Seyyed Arash Mousavi

Let say this "your results are a function of initial condition". Many simulations in CFD, Specially turbulence flow, they show the same behavior.

In my view point, we can not say which one is more accurate. First of all:

- Maybe your physics hasn't unique answer, which is very normal for fluid flow.

- If i want to choose, I will select first one. Because in this condition, transient flow is more realistic, then is better choice for initial condition.

- If you are doing 2D simulation not 3D, then because your result is a function of initial conditions, maybe you should think about simulation in 3D not 2D.

Hi Seyyed,

From my experience, if you are working in the laminar regime, then your steady state solution should match with the transient solution after sufficient time (when the flow field becomes constant with time).

Since you are working with air (with very low viscosity), your Grashof number tends to be high in general. So if your Grashof number already indicates that it is in the turbulent regime, your steady state laminar solution would not be valid. In that case, a turbulent formalism should be used.

I hope it helps.

Rituraj Borah Thank you very much for sharing your answer.

I am working in the laminar regime. Although the working fluid is air, due to the low temperature difference, the Grashof number is very smaller than the critical number. Therefore, turbulent flow is not the case.

You mentioned that the transient solution after sufficient time is when flow field become constant with time. What about heat solution? Does it need to become constant too?

Thank again.

Amir Reza Ansari Dezfoli Thank you for your answer. How about for the basic and simple problems? For instance, a flat hot vertical wall with a constant temperature which is causing a laminar free convection. The correlation for convective heat coefficient for the vertical plate in 2D is provided in reference books. Should I be getting same results from Comsol matching with the correlation?

Thank you.

Hi Seyyed,

Yes. Indeed. The transient solution must reach a steady state (both the velocity and temperature fields) which is actually the steady state solution. Since for natural convection, the velocity and temperature fields are coupled by the Boussinesq approximation term, both the fields will converge to constant values simultaneously.

Rituraj Borah Understood. Great! Now my plan is to keep it a simple geometry and try both methods. Hopefully I get same results and then I will proceed with the more complicated cases. Thanks again for your help.

Best regards.

Seyyed Arash Mousavi

If you are using the same boundary conditions and the same assumption (governing equations, simplifications of NS equation or energy equation, materials (Air) properties ...) with " reference books ", then YES. Otherwise NO. Of course still we expect close results to the reference.