Delft Series Resistance Equations - Variety

Discussion in 'Hydrodynamics and Aerodynamics' started by DCockey, Aug 4, 2011.

  1. Mikko Brummer
    Joined: May 2006
    Posts: 574
    Likes: 83, Points: 28, Legacy Rep: 258
    Location: Finland

    Mikko Brummer Senior Member

    With a loose connection to the subject of this thread, I ran some test freesurface simulations last fall with the SYSSER model 50. I chose the 50 as it was rather modern in looks. According to some information it would be based on one of the first Farr designed IMS-boats Gaucho from 1991. The Youtube video shows the simulation at rather a high Fn= 0,65.

     
  2. Leo Lazauskas
    Joined: Jan 2002
    Posts: 2,696
    Likes: 155, Points: 63, Legacy Rep: 2229
    Location: Adelaide, South Australia

    Leo Lazauskas Senior Member

    Nice work, Mikko.
    There seem to be some waves at the walls (at the same longitudinal position as
    the boat) that start and then die out and then appear again. They are not due
    to the V-shaped wake but upstream of them at the wall. Any explanation?
     
  3. Mikko Brummer
    Joined: May 2006
    Posts: 574
    Likes: 83, Points: 28, Legacy Rep: 258
    Location: Finland

    Mikko Brummer Senior Member

    You tell me, Leo! Perhaps the narrow & shallow channel, contracted locally by the model? Also, the boundary conditions in a "circulating" water channel as in the simulation tend to introduce extra turbulence at the inlet (front end) once the turbulence from the model reaches the outlet (back end), to maintain the mass flow. A longer channel, or better still, running the whole simulation as internal in a closed basin (box), with the water still & model moving. More demanding computationally, as if the re-circulating channel wasn't demanding enough.
     
  4. Leo Lazauskas
    Joined: Jan 2002
    Posts: 2,696
    Likes: 155, Points: 63, Legacy Rep: 2229
    Location: Adelaide, South Australia

    Leo Lazauskas Senior Member

    It's a well known problem in reality, i..e that the boundary layer forming on
    the sides and bottom of the tank can become unstable, and a variety of
    other non-linearities can enter into consideration.

    I feel your pain about computation times. Flotilla can sometimes take more
    than 3 minutes to produce a wave pattern ;)
     

  5. Mikko Brummer
    Joined: May 2006
    Posts: 574
    Likes: 83, Points: 28, Legacy Rep: 258
    Location: Finland

    Mikko Brummer Senior Member

    The simulation is relatively low resolution, even if there are some 2,5 million particles. The distance between the nodes of the particle lattice close to the water surface and the model is about 15 mm, while the draft of the hull is about 100 mm. Yet the force results agree pretty well with the experiment - for forces, wall functions are used to resolve the boundary layer, while the particles resolve the outer flow.

    I have taken a few shortcuts: The pitch is fixed to the number given in the experimental results, so the model is free to move only in heave (1-DOF or degree of freedom). This way no need to deal with center of gravity and tow point created moments, more demanding both computationally and in time. If pitch is released the simulation will take much longer to stabilize itself. Heave is necessary to be free, for correct displacement - this too takes computational time, you can see how the model bounces in heave at the beginning of the run.

    The channel is narrower and shallower than in the experiment: 4 m x 1,5 m in depth, with the model bwl 0,6 m. This gives an Fn-depth of about 0,75, still subcritical but much faster you could not go with a 1,5 m depth. On top of all, the simulation is not really two phase (air-water). Instaed, it's water against void, albeit void with athmospheric pressure. The solver can handle true 2-phase, but again it would add heaps of elements and demand computational resources. As it is, 1s of simulation already demands 12 hours, on 2 processors & 8 cores.

    The flow is all turbulent, as the solver doesn't handle laminar-turbulent transition yet. With the Delft models sand grits close to the bow this is maybe not too bad, although I know Uli has looked into this more profoundly. And for the real world case I believe all turbulent is fine, with the initial turbulence in the sea due to even small waves.

    In the grafs attached, total drag vs. speed in the first one, and coefficient form in the second. The skin friction predicted by the sim appears to be pretty close to the ITTC-70% Lwl line also plotted, although in the transient simulation cf varies a lot from one frame to another, and there is no simple way to get an average cf over time, so the numbers are not very reliable. It would appear that also in this case, Delft would predict negative residual resistance at Fn below 0,1.
     

    Attached Files:

Loading...
Forum posts represent the experience, opinion, and view of individual users. Boat Design Net does not necessarily endorse nor share the view of each individual post.
When making potentially dangerous or financial decisions, always employ and consult appropriate professionals. Your circumstances or experience may be different.