3D External Flow of Submerged Vessel leads to strange results in Fluent?

Sovereignty · February 1, 2019, 02:49

Hello Everyone,

I'm currently investigating the 3D imcompressible flow for a submerged submarine moving at 20 knots (=10.29 m/s), in which it cruises horizontally in seawater to determine the amount of drag exerted upon the body. I made some assumptions, wherein the motion of the submarine is steady and horizontal (for further simplicity), and for the flow to be turbulent, since the underwater vessel is quite large (I'm talking at least 500 m for the length of the submarine.), and steady-state.

I developed a CAD model of the hull, and made a 3D fluid domain by using DesignModeler to create an enclosure (its dimensions are 3474.72 m x 1737.36 m x 2316.48 m), using the Boolean tool and "Slice" tool to create a split fluid domain across the hull. I identified several named selections for the fluid domain before moving on to ANSYS Meshing.

I then used ANSYS Meshing, with a LOT of trial and error, to develop a 3D unstructured tetra mesh from the geometry, where I ended up with a minimum orthogonal quality = 0.30229 (average is 0.75) and the maximum skewness = 0.69771 (average skewness is 0.24834). Unfortunately, I was unable to generate a inflation boundary layer around the hull since it led to high skewness.

Upon exporting the mesh to ANSYS Fluent, I used a realizable k-epsilon model, along with enhanced wall function, to account for the y+ wall condition and boundary layers for laminar/turbulent flows around the hull. I setup the upstream as the "velocity-inlet", the top/bottom/side/symmetry planes as "symmetry", and the downstream as "pressure-outlet", with a gauge pressure of 0. The surface of the hull was set to "Wall" for no-slip. I set the Reference Values as well, in which I identified the surface area of the hull (cut in half) as 70516 m^2, and the length of the hull as 579.12 m, since those values would be used in calculations for the Cd and Cl. Upon setting up the initialization and running the iterations for the solutions of convergence for drag and drag coefficient, I ended up with these plots:

The results from the plot are not what I expected, or at least, didn't imagine where they would end up. I find it strange how the values for both drag and drag coefficient level off around zero as the iterations continue on. I was wondering if anyone knows what issues I might be experiencing with my CFD analysis, and whether or not there's changes to be made? I've spent the past month trying to build a good geometry, mesh, and setup for the analysis, but I have no idea if there's any convergence for the drag.

Best Regards,

Jeff

Sovereignty · February 1, 2019, 03:34

What I found interesting about my following approach to this was that I changed the initialization from standard to hybrid, changed the velocity from 10.289 to 5.14 m/s, and these changes led to slight changes in the plots for drag and Cd:

I'm not sure whether these changes are related to whatever errors are affecting my results, is it possible that the setup for initialization might be another factor into this as well?

February 1, 2019, 02:49	3D External Flow of Submerged Vessel leads to strange results in Fluent?	#1
Sovereignty New Member Jeff Join Date: Jan 2019 Posts: 3 Rep Power: 7	Hello Everyone, I'm currently investigating the 3D imcompressible flow for a submerged submarine moving at 20 knots (=10.29 m/s), in which it cruises horizontally in seawater to determine the amount of drag exerted upon the body. I made some assumptions, wherein the motion of the submarine is steady and horizontal (for further simplicity), and for the flow to be turbulent, since the underwater vessel is quite large (I'm talking at least 500 m for the length of the submarine.), and steady-state. I developed a CAD model of the hull, and made a 3D fluid domain by using DesignModeler to create an enclosure (its dimensions are 3474.72 m x 1737.36 m x 2316.48 m), using the Boolean tool and "Slice" tool to create a split fluid domain across the hull. I identified several named selections for the fluid domain before moving on to ANSYS Meshing. I then used ANSYS Meshing, with a LOT of trial and error, to develop a 3D unstructured tetra mesh from the geometry, where I ended up with a minimum orthogonal quality = 0.30229 (average is 0.75) and the maximum skewness = 0.69771 (average skewness is 0.24834). Unfortunately, I was unable to generate a inflation boundary layer around the hull since it led to high skewness. Upon exporting the mesh to ANSYS Fluent, I used a realizable k-epsilon model, along with enhanced wall function, to account for the y+ wall condition and boundary layers for laminar/turbulent flows around the hull. I setup the upstream as the "velocity-inlet", the top/bottom/side/symmetry planes as "symmetry", and the downstream as "pressure-outlet", with a gauge pressure of 0. The surface of the hull was set to "Wall" for no-slip. I set the Reference Values as well, in which I identified the surface area of the hull (cut in half) as 70516 m^2, and the length of the hull as 579.12 m, since those values would be used in calculations for the Cd and Cl. Upon setting up the initialization and running the iterations for the solutions of convergence for drag and drag coefficient, I ended up with these plots: The results from the plot are not what I expected, or at least, didn't imagine where they would end up. I find it strange how the values for both drag and drag coefficient level off around zero as the iterations continue on. I was wondering if anyone knows what issues I might be experiencing with my CFD analysis, and whether or not there's changes to be made? I've spent the past month trying to build a good geometry, mesh, and setup for the analysis, but I have no idea if there's any convergence for the drag. Best Regards, Jeff

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Simulating a forced external flow loop with 2 phases	CeesH	FLUENT	1	August 27, 2018 08:31
[ANSYS Meshing] Creation of regions within a mesh (for external flow) Ansys Fluent	S_Gut	ANSYS Meshing & Geometry	0	March 16, 2017 10:37
Get Maxwell ansoft results in ansys workbench and find the flow velocity in Fluent, i	rrahman	FLUENT	0	April 16, 2016 22:11
External Flow Analysis of a Rotating Wheel at FLUENT	ericclaro	FLUENT	2	March 22, 2015 21:36
Exporting Fluent results as 4D images in an interactive window	Luca95314	ANSYS	0	October 11, 2010 18:16

February 1, 2019, 03:34		#2
Sovereignty New Member Jeff Join Date: Jan 2019 Posts: 3 Rep Power: 7	What I found interesting about my following approach to this was that I changed the initialization from standard to hybrid, changed the velocity from 10.289 to 5.14 m/s, and these changes led to slight changes in the plots for drag and Cd: I'm not sure whether these changes are related to whatever errors are affecting my results, is it possible that the setup for initialization might be another factor into this as well?