How Does the Airfoil Mechanism Create Lift and Drag Forces?

[jmex]

Hello,

I read about airfoil mechanism that there happens to be pressure difference created above and below the airfoil which result in lift force.

My questions are:
1. What actually happens in the speed of air. I mean how can one side of the air travels with greater speed when they have the same speed in the beginning? And how this whole process happens.

2. What is drag force and how is it created.

3. How is angle of lift help in vawt?

If I want to start from the beginning to study and go deep into this topic guide me in right direction. Also help me with online materials available. I would like to stretch my all time into this project. Just needed a perfect direction to work.

 

[Simon Bridge]

You only have part of the story - another contributing effect is conservation of momentum as the airflow is directed downwards.

1. consider: the air leaves the trailing edge at the same flow rate as it arrives at the leading edge - but it has to take a longer path one way. It follows that it must travel faster along the longer path. It has to get that extra energy from someplace - what is interacting with the air that could transfer energy to it in some way?

The rest of the questions sound more like standard homework questions.
You should have some notes about these in your coursework.
You can also get to the answers by thinking of the air as a fluid that you must somehow shove a big object through.

You experience the drag force yourself when you stick your hand out a car window when you are driving for example: how does that happen? You can also make a crude airfoil with your hand (moving car again) and experiment with changing the angle of attack.

 

[cjl]

1. consider: the air leaves the trailing edge at the same flow rate as it arrives at the leading edge - but it has to take a longer path one way. It follows that it must travel faster along the longer path. It has to get that extra energy from someplace - what is interacting with the air that could transfer energy to it in some way?

It only follows that it must travel faster if you make the implicit assumption that the air takes an approximately equal amount of time to transit over the top of the airfoil and below the bottom. This is not the case. There are good reasons why the air travels faster over the top of the airfoil, but the longer path is not one of them.

In reality, the air travels faster over the top because the sharp trailing edge of the wing ensures that the flow will separate at the trailing edge (rather than somewhere else along the wing). In order for the overall flow field to have a stagnation point at the trailing edge and behave in a way consistent with how fluids behave, the air must have some circulation. This circulation superimposed on the steady flow field causes the flow velocity to be higher on top of the wing, usually to such a great extent that the transit time for fluid traveling over the top of the wing is substantially shorter than the transit time below the wing (somewhat counterintuitively).

 

[Simon Bridge]

Thanks cjl. I think jmex is doing coursework though.
Have you had a go answering the questions in post #1 with that approach in mind?

The reality is that nobody knows exactly how an airfoil works. The questions in post #1 are suggestive of a particular simplified approach being taken in an introductory aerodynamics course.

Without undermining the course, the usual, better, overview goes like this:
http://hyperphysics.phy-astr.gsu.edu/hbase/fluids/airfoil.html
... even that is simplified.

 

[jmex]

Haha
True.. Actually I'm trying to deduce or just say simplifying what clj just said. Whatever I read is making me more confused about its working. BTW thanks for your wonderful example of drag force. And the energy you are talking about is messing my mind.

 

[Simon Bridge]

Well, answering the questions will help me to help you.
However, you should be better able to work these things out now.

 

[rcgldr]

A wing has an effective angle of attack relative to the direction the wing travels thorugh the air, and the flow relative to the wing is diverted downwards. A wing pulls air downwards from above the wing, and/or pushes air downwards below the wing. A reasonable angles of attack, the air is pulled downwards in order to fill what would otherwise be a void created by a receding (relative to the air) upper surface of the wing, once past the peak of the upper surface. The air is pushed downwards off the bottom surface of a wing. For the flow that goes over a wing, just ahead of the wing, the flow separates, and the upper part of the flow initially travels upwards, then curves back downwards in order to follow the upper surface of a wing. This curvature corresponds with a pressure gradient (higher to lower) that causes the air to accelerate towards the wing in order to follow that curve. The low pressure air above a wing also corresponds to air from higher pressure areas to accelerate towards the lower pressure areas, and this results in the air accelerating to a higher speed as it travels over a wing until the pressure starts to increase again, corresponding to the air decelerating (also the flow usually becomes turbulent once the pressure starts to increase).

All of this faster air above a wing stuff is from the wing's frame of reference. From the air's frame of reference, the wing passes through a volume of air, accelerating the air mostly downwards (lift) and somewhat forwards (drag). Typically the air is moving fastest just a bit behind and below the trailing edge of a wing as it passes through a volume of air. From the air's perspective, a wing performs work on the air, changing it's speed from zero to some non-zero downwards and forwards flow. This work occurs due to mechanical interaction between the air and the wing, and most of the work done is related to an increase in pressure as the air is forced downwards from a lower pressure zone above the wing into a higher pressure zone below. The work performed by the wing on the air violates Bernoulli, but Bernoulli applies to the flow further away from the wing.

From the wing's perspective, an ideal wing diverts a relative flow without changing it's energy, in which case Bernoulli would not be violated. A real wing diverts and slows down the relative flow, removing some energy from the relative flow, which also violates Bernoulli, but if the wing is efficient, and the losses small, then Bernoulli can be used to approximate pressures (and in turn lift) if the relative flow speeds are somehow determined.

 

[boneh3ad]

Thanks cjl. I think jmex is doing coursework though.
Have you had a go answering the questions in post #1 with that approach in mind?

It matters not whether this is for coursework or not, the reasoning you originally provided was based on an implicit assumption that does not follow from physics. The increased velocity over the upper surface is unrelated to the longer flow path and is more correctly related to the airfoil geometry and viscosity.

The reality is that nobody knows exactly how an airfoil works. The questions in post #1 are suggestive of a particular simplified approach being taken in an introductory aerodynamics course.

This is also not true. From the aerodynamicist's perspective lift is a lot more complicated than it outwardly appears, but it is fairly well-understood. If it was as poorly understood as you claim, we would have a lot more difficulty calculating lift to the accuracy that we routinely do. This question is also not reminiscent to me of the approach taken in an introductory aerodynamics course (at least not a good one). Usually, introductory aerodynamics courses will discuss Bernoulli's theorem and introduce lift already assuming that something causes the air to move more quickly over the upper surface and then only later on go back and talk about boundary-layer separation and the Kutta-Joukowski theorem. To me, it sounds like the OP is perusing the internet for answers, and especially on this topic, the internet has a lot of bad and conflicting information.

The hyperphysics link is a good one and explains the viewpoints and mentions how both are valid. However, both viewpoints are observational and operate under the assumption that there is a downwash or a velocity difference to begin with and neither addresses why these things occur, which was the nature of the OP's first question. The actual reason is as touched-on by cjl. To go deeper requires essentially either the full Navier-Stokes equations or else a perturbation solution accounting for the boundary layer. Such an analysis (which cannot be done analytically) would show that the sharp trailing edge enforces the location of the trailing stagnation point up until the point of stall, leading to the lift phenomenon.

As to the second question, drag is simply a fluid dynamic force that opposes motion. It comes from a number of sources depending on the velocity and shape of the object. For example:

• There is a component of drag that comes from the pressure difference between the forward-facing surfaces of an object and the rearward-facing surfaces, leading to a net force. That could be caused by a separation bubble behind the object or, depending on which way you prefer to view lift, it is one way to look at lift-induced drag (which can be measured this way).
• There is also a component of drag associated with the viscosity of the fluid. Basically, fluids tend to stick to surfaces (the no-slip condition) and as you move an object through a fluid, it tends to pull a bit of that fluid with it. This requires energy and results in a loss of energy from the movement of the object, or in other words, a non-conservative drag force retards its motion. This is viscous drag.
• If an object is moving supersonically, shock waves form, which induce a very large amount of drag called wave drag. The issue with wave drag is that the object is moving so fast that the air doesn't have a means of smoothly moving around the object, so a shock forms and causes large pressure increases and energy losses. This is the similar to the drag resulting from the wake of a large ship moving through the water.

A real wing diverts and slows down the relative flow, removing some energy from the relative flow, which also violates Bernoulli, but if the wing is efficient, and the losses small, then Bernoulli can be used to approximate pressures (and in turn lift) if the relative flow speeds are somehow determined.

Provided the wing is not stalled, you can easily calculate the velocities regardless of any of the issues you cite. If you simply assume the flow is inviscid and apply the Kutta condition (i.e. enforce the trailing-edge stagnation point artificially), then you can come up with the fully inviscid flow field, including the relevant pressure gradients at the surface. From those you can run a boundary layer analysis and use that to adjust the effective shape of the airfoil and continue iteratively until you arrive at a solution where the inviscid outer flow approximates reality to effectively machine precision, and from that you can get the pressure distribution and the forces on the wing as a result of pressure. Viscous drag would require the boundary layer analysis of the near-wall flow as well.

The only real difficulty is that of laminar-turbulent transition, which is unpredictable. This tends to torpedo many efforts to get an accurate estimate of viscous drag. It affects lift a bit as well since turbulent boundary layers are thicker, but this is a much more subtle effect than that on viscous drag. At any rate, the moral of the story is that you can get a very good approximation of the lift and pressure drag on an arbitrary airfoil shape of essentially any efficiency provided the boundary layer is attached over the entire chord. If the boundary layer separates, life gets a lot more complicated.

 

[rcgldr]

A real wing diverts and slows down the relative flow, removing some energy from the relative flow, which also violates Bernoulli, but if the wing is efficient, and the losses small, then Bernoulli can be used to approximate pressures (and in turn lift) if the relative flow speeds are somehow determined.

Provided the wing is not stalled, you can easily calculate the velocities regardless of any of the issues you cite.

The issue here is that with the loss of energy, somewhere during the wing's interaction with the air, there's a decrease in pressure without a corresponding increase in speed or a decrease in speed without corresponding increase in pressure. Also, trying to calculate the velocities for a wing generally requires something more complicated than Bernoulli, such as a simplified version of Navier Stokes (but not so simplified as to be just Bernoulli equation).

 

[boneh3ad]

Like I said, you have viscous corrections that adjust the inviscid flow field such that it is essentially identical to that which is produced by the airfoil with viscosity effects included. The effects you cite are viscous in nature and the mechanism you cite is viscous dissipation. Outside of the boundary layer, viscous dissipation is zero and Bernoulli's equation holds. What I described takes that into account. It is a well-known process. There are, of course, a number of similar methods of doing this. Take a look at XFOIL, for example.

The reason Bernoulli's equation still works is that the flow can be broken down essentially into two regions, the free stream and the boundary layer. The free stream behaves as an idea fluid and therefore has no viscous dissipation and Bernoulli's equation applies. The boundary layer represents the small area near the surface where viscous dissipation is non-negligible, and Bernoulli's equation does not apply. As luck would have it, the pressure gradient through a boundary layer in the direction normal to the wall is effectively zero, so the pressure predicted at the boundary layer edge by the inviscid flow field is the same as the pressure experienced by the surface immersed in the boundary layer. Thus, even though energy is dissipated near the airfoil and Bernoulli, therefore, does not apply in the strictest sense, this is confined to the boundary layer. Due to the nature of the boundary layer pressure gradient, we have simply lucked out in that nature dictates that it doesn't matter as long as we can get the inviscid outer flow correct and then use Bernoulli on that. Every once in a while, you get lucky.

 

[rcgldr]

The effects (losses) you cite are viscous in nature and the mechanism you cite is viscous dissipation.

I only have a vague memory of this, but to simplify what I remember ... Using the wing as a frame of reference, assume the velocity of the input flow V_initial = Vx_initial. The diverted flow has x and y components and the velocity of the diverted flow = V_diverted = sqrt((Vx_diverted)^2 + (Vy_diverted)^2). For a real wing, the process of diverting the flow reduces the velocity of the flow, V_diverted < V_initial. Part (or most?) of this decrease in velocity is related to wing tip and wing trailing edge vortices, (which consume mechanical energy without contributing to net velocity) and part due to heat (friction).

So you're stating that all of this decrease in velocity is due to viscosity in the boudary layer? One issue would be where the vortices that flow off the trailing edge begin to form relative to the leading and trailing edges of a wing. Complicating matters (at least to me), is the fact that delta wings can take advanted of leading edge vortices that move outwards across the leading portion of a delta wing, helping to contribute to lift and allowing for large angles of attack (20° or so) before stalling.

 

[boneh3ad]

I only have a vague memory of this, but to simplify what I remember ... Using the wing as a frame of reference, assume the velocity of the input flow V_initial = Vx_initial. The diverted flow has x and y components and the velocity of the diverted flow = V_diverted = sqrt((Vx_diverted)^2 + (Vy_diverted)^2). For a real wing, the process of diverting the flow reduces the velocity of the flow, V_diverted < V_initial. Part (or most?) of this decrease in velocity is related to wing tip and wing trailing edge vortices, (which consume mechanical energy without contributing to net velocity) and part due to heat (friction).

So you're stating that all of this decrease in velocity is due to viscosity in the boudary layer? One issue would be where the vortices that flow off the trailing edge begin to form relative to the leading and trailing edges of a wing.

The decrease in velocity is probably more apparently in the frame of reference you often like to cite: watching the shape pass through a stationary medium. Due to the presence of the boundary layer, it will tend to drag some air with it. With a suitable shape, it will also direct some of that downward, generating lift.

The vortices you seem to be describing are a result of boundary-layer separation, which is one situation I mentioned where my previous statements generally break down. For a simple, 2-D wing under normal operating conditions, you will have no separation and the only vortex will be the lift-induced vortex which is shed by the wing due to the conservation of angular momentum (or conservation of vorticity, if you will) due to the circulation set up around the wing in generating lift.

However, this again would not occur without viscosity, and the losses in such a system are still viscous and confined to the boundary layer. The actual vortices shed from this system can generally be described inviscidly as long as some measure was taken in the calculations that acknowledges viscosity's role in fixing the trailing edge separation point.

Complicating matters (at least to me), is the fact that delta wings can take advanted of leading edge vortices that move outwards across the leading portion of a delta wing, helping to contribute to lift and allowing for large angles of attack (20° or so) before stalling.

These vortices are special case and are a result of boundary-layer separation. The phenomenon is called leading edge separation, and delta wings rely on it for lift in high angle of attack maneuvers. These, again, involve separation, which was on condition under which my previous statements were invalid.

 

[rcgldr]

The decrease in velocity is probably more apparently in the frame of reference you often like to cite: watching the shape pass through a stationary medium.

In the frame of the medium it's an increase, from zero, to some non-zero downwards (lift) and a bit forwards (drag) velocity. In the case of a glider in a stead descent, per unit time, the decrease in gravitational energy equals the increase in energy of the affected air (and losses to heat).

The vortices you seem to be describing are a result of boundary-layer separation, which is one situation I mentioned where my previous statements generally break down. For a simple, 2-D wing under normal operating conditions, you will have no separation and the only vortex will be the lift-induced vortex which is shed by the wing due to the conservation of angular momentum (or conservation of vorticity, if you will) due to the circulation set up around the wing in generating lift.

My focus was more on a real world 3-D wing. I thought XFOIL handled some amount of turbulence and separation issues.

 

[boneh3ad]

My focus was more on a real world 3-D wing. I thought XFOIL handled some amount of turbulence and separation issues.

It does in a very rudimentary way. It predicts transition using what is called the e^N method, which is flawed but about the best you can do without substantial computational resources. Once you have a transition point nailed down (or just assume entirely laminar or turbulent flow), you can predict separation reasonably well many times. The trick is whether or not the transition assumptions were valid. This is why wind tunnel tests are still important. XFOIL does have some capability in that area. I don't know how well it would do with a separation-dominated flow like a high α delta wing, though. I'm not familiar enough with it.

My point in bringing up XFOIL in the first place was to illustrate how an inviscid analysis (in this instance a vortex panel method) can provide essentially correct values for lift.

 

[WiseGreatTrixie]

In reality, the air travels faster over the top because the sharp trailing edge of the wing ensures that the flow will separate at the trailing edge (rather than somewhere else along the wing). In order for the overall flow field to have a stagnation point at the trailing edge and behave in a way consistent with how fluids behave, the air must have some circulation. This circulation superimposed on the steady flow field causes the flow velocity to be higher on top of the wing, usually to such a great extent that the transit time for fluid traveling over the top of the wing is substantially shorter than the transit time below the wing (somewhat counterintuitively).

I am not sure words are the best at explaining concepts. What do you mean by the trailing edge? Do you mean relative to the side, or the top view? I think a picture or diagram would explain it much better.