parameter regarding the mesh. Do you how to add it?
you cant input the mesh parameters in parameter set, only you can get the information regarding the no of elements/nodes generated.
how can we do parametric for boundary condition in setup
We had a tutorial on that.
ruclips.net/video/gIWqNEAcmsw/видео.html
Hi! Why does the "parameter set" window sometimes appear and sometimes not? What does it depend on?
Have you set the parameter? It would only appear if you have set any input, output parameter or both.
@AnsysTutor
Have a look. You opened a "mesh" regime, created a mesh, closed this window, and then found that the "parameter set" appeared in Workbench.
I don't understand what specific action in the grid creation process led to this.)
(When I created grids before, this did not happen.)
@AnsysTutor could you help me? 👆
Its is the mesh statistics which has been added as parameter.
@AnsysTutor i found it.
2:40
thank you for this tetorial , how can we compare cases like flow or geometry in fluent
Thank You. You can do parametric analysis to that also, we had a tutorial on that please check.
@@AnsysTutor can you please provide the link for the same .
Whenever I'm clicking on update in the 'Solution' tab it starts to solve again. Why?
You need to save the solution in fluent setup after you complete the calculation.
Ansys-Tutor I did, I saved the solution with the future calculation option as shown in your video. But it still restarts don't know why!
As its a parametric setup it would be like that, dont worry about it, after you finish the calculation for all the parameters , it would be show that green tick mark. (All the solutions updated)
@@AnsysTutor Yeah, okay! Thanks anyways!
Great Video.
how can i do this in CFX?
When I save my project the option "Use settings changes for current and future calculation" is not available. Can you pls help me?
What is the option are you getting instead of that.
@@AnsysTutor Use settings for current calcilation only
Yea even I had the same issue what did u do to overcome this??
can i make the mesh nodes and elements for the input parameter? how can i make this possible? thank you for your answer.
Its not possible as per my knowledge, but you can only vary the element size .
@@AnsysTutor Alright, thankyou sir for your answer :)
For multiphase flow, is it okay to conduct mesh independence test with single phase?
No it's not feasible, as in both cases the governing equations are different.
"Use settings changes for current and future calculation" option is blocked for me. What do I do?
Hello, it's been 2 years. If you still remember, can you tell me how to solve this?
Can we do this with periodic boundary condition ?? Or should I apply normal B.C for this test ?
yes, you can apply it with periodic BC, the boundary conditions are on the Boundary surface, it does not get affected by the numbers of cells on the boundary.
a quick question sir, what if the result I obtained provides no rapid variation at all for either of the output parameters?
Its not necessary that in every mesh there would be rapid variations. If you want to obtain rapid variations you can try with a coarse mesh.
@@AnsysTutor Alright thanks for your reply sir
Excuse me sir, is the division number just a division number or is there a unit? Then is there a limit for the division number or what, sir? Thank You
Is is the simple division of sides it has no units. There is no such limits for division, but note that if you increase the division, then the meshing would generate smaller cells which would eventually increase the computation time.
Thank you sir, so what about the aerodynamics simulation that I am doing for my final assignment at university, sir? so in the simulation I use domain, body of influence and body sizing. Does the application of number of division apply if we add body of influence, body sizing, sir? please guide
Or, can I ask for your email? for discussion
Hi, can i ask do we need to change the relevance center sizing from coarse to fine? Or just insert the sizing method and ignore the relevance center sizing without changing anything?
You dont need to change the relevance center , as it cant be parameterized.
@@AnsysTutor alright thank you so much 👍
You are very welcome.
hello again, just want to ask. for this mesh independence study, can i change the fluid into something else and used turbulent model instead of laminar? or i must used laminar and water to perform this test?
i mean, used my on fluid properties instead of using water as shown in your vid
You can use any material as per your study, its not mandatory to use the same material as shown in the tutorial.
@@AnsysTutor thanks for your help sir
sir i have few queries related to ansys so how can i contact you actually i am working on my project so i need some help as i am new to ansys....
You can ask in our Facebook page chat, email etc. All the links given in the about tab of this youtube page.
Hi. Thanks for the very informative video. However, if let say I run the simulation and may I know how to plot a graph of pressure across the cylinder against length of cyliner, for few design points using parameters feature?
As I tried in post processing only know how to create for one design point (current) only. Is that possible to insert all the results (pressure along cylinder vs cylinder length) on the same chart?
Thank you in advance.
yes, its possible . you need to insert it in the new series .
yes, its possible . you need to insert it in the new series .
Sir.. Is this applicable for any fluid flow through any structure and physics..
Yes. This application (Ansys fluent) is based on CFD.
Can't thank you enough....._/\_
Thanks
Great video but how do we select the correct sizing?
If you have any heat transfer calculation over the pipe you need to create the pipe domain or you can define the wall thickness in boundary conditions.
@@AnsysTutor thats the heat flux wall thickness is part of the surface right?