Prevent Flow Separation | Impact on Drag & Efficiency of Cars & Wings

[dtms1]

Why is it important to ensure flow separation does not occur in the back of a vehicle or wing?

I know that when flow separation occurs its when you have a large differential pressure which in turn cause turbulence and drag.
I understand that with more drag you have less efficiency.

But if it it occurs behind the object (tail end of the car for example) wouldn't be okay since the object in motion is already in front of the flow separation?
How is the wake behind the object physically affects the car or wing?

 

[etudiant]

The wake creates a pressure differential, as you pointed out.
That pressure differential makes it harder to move forward. That is why it is called drag.

 

[munster]

You need to realize that air pressure pushes on an object in all directions front and back. In motion, more pressure piles up in front, but if the flow was able to avoid separation completely, it actually balances the frontal pressure completely. Then all forces cancel out and there is no drag. But in real life, flow separation always occurs due to the formation of a boundary layer that weakens as the flow goes over to the back. This separation destroys the formation of the equalizing back pressure and replaces it with turbulent air, which is weaker. The excess of frontal pressure over the weakened back pressure is what we call drag.

 

[boneh3ad]

Then all forces cancel out and there is no drag.

No. If the fluid were completely inviscid then this would be true, but unless you are flying/driving/moving through something like superfluid helium, you will have viscous drag.

But in real life, flow separation always occurs due to the formation of a boundary layer that weakens as the flow goes over to the back.

This also isn't true. The formation of the boundary layer is required for separation to occur since it is, by definition, a boundary-layer phenomenon, but not all boundary layers separation. Far from it, actually. It all depends on the pressure gradient. Also, saying the boundary layer weakens really has no meaning. There is no such concept as boundary layer strength.

The excess of frontal pressure over the weakened back pressure is what we call drag.

It is what we call form or pressure drag. It is hardly the only component of drag, though. For example, there is also viscous (skin friction) drag, lift-induced drag and wave drag.

 

[munster]

boneh3ad is technically correct - perhaps I oversimplified and should have qualified my post. But the original question is essentially clarifying the impact of flow separation on drag and that impact is essentially on pressure drag.

The cancellation of front and back forces is what gave rise to D'Alembert's paradox - that there is no drag in inviscid flow while real flows do experience drag. Skin friction, induced and wave drag all have their place. but in practice, skin friction drag is usually a very small component of total drag, induced drag occurs only when the object generates tip vortices and wave drag occurs only at transonic and above speeds - I omitted mention of these other kinds of drag to focus on pressure drag.

"Weakening" is perhaps a poor choice of word to describe what is happening to the boundary layer. I meant the reduction in flow velocity that eventually causes flow reversal in the boundary layer, resulting in flow separation.

Boundary layers separate when faced with an adverse pressure gradient. If there is none e.g. a semi-infinite flat plate, then there is nothing to cause separation. But in practice, boundary layers over bodies in a flow do separate (assuming a viscous continuum fluid). (At subsonic speeds,) a fluid flow encountering a body "sees" an increasing cross-sectional area as it proceeds past the body until a maximum area is reached at the thickest part of the body. In this front portion of the body, streamlines converge as they proceed, flow velocity accelerates and pressure decreases. Past the maximum point, the flow sees a decreasing cross-sectional area until it reaches zero area at the point that the fluid exits the body. Streamlines diverge, flow velocity decelerates and pressure "recovers" (i.e. increases). Pressure recovery creates an adverse pressure gradient i.e. the fluid experiences increasing pressure as it proceeds. This adverse pressure gradient is inevitable for bodies because the body has to end at some point. The velocity in a boundary layer is already slow, especially near the surface, and an adverse pressure gradient causes it to slow down further (weaken?) to a point when the velocity reverses direction. Fluid is now flowing into the boundary layer between the surface and the boundary layer - and this is what is known as flow separation.

As the original question was referring to vehicles and wings, there will always be flow separation at some point "behind". The fluid that fills the space left by the separated flow has less pressure than a flow that stayed attached - and the missing pressure is what causes pressure drag.

So to restate what was meant about flow separation and the boundary layer: "... flow separation always occurs due to (the inevitable) formation of a boundary layer that [STRIKE]weakens [/STRIKE] will inevitably encounter an adverse pressure gradient as the flow goes over the back"

By design, wings have a gradual taper that minimizes the adverse pressure gradient, but "minimal" is still enough to cause flow separation except that it is delayed to occur near the trailing edge (at low angles of attack), allowing most of the flow around the wing to look like an inviscid flow. As in inviscid flow, the pressure over the rear portion of the wing cancels the frontal pressure up to where the flow separates and hence the pressure drag is very low if flow separation occurs at the thin end of the wing. Skin friction drag is unaffected and hence becomes significant when compared to the low pressure drag. At high (i.e. stalling) angles of attack, flow separation happens right up front near the thickest part of the wing - and drag increases sharply as pressure drag takes over.

Wings have induced drag too, but perhaps that is something for another discussion. Most other bodies (cars, balls, frying pans etc) would suffer from flow separation over a large part of the body if they flew through the air - but if a gently tapered shroud was added, the pressure drag would be reduced significantly.

 

[boneh3ad]

in practice, skin friction drag is usually a very small component of total drag

This is not true. On, for example, a Boeing 737, the skin friction drag can account for 15% to 20% of total drag on the vehicle.

boneh3ad is technically correct - perhaps I oversimplified and should have qualified my post. But the original question is essentially clarifying the impact of flow separation on drag and that impact is essentially on pressure drag.

I omitted mention of these other kinds of drag to focus on pressure drag.

The cancellation of front and back forces is what gave rise to D'Alembert's paradox - that there is no drag in inviscid flow while real flows do experience drag.

And this is all fine except you made some sweepingly general statements about drag as a whole, which, to the uninitiated such as the OP, would give the wrong idea, especially when mentioning D'Alembert's paradox which doesn't even hold for unseparated Stokes flow given the fact that viscosity exists.

This adverse pressure gradient is inevitable for bodies because the body has to end at some point. The velocity in a boundary layer is already slow, especially near the surface, and an adverse pressure gradient causes it to slow down further (weaken?) to a point when the velocity reverses direction. Fluid is now flowing into the boundary layer between the surface and the boundary layer - and this is what is known as flow separation.

As the original question was referring to vehicles and wings, there will always be flow separation at some point "behind". The fluid that fills the space left by the separated flow has less pressure than a flow that stayed attached - and the missing pressure is what causes pressure drag.

So to restate what was meant about flow separation and the boundary layer: "... flow separation always occurs due to (the inevitable) formation of a boundary layer that weakens will inevitably encounter an adverse pressure gradient as the flow goes over the back"

I was with you right through the first sentence in this quote. Yes, on just about any two- or three-dimensional body, there will almost certainly need to be a pressure recovery region and there will be some degree of adverse pressure gradient. This, however, does not guarantee flow separation, and there are plenty of streamlined bodies that do not undergo flow separation at their design conditions. Shoot, in one of the wind tunnels I work in, we have such a model. Several, actually. Additionally, the same shape may get different results based on a laminar or turbulent boundary layer, as turbulent boundary layers are quite a bit more resistant adverse pressure gradients than laminar boundary layers. The bottom line is, over a streamlined body, the flow does not always separate.

 

[cjl]

This is not true. On, for example, a Boeing 737, the skin friction drag can account for 15% to 20% of total drag on the vehicle.

I would have thought it was higher than that, actually. I'm surprised it's that low (at least during high speed flight - at low speeds, the majority of the drag is induced drag of course).

 

[GuillermoH]

When the flow stops separating from the trailing edge of a wing, it does not only produces more drag, but it also decreases the production of lift. That is the most important reason why that situation needs to be avoided.
Wings produce lift because the sharp trailing edge gives continuity to the circulation from the wing into the wake. When the flow separates elsewhere, the circulation of velocity around the wing becomes much weaker and the lift decreases dramatically. This occurs a high incidence angle.
Because most wings are not specifically designed to be used beyond stall (except maybe from delta wings), there is no interest in what will happen during stall, but rather when stall will actually occur.
What really becomes an important performance issue in wings is the transition of the boundary layer from laminar to turbulent, because this affects the viscous drag. However, the boundary layer is sensitive to the pressure gradient. Here is where the geometry of the airfoil plays an important role, since it is a delicate balance between the pressure gradient and the boundary layer what keeps the flow attached at high incidence.
Designing the right profile that would produce a long laminar boundary layer and still have a good stall behavior is a sort of art.
An example of profiles that have been designed with the aim of keeping a long favorable (zero) pressure gradient and a long laminar region are the 6 digit NACA. However, as far as I read, it seems they are very sensitive to the quality of the surface.

 

[GuillermoH]

I only wanted to add that induced drag is not only a consequence of the tip vortices. It is a consequence of the span-wise variation of circulation, which is indeed the result of a finite aspect ratio. The induced drag is distributed all over the wingspan, although it does jump in regions of spanwise circulation discontinuity, like the wing tips or the slot between flapped and unflapped regions.
:)