[Fauster]

Dear Foamer,

I have two questions about the dynamic motions libraries of OpenFOAM. As far I understand there is two different libraries for calculating the movement of a rigid body.

src/rigidBodyDynamic

and

src/sixDoFRigidBodyMotion

For example in the DTChull tutorial we have two different kind of dictionary for one or the other.
What are the reason for having this two libraries ? What are the strengths and weaknesses for each one ? In which cases it's advisable to use one or the other ?

Another interesting point about restraints implementation :
In rigidBodyDynamic if you use a linearSpring it's seems that the spring generate a force and a torque on the body according to the source code :

// Force and moment on the master body including optional damping
vector force
(
(-stiffness_*(magR - restLength_) - damping_*(r & v))*r
);

vector moment(attachmentPt ^ force);

But in sixDoFRigidBodyMotion the restraintMoment is null.

restraintForce = -stiffness_*(magR - restLength_)*r - damping_*(r & v)*r;
restraintMoment = Zero;

Is this difference due to the way the six dof motion is solved ? I guess I need to carried out a test case to be sure that the results are equivalent between both libraries but If someone has alrady done the verification it would be amazing !


Thanks for your thought.
Paul

[Krapf]

Hi Fauster,

the difference between the two libraries is explained here: Flow induced 3D motion OpenFoam

Kind regards,
Krapf

[Fauster]

Originally Posted by Krapf

Hi Fauster,

the difference between the two libraries is explained here: Flow induced 3D motion OpenFoam

Kind regards,
Krapf

Thanks for you answer.

For me there are also important differences. The way you define the constraints is very different between both libraries. From my point of view the restraints definition of rigidBodyDynamic.so is closer to commercial CFD packages.

The sixDoFRigidBodyMotion.so is based on this paper : http://link.aip.org/link/?JCP/107/5840/1

For the rigidBodyDynamic.so I didn't find the reference.

Since my last post I find out why there is no moment returned by the restraints in the sixDoFRigidBodyMotion.so library.

The explanation can be found in the sixDoFRigidBodyMotion.C file (inside function applyRestraints)

// Restraint position
point rP = vector::zero;

// Restraint force
vector rF = vector::zero;

// Restraint moment
vector rM = vector::zero;

// Accumulate the restraints
restraints_[rI].restrain(*this, rP, rF, rM);

// Update the acceleration
a() += rF/mass_;

// Moments are returned in global axes, transforming to
// body local to add to torque.
tau() += Q().T() & (rM + ((rP - centreOfRotation()) ^ rF)); <-- the moment generated by the force is calculated here

Paul

[Krapf]

"Rigid body dynamics algorithms" from Roy Featherstone is the source for rigidBodyDynamic: https://github.com/OpenFOAM/OpenFOAM...873570d1e7c8dc

Kind regards,
Krapf

[kiski]

Originally Posted by Fauster

Thanks for you answer.

The sixDoFRigidBodyMotion.so is based on this paper : http://link.aip.org/link/?JCP/107/5840/1

Paul

Could you give the full citation please? the link is broken

[roenby]

Regarding the dead link to the reference above (I know, but better late than never):
I believe the paper referred to is this one: https://doi.org/10.1063/1.474310
As far as I can see, it does not document how the sixDoFRigidBodyMotion library works - only how the symplectic body motion integrator works.