[say_cfd]

In FloEFD the computational domain around a body (say an led assembly having a constant heat flux source subject to natural convection) is automatically created. While simulating a flow situation of fully external flow past such a body, what boundary condition is applied on all the six bounding faces of the computational domain?

[Boris_M]

If you haven't specified a certain velocity for in or outflow then the BC behind that is a pressure with the specified value and if gravity is enabled and the pressure potential option activated, the pressure varies from the defined pressure in the initial condition as the value in the global coordinate system center variing in gravity direction.
This is barely noticeable with gasses due to the density of course but you would see it in a larger simulation such as a submarine under water or so.
But the boundary conditions can be partially violated due to the flow downstream. For example if you define a mixture in the initial condition this mixture is always entering the computational domain but it is not forced that way to leave it of course as another source could lie within the domain with a different fluid adding to the mixture with which is initialized or coming into the domain.

Do you have any problems or were you asking just out of curiosity?

Boris

[say_cfd]

Thanks Boris.

It was out of curiosity as I could not see the clear picture of how FloEFD is treating the value of pressure and velocity at the faces of edge cells. Basically for solving the governing equations and necessary pressure & velocity corrections, any cfd solver would consider/ assume an initial value of the pressure and velocity at all the edge cell faces.
Could you also give some idea about how FloEFD takes care of changes in the velocity due to natural convection ?
I have seen people using Ansys, who create a big far field (computational domain) around a heat emitting body and declaring the edges of computational domain as walls (like du/dy = 0, dv/dx = 0 etc.). And after simulation the velocity vectors shows recirculation (indicating the rise of hot air due to buoyancy which hits the walls and comes back) whereas in FloEFD such a recirculation is not shown instead a much better velocity plots are observed (indicating the rise of hot air but it leaves the domain). How exactly are the velocity components dealt with in FloEFD ?

[Boris_M]

Well you can specify a velocity in x, y and z direction as an initial condition which will then be kept constant as an inflow during the calculation or if defined time dependent then of course not constant. But for example in case of an external aerodynamics case you define the flow in x direction and the computational domain in +x and -x are then the main inlet and outlet faces for the flow except of course for any flow that escapes in y or z direction due to blockage in the flow which is why in some cases you should use a larger far field but in natural convection the walls are treated more as pressure opening if no flow is forced. So the flow leaves the top of the computational domain due to boyancy and new fluid enters from the sides and bottom due to conservation laws. You could also define a box with pressure openings and let the simulation take its turns.
In some cases this is used such as in a case where for example you have a vacuum cleaner and it sucks air in but you don't want to model the whole exterior. In this case you can specify a plenum with a pressure opening. Just as an example.

Boris