|
[Sponsors] |
January 13, 2014, 19:48 |
dfbi,resistance calculation
|
#1 |
New Member
Join Date: Jan 2014
Posts: 3
Rep Power: 12 |
Dear all,
I would really appreciate your help in a question that may sound a bit 'stupid'... I model the half of the geometry in STAR CCM+ in accordance with the DFBI tutorials and everything is set up properly. The resistance that I will calculate will be for the half model or for the whole? I haven't really cleared in my mind the function of the symmetry plane. And another thing,how many cells do you usually use just to calculate the calm-water resistance? Thank you again. |
|
January 14, 2014, 12:07 |
|
#2 |
New Member
Join Date: Jan 2014
Posts: 3
Rep Power: 12 |
Well, seeing no reply I thought to elaborate a little.
When I started messing with Star CCM+ an experienced on CCM+ friend's friend (which I cannot reach anymore),advised me not to use more than 2M cells to simulate my cases (for the half of the geometry-symmetry plane,model size).Having domain dimensions equal to -2L<x<2L,0<y<L,and 1.5L from to top to bottom boundary, I generated grids from 1M to 1.8M cells.The calculated resistance was the half of the experimental one which I thought that made sense since only the half boat was modelled.By refining the grid I had accuracy from 5% to less than 1%.The value for trim was very close but I had large deviation to the sinkage value (about 40%).I set it aside for a while and in the meantime a colleague of mine made some simulations with a similar geometry with about 3.5M to 4.5M cells(similar domain size). The resulted resistance he ended up with was for the whole boat not just the half .Since he hasn't experimental results for sinkage and trim we cannot evaluate the other parameters but I believe that my mesh is to coarse and my results were just coincidental. On the other hand,by viewing various grids from Adapco's presentations over the internet, I understand that their grid density is much similar to what I have generated because with twice the cells the grid becomes much much finer. So I turn to you for your help.Please advise me. |
|
January 14, 2014, 22:18 |
|
#3 |
Senior Member
KHB
Join Date: Aug 2010
Location: Singapore
Posts: 118
Rep Power: 15 |
Dear Antares,
If you are simulating half of the domain, the Resistance value that you will obtain is only half of the 'real' resistance. This is because the pressure and shear force will be integrated and since only half of the surface that is present hence half of the total force that you will be getting. You need to remember that the mass, inertia that you need to include in the DFBI is also half of the physical thing. Regarding the sinkage value, make sure the sinkage is using the same reference, some tank test result based their sinkage on the static VCG, and maybe your simulation reference is from certain reference line, just make sure the simulation and test both are using the same reference. Mesh count I believe is largely dependent on your geometry complexity as well the speed that you are running. But I think 1-2M you should be able to get sensible result for simple shape (not like 100% error ) |
|
January 14, 2014, 22:32 |
|
#4 |
New Member
Join Date: Jan 2014
Posts: 3
Rep Power: 12 |
Dear Lava,
I was hoping that you would reply, thank you very,very much. |
|
January 15, 2014, 01:27 |
|
#5 |
Senior Member
KHB
Join Date: Aug 2010
Location: Singapore
Posts: 118
Rep Power: 15 |
You are welcome
|
|
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
Transfer output data as input for second calculation | mannobot | Main CFD Forum | 3 | December 7, 2018 09:37 |
Questions on Boundary Layer Spring Method for 3D Calculation | didiean | FLUENT | 15 | December 18, 2014 21:09 |
MRF and Heat transfer calculation | Susan YU | FLUENT | 0 | June 2, 2010 08:46 |
Defining output as input for second calculation | mannobot | FLUENT | 1 | June 2, 2010 04:20 |
Warning 097- | AB | Siemens | 6 | November 15, 2004 04:41 |