Problems Encountered While Modelling A Vertical Floating Cylinder In Ansys Aqwa 13.0

Discussion in 'Software' started by erloroloco, Oct 28, 2013.

  1. erloroloco
    Joined: Oct 2013
    Posts: 6
    Likes: 0, Points: 0, Legacy Rep: 10
    Location: Canary Islands

    erloroloco Junior Member

    There is a Youtube Video

    http://www.youtube.com/watch?v=94V1wgKQbI0

    that shows exactly what Im trying to model.

    Problem is Im struggling to get the body to float in AQWA, and I dont know where the mistake is. Thats why Im resuming a step by step procedure on which parameters I set for the harmonic time response simulation, and whats the outcome of the computation (see the graph attached).

    (note that I work with the Workbench module manager, so everything starts by choosing a harmonic Diffraction module and connecting it to the corresponding harmonic time response module)

    1. DesignModeler: I extrude the cylinder (r=11,5m,L=40m) and slice it by the xy plane. Then I split it into a floating and a submerged part. Finally I shell both parts i.o. to get a thin structure ( ready for a BEM program).
    2. Model setup: Choosing fiberglass (density:36 kg/m3), my mass results in m=5,94e5 kg and the moments of inertia Ixx=Iyy=9,87e7 kgm2 and Izz=3,9e7 kgm2, while the coG is set at (0,0,0). I type these infos into the point mass option, then I autom. run the mesh.
    3. Harmonic Diffraction module modifications:
    - I set the sea grid factor to 4 (I noticed that this is the option that enlarges the animated area :)
    - Wave directions: Intermediate Wave directions:15
    - Wave frequencies: Total number of 8
    4. Insert Hydrostatic Result.
    5. I use the result from the Hydrostatic Calculation: "Actual Volumetric Displacement:8246 m3" and the coB: (0,0,-10) to define in a Point buoyancy.
    6. Harmonic Time response module modifications:
    - Analysis settings: Irregular Wave response; start time: 0 s, time step: 0,001 s, finish time: 50s; "no" cable dynamics and "yes" to convolution (I read it in an ANSYS presentation)
    - Insert Irregular Wave "Pierson Moskotwitz":start freq: 0,1, finish freq: 2 Hz; Hs: 0,88 m; zero crossing period: 10s
    - Insert Structure position graph on the z axis.

    The simulation runs smoothly, only two warnings appear:
    . Mass/Displaced mass are significantly different for structure1. Difference=100,2%
    . Negative Damping detected - STR#1 MODE FREEDOM#3 at different frequencies (0,51-1,03 Hz)

    Now, if you look at the graph I attached, it looks as if the cylinder sinks over time (I double checked it with the AQWAGS plot viewer and it also shows the cylinder dropping down).
    Can you tell me why the hell is doesnt simply oscillate over time, heaving and maybe pitching or rolling a bit?

    You are the only ones I know that can help me get this running, so I hope you can at least give me some advice if Ive forgotten somehting or made some mistakes,

    Look forward to hear from you,

    best regards, Lorenzo.
     

    Attached Files:

  2. jehardiman
    Joined: Aug 2004
    Posts: 3,762
    Likes: 1,152, Points: 113, Legacy Rep: 2040
    Location: Port Orchard, Washington, USA

    jehardiman Senior Member

    Looks like AQWA ran perfectly....Of course it sinks....and it should be obvious why if you just read your post. If it isn't, and you are a student, go back and study your basic hydrostatics before trying to use a program that will do exactly what you tell it to. If this is a commerical venture, you should really seek some professional help with this.

    Anyway, the cylinder has an obvious buoyant displacement of 16,619 kg (i.e. pi*R^2*L...pi*11.5^2*40 = 16,619 m^3), yet you say you entered a mass (weight) of 594,000 kg (i.e. 5.94 * 10^5)....

    Your weight should be the surface area * thickness * material density....not volume * material density. And if you want it to float with half the cylinder exposed you need the weight to be exactly 8309 kg for fresh water, 8516 kg for salt water.

    FWIW, it is very hard to get a cylinder to float vertically in waves without the CG being very close to one end. More that a z offset of -10 is needed.
     
  3. erloroloco
    Joined: Oct 2013
    Posts: 6
    Likes: 0, Points: 0, Legacy Rep: 10
    Location: Canary Islands

    erloroloco Junior Member

    Hi there,

    thank you very much for the given advices. From your recommendations, I suppose I should modify the following parameters:

    m_cyl = 8516 kg
    Displaced_Volume = 16,619 m^3
    coG_z= -10 < coG < -20

    From basic hydrostatics though, I assume that my case reflects the partially submerged structure, so

    Weight of liquid displaced by partially submerged object = weight of object

    Hence, theoretically m_cyl= (density of sea water)*pi*R^2*draught = 1027*pi(11,5)^2*20= 8,46e6 kg (so what should I choose?)

    Up to now, I´ve strictly followed your advice, modified the m_cyl (recalculated then Ixx=Iyy and Izz) and the Displaced_Volume, while choosing the same simulation parameters I described at the beginning.

    Unhappily, the results show the cylinder again sinking into the ocean (see the attached). Again, two warnings appear:
    . Mass/Displaced mass are significantly different for structure1. Difference=100,2%
    . Negative Damping detected - STR#1 MODE FREEDOM#3 at different frequencies (0,51-1,03 Hz)

    Could you tell me from my description, whether Im doing something wrong with my simulation parameters? Yet, I set the Stucture Fixity options to "Free to move" and "Program controlled" Submerged Structure Detection.

    Im sorry but this is a possible commercial venture, but up to now, Im on my own, with little theory support, and I need to get it runnning.

    Hope you can help me a little bit further. Thx a lot in advance.
     

    Attached Files:

  4. jehardiman
    Joined: Aug 2004
    Posts: 3,762
    Likes: 1,152, Points: 113, Legacy Rep: 2040
    Location: Port Orchard, Washington, USA

    jehardiman Senior Member

    Yeah, my bad, should be Tonnes not kg .... 1 m^3 salt water = ~1025 kg not 1.025 kg
    so 851,600 kg

    Anyway, it is becoming too heavy in either case so now I would go look at the model to make sure that the body is actually closed (i.e. your shelled surfaces have lids). I would also check your surface normal vectors (they should be out). Does the top half of the cylinder have buoyancy (it should, but no weight)?
     
  5. erloroloco
    Joined: Oct 2013
    Posts: 6
    Likes: 0, Points: 0, Legacy Rep: 10
    Location: Canary Islands

    erloroloco Junior Member

    Hello jehardiman,

    shouldnt it be m_cyl=8516*1000=8516000 kg and not 851,600 kg as you propose? Then it would nearly match the value of m_cyl=8,46e6 kg I previously calculated.

    Anyway, Ive checked that the normal vectors point outwards. How do I check whether the surface is closed or not?And moreover, do I really need to define a point buoyancy with m=0 kg, Ixx= Iyy=0 kgm^2 and Izz=0 kgm^2 for the upper body part?

    Again, thank you very much for leading me. I think Im getting close to a solution :)

    Regards, lorenzo.
     
  6. erloroloco
    Joined: Oct 2013
    Posts: 6
    Likes: 0, Points: 0, Legacy Rep: 10
    Location: Canary Islands

    erloroloco Junior Member

    Jehardiman,

    Ive now modified the mass into m_cyl=8516000 kg while modifying the inertia values and leaving the Displaced_Volume = 16,619 m^3.

    The simulation works, as you can see from the attached results! The "difference between mass/displaced mass" is still at 50%, but there is no warning about "negative damping detected". From the structure position in the Z coordinate plot it is pretty clear that the cylinder heaves over time and now doesnt sink into the ocean, so it floats. Curiouhsly, Ive checked the animation in AQWAGS and the cylinder capsizes (turns upside down). I guess, it has something to do with the position of the Metacentric Height and hence with the defined coG = (0,0,0).

    Conclusion up to now: There is no need to add an additional point buoyancy on the upper part of the cylinder. But it is fundamental to define correctly the values of mass, coG and inertia, as well as the displaced volume. One should also definitely check that the normals point outwards the panels in DesignModeler (once you select the surface, it should be highlighted in green).

    I still look forward to hear what you got to say about the last 2 posts. And maybe you have some recommendations for the capsizing issue.

    Thanks again,

    Lorenzo
     

    Attached Files:

  7. musiice
    Joined: Sep 2013
    Posts: 17
    Likes: 0, Points: 0, Legacy Rep: 10
    Location: Europe

    musiice Junior Member

    Looks very messy for such a simple task.

    First of all start by doing some hand calculations to determine displaced volume, mass, thus buoyancy, water plane area... etc.

    You don't need so many directions. The number of wave directions you need depend on the geometry and its symmetry.

    Do not cut on the number of wave frequencies, you need to increase it. The maximum is 50 in AQWA. A low number of frequencies results in poor analysis. The range of frequencies, maximum and minimum value, depend very much of your needs and the size of the mesh, the smaller is the mesh the wider it becomes.

    Point Mass is required
    Point of Buoyancy not really unless you are trying to represent elements not shown in the geometry, the panels and/or tubes you define already include its own buoyancy according to their volume.

    I see you have tried to make a lot of steps in once, when you must follow the stages in order.
    LINE -> LIBRIUM -> NAUT/DRIFT/FER


    Regards
     
  8. NavalSArtichoke
    Joined: Oct 2013
    Posts: 431
    Likes: 9, Points: 0, Legacy Rep: 83
    Location: GulfCoast

    NavalSArtichoke Senior Member

    Instead of specifying both the weight and the displacement, shouldn't say just the weight be specified and then let the program find the corresponding waterline and buoyancy? I'm not familiar with the capabilities of AQWA, but similar programs provide a means whereby a loading condition or a waterline is specified, but not both.
     
  9. musiice
    Joined: Sep 2013
    Posts: 17
    Likes: 0, Points: 0, Legacy Rep: 10
    Location: Europe

    musiice Junior Member


    Actually is like you say.

    He should define a point mass in the CM of the Cylinder where he inputs mass, CM coordinates and Moments of Inertia. The buoyancy terms are computed internally to match the geometry.

    On top of that because of incorrect modelling
    Mass/Displaced mass are significantly different for structure1. Difference=100,2%
    This means that the cylinder is completely under the water level.
     
  10. erloroloco
    Joined: Oct 2013
    Posts: 6
    Likes: 0, Points: 0, Legacy Rep: 10
    Location: Canary Islands

    erloroloco Junior Member

    Hi there,

    I recalculated my values again, and here is what I get:

    Cylinder mass:
    - m_cyl= (density of sea water)*pi*R^2*draught = 1027*pi*(11,5)^2*20= 8,46e6 kg

    Moments of inertia:
    - Ixx=Iyy=7,05e9 kgm^2
    - Izz=5,56e8 kgm^2

    - coG=(0,0,0); (xy plane, half the height of the cylinder)

    Now, if I just enter these values into a point mass, while not considering the Point Buoyancy option, the simulation crashes ( the rest of the simulation parameters are the same as stated in the first post ). The error says "Frequency number 1-8 has been omitted. Solve aborted"

    So I define a Point Buoyancy (although I dont have hidden elements in the model) and I have to set two additional parameters, namely:

    Center of Buoyancy
    - coB=(0,0,-10); ( assumed by the geometry, half the cylinder height below water level)
    - Displaced_Volume=pi*R^2*draught=8309,52 m^3

    Now you can see that the simulation runs for t=100s, while the cylinder floats and heaves in the z direction (attached results). Still, two warnings appear:
    1- Mass/Displaced mass are significantly different for structure1. Difference=-50,0%
    2- Read equilibrium position from unconverged Librium run

    I hope this is not so messy anymore :-=) However, I would appreciate some comments or infos on
    - why does the simulation with the point mass option crash?
    - why do I still have a difference of mass/displaced mass?
    - how do I avoid the capsizing issue of the vertical floating cylinder?

    best regards,
    Lorenzo

    P.S: Im running the program from the Workbench, so I guess the program follows the "LINE -> LIBRIUM -> NAUT/DRIFT/FER" order musiice mentioned
     

    Attached Files:

  11. musiice
    Joined: Sep 2013
    Posts: 17
    Likes: 0, Points: 0, Legacy Rep: 10
    Location: Europe

    musiice Junior Member

    Wrong.

    First of all:

    Your Point mass for the cylinder is:
    M=V x rho (kg)
    V = Cylinder volume, is it solid? is it a shell with some thickness?
    rho = cylinder material density (kg/m3)

    Secondly:
    Water density 1025 kg/m3


    Pretty basic stuff.
    Best regards.
     
  12. erloroloco
    Joined: Oct 2013
    Posts: 6
    Likes: 0, Points: 0, Legacy Rep: 10
    Location: Canary Islands

    erloroloco Junior Member

    Musiice,

    for the point mass option, I dont have to declare a volume. But just i.o. to keep the calculations with an overview, the cylinder volume is equal to

    V=pi*R^2*L=pi*11,5^2*40=16619,4 m^3

    The submerged volume, which in this case I assume it to be the displaced Volume is exactly half of V, so

    V_displ=V*0.5=8309,51 m^3

    And with respect to the mass (considering fiberglass with 36 kg/m^3) I agree that the general mass formula is

    m_cyl=(material_density)*V = 598284,9 kg

    though I read in a thesis that a floating cylinder has a mass equal to

    m_cyl=(density of sea water)*pi*R^2*draught, which agrees to the archimedes principle. Thats why I choose here 8,46e6 kg instead

    By the way, the cylinder is considered to be solid, but for the BEM calculations, it has been shell with a thickness of 0 in the DesignModeler.
    Note that the water density equals 1027 kg/m^3 since it is salt water we are calculating with.

    Any comments on the simulation parameters /output?
     
  13. UTKARSH RAMAYAN
    Joined: Oct 2012
    Posts: 9
    Likes: 0, Points: 0, Legacy Rep: 10
    Location: IIT MADRAS

    UTKARSH RAMAYAN Junior Member

  14. UTKARSH RAMAYAN
    Joined: Oct 2012
    Posts: 9
    Likes: 0, Points: 0, Legacy Rep: 10
    Location: IIT MADRAS

    UTKARSH RAMAYAN Junior Member


  15. CWTeebs
    Joined: Apr 2011
    Posts: 232
    Likes: 15, Points: 0, Legacy Rep: 171
    Location: Maine

    CWTeebs AnomalyGenerator

    Post the Aqwa workspace (if you're still having trouble). The negative damping issue is probably due to irregular frequencies, and/too few frequencies (you said 8 were used, use at least 20).
     
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.