VIIC 1:45 modell

Show off your FreeCAD projects here!
thschrader
Posts: 948
Joined: Sat May 20, 2017 12:06 pm
Location: Germany

Re: VIIC 1:45 modell

Postby thschrader » Fri Jun 15, 2018 7:21 am

Hello Heiko,
thanks for the openfoam case. I had a first look, some hints:

In „constant/transportProperties“ the kinematic viscosity is set
to 1.0e-05, should be 1.0e-06 for water.

In „system/fvschemes“ a steady-state analysis is defined.
Then you should change the time entries in „system/controlDict“.
When running a steady-state calculation, the time is not a physical time
(as it is when doing a transient calculation), the parameter time works as an
incrementation counter for the iterations. Your entries should be
startTime=0
endTime=3000 (for ensuring convergence)
deltaT=1
writeInterval=20

As a first start I would not use a turbulence model. When calculating
the turbulence parameters „by hand“ or using this tool
https://www.cfd-online.com/Tools/turbulence.php
the turbulent kinetic energy/dissipation is nearly zero.

I will try to get the case running.
Thomas
simplefoam.JPG
simplefoam.JPG (28.17 KiB) Viewed 235 times
turbulence.JPG
turbulence.JPG (35.22 KiB) Viewed 235 times
thschrader
Posts: 948
Joined: Sat May 20, 2017 12:06 pm
Location: Germany

Re: VIIC 1:45 modell

Postby thschrader » Fri Jun 15, 2018 8:44 am

Did a steady-state calculation without turbulence.
Solution converges after 180 iterations. The calculated
dragZ-force should be 100 times higher. Maybe a
problem with the mesh? When zooming in the mesh looks corrupt.
The computing-domain is a little bit small.
Time to get the problem into the FC cfd-wb... 8-)
sim001.JPG
sim001.JPG (126.41 KiB) Viewed 220 times
computing_domain.JPG
computing_domain.JPG (97.15 KiB) Viewed 220 times
mesh.JPG
mesh.JPG (98.4 KiB) Viewed 220 times
hds
Posts: 49
Joined: Tue Feb 14, 2017 8:23 am

Re: VIIC 1:45 modell

Postby hds » Fri Jun 15, 2018 8:48 am

what water speed did you use ?

you can check the quality of the mesh with checkMesh. Output says ok.

with transportProperties:
transportModel Newtonian;

nu [0 2 -1 0 0 0 0] 1e-06;
//mu 1e-3;
//nu 1e-6;
rho 1000;

These are the numbers i got so far.

viicForces875rudderDown.jpg
viicForces875rudderDown.jpg (20.74 KiB) Viewed 203 times
Last edited by hds on Fri Jun 15, 2018 11:01 am, edited 2 times in total.
thschrader
Posts: 948
Joined: Sat May 20, 2017 12:06 pm
Location: Germany

Re: VIIC 1:45 modell

Postby thschrader » Fri Jun 15, 2018 9:54 am

hds wrote:
Fri Jun 15, 2018 8:48 am
i get 1N for 0.5m/s...
inlet speed 0,5 m/s, nu=1.0e-06, rho=1000:
laminar: dragZ=0,29 N
turbulent: dragZ=0,91 N
checkMesh gives back "Mesh OK"
Obviously the turbulence has a big influence.
U used K-epsilon model, in FC cfd-wb I could use K-omega.
thschrader
Posts: 948
Joined: Sat May 20, 2017 12:06 pm
Location: Germany

Re: VIIC 1:45 modell

Postby thschrader » Fri Jun 15, 2018 3:03 pm

Hi Heiko,
there must be a point we both overlook.
At 0,5 m/s fluid speed, 0,5 m^2 projected frontal area and a drag-coefficient
of only 0,6 you get 37,5 N drag. To get 1 N drag like in your diagramm above, the drag coefficient
must fall down to 0,02. That is lower than the drag coefficient of a penguin (0,03), which
has the best coefficient of all animals. Mmmh..where is the bug?
Thomas
something_goes_wrong.JPG
something_goes_wrong.JPG (65.76 KiB) Viewed 173 times
thschrader
Posts: 948
Joined: Sat May 20, 2017 12:06 pm
Location: Germany

Re: VIIC 1:45 modell

Postby thschrader » Fri Jun 15, 2018 3:28 pm

aaah...
Heiko, the projected front area is not 0,5 m^2,
that is the value for the complete hull after using "integrate variables"-filter in paraview :oops: .
See my pic above, the slice through the flow channel. According to the blockmeshdict
the channel has an area of 0,3*0,35=0,11 m^2. The sub uses approx 1/6 of the area,
which gives 0,02 m^2 projected front area. Using this value, the total dragZ-force at 0,5 m^2/s
gives 1,5 N (cd=0,6). Not bad, your drag-diagramm is ok.
Thomas
hds
Posts: 49
Joined: Tue Feb 14, 2017 8:23 am

Re: VIIC 1:45 modell

Postby hds » Fri Jun 15, 2018 5:45 pm

Hi Thomas,

these figures in paraview are more and more vodoo to me.
I scaled the VIIC up to 1:1 in FreeCAD.

1:45 speed 0,4m/s -> Drag 0,8N
1:1 speed 4m/s -> Drag 72100N

72100/scale^3 = 0,8 as expected.
Since uploading the files will take too long from home i can do that on Monday if you like.


viicDrag1to1Scale.jpg
viicDrag1to1Scale.jpg (182.95 KiB) Viewed 156 times

Calculate Front Area in paraview (can't this be done in FC ???):
http://www.discretizer.org/node/20
That gives 0,014m^2 in 1:45 scale. Which seems correct given a max diameter of frame 54 of approx 14cm -> 0,0153m^2

viicFrontArea.jpg
viicFrontArea.jpg (211.13 KiB) Viewed 147 times
freecad-heini-1
Posts: 3956
Joined: Tue Jan 07, 2014 11:10 am
Contact:

Re: VIIC 1:45 modell

Postby freecad-heini-1 » Sat Jun 16, 2018 10:06 am

Inspired from your last picture:
phpBB [video]

That is an amazing project.
hds
Posts: 49
Joined: Tue Feb 14, 2017 8:23 am

Re: VIIC 1:45 modell

Postby hds » Sat Jun 16, 2018 12:43 pm

... intense movie indeed.

Here are some findings from the foam simulation:
Forces at 0,5m/s water speed, rudder position 0° i.e. horizontal.

Rudder profile alone without hull: front +1.3e-6N (i.e. practically zero) ; aft +0.0026N

Code: Select all

Rudder profile with hull:  front -0.005N; aft +0.042N; boat overall: -0.11N
Aft rudder semi flattend:  front -0.005N; aft +0.025N; boat overall: -0.12N
Aft rudder flattend:       front -0.005N; aft +0.006N; boat overall: -0.16N
I think one cannot judge a rudder profile on it's own. It has to be done in context with the boat.
The lift force for the whole boat with the aft rudder 17.5° up results in +0.205N. i.e. The aft profile itself produces approx 25% of this force.
The faster the boat goes the more the aft is lifted and pushes the boat downwards by the symmetrical aft profile alone.

Cool i would have never thought of that. Thumbs up for FC plus openFoam !

RuderProfilAchtern.jpg
RuderProfilAchtern.jpg (147.15 KiB) Viewed 101 times

The lift forces for a flattened profile:


AftRudderForcesTopFlatProfile.jpg
AftRudderForcesTopFlatProfile.jpg (39.44 KiB) Viewed 76 times
thschrader
Posts: 948
Joined: Sat May 20, 2017 12:06 pm
Location: Germany

Re: VIIC 1:45 modell

Postby thschrader » Mon Jun 18, 2018 6:34 am

hds wrote:
Fri Jun 15, 2018 5:45 pm
Calculate Front Area in paraview (can't this be done in FC ???):
Hi hds,
for getting the projected frontal area you can draw a sketch around the
(transparent) hull. Then upgrade the sketch to face and measure the area
with the „survey“-tool in arch-wb.

I did several calculations with the VIIC body, in the end I get a drag coefficient cd=0,31
at 0,5 m/s speed. With a „modern“ hull-form (length 1500 mm like your model)
cd is only 0,152.

Using this cd values for the original subs (upscaling factor 45) and 4 knots (2 m/s)
slow cruising speed the VIIC-type needs 38 kW power. The modern type,
which has more than double of the VIIC volume, needs only 45 kW (18% more).
regards Thomas
survey.JPG
survey.JPG (63.72 KiB) Viewed 37 times
modern_type.JPG
modern_type.JPG (45.16 KiB) Viewed 37 times
submarine_design.JPG
submarine_design.JPG (92.97 KiB) Viewed 37 times
calculation.JPG
calculation.JPG (101.95 KiB) Viewed 37 times