Moment of Inertia Issue

sugarl14sugarl14 Member Posts: 14
edited June 10 in Fluids

Hi, I am modeling an OWSC using open channel wave boundary condition and a dynamic mesh. Instead of using a UDF, I am using the built in 6DOF solver to assign rotation to the hinge of the OWSC. I am currently experiencing an issue with my OWSC model in Fluent 2021, similar to the issue found in this post:

https://forum.ansys.com/index.php?p=/discussion/comment/112606#Comment_112606

The specific issue that I am having is this: whenever I start the simulation using dynamic mesh, the OWSC immediately “falls over” with high velocity. The mass of my device is around 300 kg, and the moment of inertia about the hinge is around 1500 kgm2 according to SolidWorks. After trying to isolate the issue, here is what I found: When I run the simulation with +1500 kgm2 as the moment of inertia, the OWSC immediately falls over counter-clockwise. When I run the simulation with -1500kgm2 as the moment of inertia, the OWSC immediately falls over clockwise. When I change the moment of inertia to some large value like 100,000 kgm2, the OWSC doesn’t fall over, and actually begins to act normally by oscillating back and forth as the waves hit it. However, a moment of inertia this large seems quite unrealistic..


The discussion link that I posted above seems to have been inconclusive, and didn't really help my case. I've tried reducing the time step, but nothing seems to be working.. Any ideas?

Answers

  • sdebsdeb Posts: 225Forum Coordinator

    Hello,

    Did you test by changing the initial depth of submersion of the OWSC?

    Can you let the simulation start with a very high moment of inertia and let it stabilize a bit without toppling. Then try to lower the moment of inertia in steps to the actual value?

    Regards,

    SD

  • sugarl14sugarl14 Posts: 36Member

    Hi SD, thanks for the response.

    I ran a quick case with the MOI at 100,000 kgm^2 and let it get situated for about 1 second of simulation time, and the changed MOI to 1500 kgm^2, and then it started having some weird responses even before the wave hit the OWSC. Here's the torque plot:


    I ran another one where I incrementally stepped down from 100,000 to 5,000 and this what happened. Looks like the closer i get to my actual moment of inertia (1500 kgm2), the more the results seem to go crazy, until I get down to 0, then they level back out again.


    What are your thoughts on this?

    Also, I have not changed the depth of the OWSC. Are you recommending to change the depth so the OWSC is completely submerged?


    Thanks SD,


    Landon

  • sdebsdeb Posts: 225Forum Coordinator

    Hello,

    Is there a possibility of resonance with wave frequency matching the frequency of oscillations when the moment of inertia is set to around 1500? Although this sounds far fetched.

    Yes, can you change the depth as well and then test?

    Another recommendation can be to set the Constraints for the rotational angle. For this you can initially constrain the rotation about a single axis.

    You might also look into the Six DOF UDF for more customization [https://ansyshelp.ansys.com/account/secured?returnurl=/Views/Secured/corp/v211/en/flu_udf/flu_udf_sec_define_sdof_properties.html]

    Regards,

    SD

  • sugarl14sugarl14 Posts: 36Member
    edited June 14

    Hi SD,

    The only reason I don't believe the natural frequency and wave frequency is the issue is because the flap falls over instantly, even without the presence of waves. For example, even when I preview mesh motion the flap falls over, so I don't think the waves have an influence on the issue I am experiencing.

    I changed the water depth to completely submerge the OWSC and the results were even worse.. I lowered the timestep to 0.001s and still got a negative cell volume because the flap falls over so fast. It's almost like the moment of inertia is causing some sort of torque on the device...

    Regarding rotational constraints, I originally set constraints for the rotational angle to -30 and +30 degrees.


    I've tried writing a UDF for my case so I could apply an external buoyancy force to pull the flap back up to equilibrium, but didn't have any luck. Here's a my code:


    #include "udf.h"


    DEFINE_SDOF_PROPERTIES(flap, sdof_prop, dt, time, dtime)

    {

     Six_DOF_Object *sdof_obj = NULL;


     sdof_prop[SDOF_MASS] = 345.28985 ;

     sdof_prop[SDOF_IXX] = 0.0;

     sdof_prop[SDOF_IYY] = 0.0;

     sdof_prop[SDOF_IZZ] = 1595.98980;

     sdof_prop[SDOF_LOAD_M_X] = 0.0;

     sdof_prop[SDOF_LOAD_M_Y] = 0.0;

     sdof_prop[SDOF_LOAD_M_Z] = 16664*sin(DT_THETA(dt)[2]); /*16,664 is the hydrostatic restoring coefficient*/



     sdof_obj = Get_SDOF_Object(DT_PU_NAME(dt));


     if (NULLP(sdof_obj))

     {

      /* Allocate_SDOF_Object must be called with the same name as the udf */

      sdof_obj = Allocate_SDOF_Object(DT_PU_NAME(dt));


      SDOFO_1DOF_T_P(sdof_obj) = FALSE;    /* one DOF translation */

      SDOFO_1DOF_R_P(sdof_obj) = TRUE;    /* one DOF rotation */

      SDOFO_DIR(sdof_obj)[0] = 0.0;

      SDOFO_DIR(sdof_obj)[1] = 0.0;

      SDOFO_DIR(sdof_obj)[2] = 1.0; /*rotation about the z-axis only*/

      SDOFO_CENTER_ROT(sdof_obj)[0] = 0.0;  /* only needed for one DOF rotation. Center of Rotation is at [0 0 0] meters */

      SDOFO_CENTER_ROT(sdof_obj)[1] = 0.0;  /* only needed for one DOF rotation. Center of Rotation is at [0 0 0] meters */

      SDOFO_CENTER_ROT(sdof_obj)[2] = 0.0;  /* only needed for one DOF rotation. Center of Rotation is at [0 0 0] meters */

      SDOFO_CONS_P(sdof_obj) = TRUE;     /* constrained */


      if (SDOFO_CONS_P(sdof_obj))

      {

       SDOFO_LOC(sdof_obj) = 0.0;

       SDOFO_MIN(sdof_obj) = -0.52;

       SDOFO_MAX(sdof_obj) = 0.52;

       SDOFO_F(sdof_obj) = 0.0;       /* spring preload */

       SDOFO_K(sdof_obj) = 0.0;       /* spring constant. Also tried with 16664 Nm/rad and it didn't work either */     


       SDOFO_INIT(sdof_obj) = SDOFO_LOC(sdof_obj);

       SDOFO_LOC_N(sdof_obj) = SDOFO_LOC(sdof_obj);

      }

     }

    }



    I don't think it really matters whether I initialize using a Flat or Wavy method, is this correct?

    Please let me know if you have any more help with this. I really need to get this issue figured out!!

  • sdebsdeb Posts: 225Forum Coordinator

    Hello,

    How do you calculate the moment of inertia for this OWSC? Can you double check if the values are correct?

    It looks like the OWSC is unconditionally unstable when you provide this moment of inertia.

    Regards,

    SD

  • sugarl14sugarl14 Posts: 36Member

    Hi SD,

    I calculated the MOI from SolidWorks at the OWSC's hinge. As you can see from the screenshot, the mass of the flap is 345 kg, and the moment of inertia about the rotation axis (Y-axis) is 1596 kgm^2. The moment of inertia about the COG is 462 kgm^2. I've also tried using that in the SDOF Solver and I still get the same instability issues.

    Here's a snippet of my SDOF solver settings. Just to let you know I have calculated the restoring force need to restore the OWSC back to its upright position and used that for the Spring Constant (Nm/rad) term, but it also made no difference in the OWSC instability issue. To reiterate, the only thing that seems to cause the OWSC to act more stable is by setting MOI to 0 kgm^2, or to some large value like 100,000 kgm^2.



    Thank you again for your help and patience. It is very much appreciated!


    Landon

  • sdebsdeb Posts: 225Forum Coordinator

    Hello,

    Can you calculate the net torque acting on the body initially?

    You might want to constrain the rotation completely and calculate the torque just when you start the simulation.

    This might provide some clue as to why it behaves in this unstable fashion.

    Also from the image of the OWSC, it looks as if a slight rotation will cause the gravity vector to fall outside of the body which will then eventually make it unstable.

    Essentially in still water situations, the relative positions of the center of buoyancy and center of gravity will dictate the stability of the object.

    I would suggest to check these conditions again.

    Regards,

    SD

Sign In or Register to comment.