|
[Sponsors] |
March 14, 2017, 15:54 |
Meshing help for very large domain
|
#1 |
New Member
Join Date: Mar 2017
Posts: 2
Rep Power: 0 |
Hi guys, im relatively new to ANSYS since it is required for my final year project in engineering and its been some months on it. Getting on to the main issue at hand. I have a very large flow domain , it extends vertically up to 1000m and horizontally it isnt much around less than 1m. My domain is the flow domain for an "O IL D RILL S TRING" (http://www.drillingformulas.com/wp-c...g-component.jp <---- this picture can be looked at as an example) . So now i have been trying to generate a mesh for more than a month (whenever i get the time to do so) however i find it very hard. I extracted the flow domains from within each solid body then i selected all flow domains and formed a new part so that it becomes one single domain. I tried every method to mesh i could find however some or problem comes up. as a summary some methods i tried are mentioned below.
1.) I tried to patch comform all parts and meshed each with suitable body sizing, got a suitable mesh for bit flow but for pipes it was very coarse as the geometry is very long as mentioned before, when taken to fluent it kept loading forever. 2.) I tried hex/tri for pipes and tri patch conforming for bit , but got error message as "mesh failed in translation...etc etc" 3.) i tried learning ICEM CFD but that was SUPER HARD and my time got drifted away. 4.) i tried method 1 again and did a somewhat ok mesh and got 22million mesh size but it was so huge fluent wasnt able to open. I did many other things but as of now i cant recollect them into my thoughts. As of now i believe my main issue is the geometry is so big there will be issues always. is there any way i can solve this so maybe i can have same shape but maybe scale geometry and have lesser mesh. or if anything else please help me as i have 2 weeks left for presentation X_X. I will attach my geometry and flow domain if you want to look into it. Best Regards |
|
March 15, 2017, 10:51 |
|
#2 |
Senior Member
Sebastian Engel
Join Date: Jun 2011
Location: Germany
Posts: 566
Rep Power: 20 |
Hi iluv,
an easy way would be just meshing the head section and sweep the interface mesh along the 1000m. You already know how to do the tet mesh. Then search the help how you can create prismatic layers over a surface mesh or with other words how to extrude a mesh (i can't help you with Ansys meshing, i can only guide in ICEM). regarding your 2 weeks deadline... get someone experienced (locally) to help you... you haven't even started the simulation part. With regards, Sebastian |
|
March 23, 2017, 01:13 |
|
#3 |
New Member
Join Date: Mar 2017
Posts: 2
Rep Power: 0 |
Hi BlueBase,
thanks for the awesome advice. Im looking into it , however i was able to generate quite useable mesh through ansys meshing itself. i achieved hex on the long cylinders through multizone and some sizing options and did tet patch conforming on non sweepable bodies, all this by slicing through sections in the bodies to get as many as sweepable bodies as possible. now what i want to do in fluent is simulate two different flows. to explain exactly what i mean . this is how it goes, the DRILL FLUID flows in the inner cylinder exits the drill bit nozzles, joins up with the DRILL MUD and these two different fluids flow together in the outer cylinder till the exit. to do this as i mentioned above, since i sliced the flow domains into many sections thus resulting in many domains. from here i selected those domains that are for fluid and FORMED a new part, similarly selected those domains that form for mud and FORMED a new part, now i have TWO major domains. when i went to fluent there werent two domains anymore though, instead there were all domains listed. so i didnt stop and in materials i created the fluid and mud materials and then in cell zones in edited the materials for each domain and now the simulation is currently running . however im not sure how correct my method is to achieve a model with two different fluids flowing. can u tell me if im correct? |
|
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
decomposition of multi domain cases for large simulations | romant | OpenFOAM Pre-Processing | 6 | May 10, 2023 21:27 |
Domain Imbalance | HMR | CFX | 5 | October 10, 2016 05:57 |
injection problem | Mark New | FLUENT | 0 | August 4, 2013 01:30 |
simpleFoam convergence on large domain | kwiik | OpenFOAM | 43 | May 31, 2013 10:10 |
Error finding variable "THERMX" | sunilpatil | CFX | 8 | April 26, 2013 07:00 |