[lordvon]

I am preparing a case where OpenFOAM is called from a script to run one iteration at a time. This is to extract forces and moments each timestep to calculate a new RPM and rewriting dynamicMeshDict.

Can someone elaborate on the runtime of this program, stopping and starting OpenFOAM every iteration? How much slower would it be than if I could better integrate the RPM change in the OpenFOAM software?

[protarius]

Actually I'm interested in this problem too.

In the simple case of a turbine rotating around a fixed axis, I ask if it is possible to use a ODE solver for the equation

I \dot{\omega} = M

where \omega is the angular velocity (of the turbine, hence of the GGI), I is the fixed moment of inertia and M is the torque (referred to the rotation axis) that comes from "forces" library.

The angular velocity in the dynamicMeshDict should be "linked" to the above equation.

I'm new in OF, so I don't know if it is correct and possible.

Regards

[lordvon]

Its definitely possible. See here: http://www.youtube.com/watch?v=wjr7NUe5hfg

I was just going to use a form of what you had stated:
(torque) = (rotational inertia) * (angular acceleration)

take (angular acceleration) = (change in angular velocity) / (time step)

then (change in angular velocity) = (time step) * (torque from forces output) / (rotational inertia)

[protarius]

Thanks.

I think this is the point:

> (change in angular velocity) = (time step) * (torque from forces output) / (rotational inertia)

Can I get a reliable result applying only this rough formula? Which is the "right" time step?

I was thinking about applying also some ODE solver to get a better control of the time step, but I don't know if it is necessary.

Regards