[VarunV]

Hello,

We are simulating some test cases for a propeller, which requires that we do a transient simulation. The common advice on sizing of the mesh is that the pressure outlet should be far enough so that the wake dies out before it reaches it. But since this is a transient simulation, it takes a lot of time steps for the wake to fully reach the outlet. And even when it is doing so, I notice on my lift and drag monitors that the forces are converged. Therefore, I was thinking if I should run a steady state simulation first, wait for wake to fully develop, and then switch to transient?

[FMDenaro]

Originally Posted by VarunV

Hello,

We are simulating some test cases for a propeller, which requires that we do a transient simulation. The common advice on sizing of the mesh is that the pressure outlet should be far enough so that the wake dies out before it reaches it. But since this is a transient simulation, it takes a lot of time steps for the wake to fully reach the outlet. And even when it is doing so, I notice on my lift and drag monitors that the forces are converged. Therefore, I was thinking if I should run a steady state simulation first, wait for wake to fully develop, and then switch to transient?

Well, it depends on what you really want to simulate. Let me clarify:


1) you want to simulate the whole transient, that is the flow problem starting from some suitable physical initial and boundary conditions.
2) you are interested only in the final fully developed unsteady flow when an equilibrium is reached.


Clearly, only case 2 can be started from arbitrary initial conditions and you can use a converged steady solution.

Be aware that, depending on the formulation you are using (DNS/LES/DES/URANS) the things can be more complex.

[VarunV]

Originally Posted by FMDenaro

Well, it depends on what you really want to simulate. Let me clarify:


1) you want to simulate the whole transient, that is the flow problem starting from some suitable physical initial and boundary conditions.
2) you are interested only in the final fully developed unsteady flow when an equilibrium is reached.


Clearly, only case 2 can be started from arbitrary initial conditions and you can use a converged steady solution.

Be aware that, depending on the formulation you are using (DNS/LES/DES/URANS) the things can be more complex.

I want to simulate the forces acting on the propeller when it is exposed to oblique flow. Barring the initial wake development, I would be interested in the force variation as the blades go through a complete revolution. So I guess case 2 is what I'm looking for.


So when I go with case 2, if I understand correctly, the steady state simulation will converge with a full wake, which the transient simulation will finally refine. Am I correct?