[jealor]

Hi everyone,
I want to simulate the motion of a turbine,

Unlike tradition operation where we set a rotation speed，I want to achieve that , the rotation speed autamatically change according to the torque generated by the fluid field(a initial speed is given). I achieved this in Fluent with 6DOF model however it takes too much time, so I want to know can this be achieved in CFX?

I have thought about using expresion to calculate the rotation speed, but I found that there is no method to get the torque and speed in the previous time step to determine the speed in current time.

Is there anyway to achieve such simulation? Thank you all very much.

[urosgrivc]

Usualy you set up a fev simulations each having diferent RPM
than you plot a graph of (residual moment (load-result from simulation)vs rpm) vhere moment is 0 that is your RPM (as that would be steady state in real life) than you run a simulation at those rpms.
there is something about this on the forum maybe if I find it I will post it here.
I doo see that this could be automated by expresions.

[ghorrocks]

This question gets asked many times - I should write an FAQ on it.

Getting the rotation speed from a 6DOF model is slow, difficult and tricky to set up. It is not recommended. A far better way of doing it is to run a series of speeds and output the torque from these simulations. Draw a graph of torque versus speed. Then draw the graph of your load versus speed. The point at which these lines intersect is the steady state operating point. This approach is MUCH easier and more reliable.

[jealor]

Originally Posted by ghorrocks

This question gets asked many times - I should write an FAQ on it.

Getting the rotation speed from a 6DOF model is slow, difficult and tricky to set up. It is not recommended. A far better way of doing it is to run a series of speeds and output the torque from these simulations. Draw a graph of torque versus speed. Then draw the graph of your load versus speed. The point at which these lines intersect is the steady state operating point. This approach is MUCH easier and more reliable.

Thanks for your reply.
Steady simulation has been conducted and I think there might be difference between steady simulation and actrual behavior, so I want to conduct the transient run. But as you said, 6DOF method is really slow, and very easy to become divergence when time step is big. Maybe this effort will result in failure.......