Friends,
I am a beginner in CFD. I have been trying to model fluid flow through a pipe. I have defined the inlet velocity, inlet pressure and outlet pressure apart from the wall boundary condition. The flow regime I have chosen is laminar. At low speeds, I am able to get the expected velocity profile which becomes fully developed as the flow progresses down stream but at higher flow speeds the velocity profile is a bit unexpected. A check of the Reynold's number reveals that it is more than 4000 which means that the flow is likely to be turbulent but since I have assumed laminar flow regime,numerically, should I not get the same velocity distribution as before (even though it may not be so in reality)?
please use the "solid works"and "Ansys" softwares for solving this problem."All the best" for your simulation.
Hi Mir,
What kind of deviations do you observe? In very smooth pipes, flow can easily be laminar with these Reynolds numbers. But, if there is some kind of wall treatment that assumes a certain roughness, that may affect the results. I could also imagine the accuracy of discretization schemes and even regularity of your mesh may affect the solution in this case, as a minor perturbation of the flow may push it towards turbulent behavior. If you can share some more info about software, setup and observations, that would be great!
Could you show the solution? Then, if you are solving the 3D time-dependent equations you are assuming nothing about laminar, transitional or turbulent character of the flow! The flow is solved in a no-model formulation which means you can have either a DNS if the grid is sufficiently fine or a LES no-model (or underesolved DNS) if the grid is coarse. Even in this latter case, the solutions can show some typical features of real turbulence but that must be assessed from the solution and from its statistics.
Then, for laminar steady state I suggest to adopt a smaller Reynolds number
Hi Filippo Maria Denaro , I agree that adopting a small enough Reynold's number would make the behavior of the numerical fluid closer to that of the real fluid but should not the CFD solver give me the solution in the hypothetical case where the flow continues to remain laminar even at high speeds? Is there an explanation for this? I would appreciate if some light is thrown this (ie. why the CFD software should not give a solution that agrees with the assumptions and the pipe flow theory even though it is not physically valid). I mean after all it is just a computer and it knows nothing about actual fluid flows.
That is not exact and dose not depend on the CFD as you are imposing that the solution of the PDE must be steady. This is an arbitrary assumption for a flow at a Reynolds number where transition to a turbulent state (unsteady and 3d) can happen.
Ideally, if you want to study the physical numerical solution of the flow in a pipe at Re=4000, you must solve the unsteady 3d equations using a grid resolution enough fine to have a cell Reynolds number of O(1). That is the Direct Numerical Simulation (DNS).
The fact that in your steady approach you get oscillations in the solution (and I suppose in the residuals) is a signal that the flow is developing an unsteady behavior.
- Badges
- Science topic