# Will "filling" areas with re-circulation always eliminate that flow separation

 Register Blogs Members List Search Today's Posts Mark Forums Read

 August 28, 2016, 16:01 Will "filling" areas with re-circulation always eliminate that flow separation #1 New Member   Michael Brooker Join Date: Feb 2016 Posts: 6 Rep Power: 9 Hi, i'm a new user here and a beginner to the field of fluid mechanics. In my little experience and human intuition when I've tried to design around flow separation I've found that "filling" these areas or essentially having the geometry follow where the natural flow wants to go seems to eliminate flow separation. For instance, in the example of a backward facing step, if you were to analyse how the flow separates and alter the geometry such that there was a transition following the natural path of the flow you would expect no separation. My question is, is this always the case (with subsonic flows and standard conditions for air), and if not, what is the reason and could you give me some scenarios? Thanks, this is just for curiosity

August 28, 2016, 17:12
#2
Senior Member

Filippo Maria Denaro
Join Date: Jul 2010
Posts: 6,672
Rep Power: 71
Quote:
 Originally Posted by Michael B Hi, i'm a new user here and a beginner to the field of fluid mechanics. In my little experience and human intuition when I've tried to design around flow separation I've found that "filling" these areas or essentially having the geometry follow where the natural flow wants to go seems to eliminate flow separation. For instance, in the example of a backward facing step, if you were to analyse how the flow separates and alter the geometry such that there was a transition following the natural path of the flow you would expect no separation. My question is, is this always the case (with subsonic flows and standard conditions for air), and if not, what is the reason and could you give me some scenarios? Thanks, this is just for curiosity
no, separation occurs for other reason, such as transition of regime..think about the flow over a flat plate, you get separation at a certain x-location due to the transition from laminar to turbulence stage

 August 28, 2016, 18:32 #3 New Member   Michael Brooker Join Date: Feb 2016 Posts: 6 Rep Power: 9 Thanks, I hadn't thought about this scenario. Is it possible that it might also work in this example? If the geometry of the flat plate started to follow the natural flow separation at that point I would imagine it would cause a locally higher pressure in that region therefore delaying flow separation due to less of an adverse pressure gradient?

August 28, 2016, 18:41
#4
Senior Member

Filippo Maria Denaro
Join Date: Jul 2010
Posts: 6,672
Rep Power: 71
Quote:
 Originally Posted by Michael B Thanks, I hadn't thought about this scenario. Is it possible that it might also work in this example? If the geometry of the flat plate started to follow the natural flow separation at that point I would imagine it would cause a locally higher pressure in that region therefore delaying flow separation due to less of an adverse pressure gradient?
The example of the flow over a flat plate consider separation due to instability of the flow that becomes unsteady and three-dimensional.

As different example, in flow over wing sections, it is quite common to work with geometries that tend to mantain a laminar condition, however separation is a quite common phoenomenon.

Trying to prevent separation or to produce a delay is a hystorical issue in aerodynamics and fluid dynamics.

Prevent separation for all types of geometries and flow conditions is not possible

August 29, 2016, 12:23
#5
Senior Member

duri
Join Date: May 2010
Posts: 245
Rep Power: 16
Quote:
 Originally Posted by FMDenaro no, separation occurs for other reason, such as transition of regime..think about the flow over a flat plate, you get separation at a certain x-location due to the transition from laminar to turbulence stage
Separation is not transition. Flat plate at zero angle of attack never separates. Flow separation can occur in laminar as well in turbulent boundary layers.

August 29, 2016, 12:36
#6
Senior Member

Filippo Maria Denaro
Join Date: Jul 2010
Posts: 6,672
Rep Power: 71
Quote:
 Originally Posted by duri Separation is not transition. Flat plate at zero angle of attack never separates. Flow separation can occur in laminar as well in turbulent boundary layers.

Flow over flat plate a zero angle of attack has separation and reattachement, unsteady and three-dimensional, due to transition! What you are thinking about is in terms of the statistically averaged velocity field that shows no separation.

August 29, 2016, 12:50
#7
Senior Member

duri
Join Date: May 2010
Posts: 245
Rep Power: 16
Quote:
 Originally Posted by Michael B I've tried to design around flow separation I've found that "filling" these areas or essentially having the geometry follow where the natural flow wants to go seems to eliminate flow separation. For instance, in the example of a backward facing step, if you were to analyse how the flow separates and alter the geometry such that there was a transition following the natural path of the flow you would expect no separation.
By filling the area you reduced the adverse pressure gradient. This will work only for bluff bodies like backward facing steps. In case of streamline bodies filling area will not reduced adverse pressure gradient it may worsen it. You can't apply this technique to airfoil. There are several separation control techniques available which can eliminate or push separation to greater extent.

August 31, 2016, 22:52
#8
Senior Member

duri
Join Date: May 2010
Posts: 245
Rep Power: 16
Quote:
 Originally Posted by FMDenaro Flow over flat plate a zero angle of attack has separation and reattachement, unsteady and three-dimensional, due to transition! What you are thinking about is in terms of the statistically averaged velocity field that shows no separation.
At separation reverse flow should occur with in boundary layer and skin friction would become zero at the point of separation. Skin friction doesn't go to zero during transition.
Vortex lift off during transition doesn't mean it is separated. It actually energize the boundary layer from main stream mixing.

 September 1, 2016, 04:42 #9 Senior Member   Filippo Maria Denaro Join Date: Jul 2010 Posts: 6,672 Rep Power: 71 yes, I was not enough clear using the term "transition". I dont want to say that separation happens "in" the transitional region but that separation happens thereafter in the onset of turbulent region due to the previous transion.

 September 1, 2016, 14:43 #10 New Member   Michael Brooker Join Date: Feb 2016 Posts: 6 Rep Power: 9 Thanks for the info guys.

 Tags backward facing step, design, flow separation, recirculation