race car aerodynamics, designing for speed 空气动力学 - 图文

Race car aerodynamics,designing for speed

Fig.1-7. Trends showing the increase of the maximum cornering acceleration over the past years for

race cars with and without aerodynamic downforce.

Fig.1-11 Schematic description of the “ground effect” that increases the aerodynamic lift of wings

when placed near the ground.

- 1 -

Fig.2-1. Streamlines in a steady-state flow over an airfoil.

By observing several streamline traces in the flow(as in Fig.2-2), it is possible to see if the flow follows the vehicle’s body shape close to its surface. When the streamlines near the solid surface follow exactly the shape of the body(as in the upper portion of Fig.2-3) the flow is considered to be attached. If the flow does not follow the shape of the surface(as seen behind the vehicle in Fig.2-2 and in the lower part of Fig.2-3) then the flow is considered detached or separated. Usually such separated flows behind the vehicle will result in an unsteady wake flow, which can be felt up to large distances behind the vehicle. As we shall see later, having attached flow fields is extremely important in reducing aerodynamic drag and/or increasing downforce.

Fig.2-2. Visualization of streamlines(by smoke injection) during a wind-tunnel test. Courtesy of

Volkswagen AG.

Fig.2-3. Attached flow over a streamlined car(A), and the locally separated flow behind a more

realistic automobile shape(B).

- 2 -

Fig.2-4. Side view of the velocity distribution near a flat plate in a free stream, V∞(velocity profile on

the upper surface is described by the V vs z graph)

Now, leaving momentarily the discussion about the basic definitions, we can observe a couple of interesting features in Fig.2-4. First, the air velocity near the surface comes to a halt! This is known as the “no-slip-condition.” The fluid particles touching the body will stick to the surface; they have no relative velocity. Farther away from the solid body the velocity increases, until it equals the local free-stream value. This thin boundary is termed the boundary layer and will be discussed with more detail later.

Fig.2-5. Fluid particle traces in laminar and in turbulent flow.

- 3 -

Fluid Properties: e.g., temperature, pressure, density, viscosity., etc.

Viscosity is, in a very generic sense, a measure of fluid resistance to motion (similar to friction), and is

designated by the Greek symbol μ.

The effect of viscosity in a fluid can be demonstrated by the simple example shown in

Fig.2-6(following the analogy to dry friction) where a viscous fluid is placed between two parallel, solid surfaces. The lower surface is stationary, while the upper one is moving to the right at a constant speed. The fluid particles near the two walls tend to stick to the solid surface and maintain a zero relative velocity (this is the previously mentioned no-slip condition).

Fig.2-6. Velocity distribution between two parallel plates, caused by the motion of the upper plate. The

lower plate is stationary, and the upper one is moved by the force F at a constant speed V∞.

The magnitude of the shear force F can be connected to the speed of the upper plate and to the

viscosity of the fluid by the relation:

VF???h A As an example, assume that the upper plate with an area of A=1m2 is being pulled at a speed of

5m/s. The fluid between the two surface is water, and the separation distance is 0.02m. Taking the value of the viscosity coefficient μ from Table 2.1 we can calculate the force F required to pull the plate as:

5?V?F?????A?1.0?10?3??1?0.25N0.02?h?

- 4 -

The Reynolds number:

The Re number represents the ratio between inertial and viscous (friction) forces created in the air and is defined by the following formula:

Re??VL?

Here ρ(pronounced rho) is the fluid density, μ is the viscosity, V represents the velocity, and L is some characteristic length (of the velocity, for example ).

An important feature of this number is that it is nondimensional, that is, the units cancel out (even if we use British, USA, or European units). For a typical numerical value of the Reynolds number we can assume a car length of 4m and a speed of 30m/sec, and use the properties of air from Table 2.1, thus;

ReL?1.22?30?4/?1.8?10?5??8.1?106

For example, for Reynolds number values(base on the car length) of less than 105 the flow over

wings will be laminar and the drag and the lift obtained at this rang may be considerably different than at the higher values of the Reynolds number. Returning to the case of a race car piercing its way through air(with very small viscosity) we find that the Reynolds number will be on the order of several millions. But if the same vehicle moves through a highly viscous fluid such as motor oil then the Reynolds number will be far less.

雷诺数代表流场中物体所受的惯性力与粘性力的比。因此，Re数越小的流动，粘性作用越大（相对

于惯性力来说）；Re数越大的流动，粘性作用越小。

Another interesting feature of the Reynolds number is that two different flows can be considered

similar if their Reynolds number are the same. A possible implementation of this principle may apply when exchanging water tunnel for wind tunnel testing, or vice versa. Typical gains are in reduced model size, or in lower test speeds. Foe example, the ratio of viscosity/density in air is about 15 times larger than in water; therefore, in a water tow tank much slower speeds can be used to test the model at the same Reynolds number(and this has been done but seems not be very practical for automobile testing). A more practical application of this principle would be to test a 1/15-scale submarine model in a wind tunnel at true water-speed condition. Usually it is better to increase the speed in the wind tunnel and then even a smaller scale model can be tested. Boundary layer:

The concept of a boundary layer can be described by considering the flow past a two-dimensional

flat plate submerged in a uniform stream, similar to the one shown in Fig.2-4.

This layer of rapid change in the tangential velocity(shown schematically by the velocity profile

in Fig.2-4) is called the boundary layer, and its thickness δ(delta) increases with the distance along the plate. The boundary layer exists on more complicated shapes, as well,(e.g., the automobile shown in Fig.2-7).A typical velocity profile within this layer is described by the inset on this figure.

The thickness of this boundary layer is only several mm at the front of a car traveling at 100

km/hr, and can be several cm thick toward the back of a streamlined car. As you will see, a thicker boundary layer creates more viscous friction drag. Furthermore, a too steep increase in this thickness can lead to flow separation, resulting in additional drag and a loss in the downforce created by a race car’s wings.

- 5 -

Sources of Drag and Lift:

总的气动阻力 外部阻力 内部阻力 形状阻力 诱导阻力 发动机冷却系 阻力 驾驶室内空调 阻力 汽车部件冷却 阻力 压差阻力和表面摩擦阻力(skin friction)的本质来自于气流的粘性。

- 16 -

Wind Tunnel Methods:

Fig.3-12. Schematic description of a basic open-return wind tunnel.

Fig.3-13. Plan view of General Motor’s closed return wind tunnel in Detroit.

Fig.3-15. Open-return wind tunnel with an open test section (Eiffel type).

- 17 -

Fig.3-14. A closed return wind tunnel with an open test section, Gottingen type.

Fig.3-20. Streamlines near a body in an open free stream(A), and when constrained by two rigid walls(B).

- 18 -

Let us start by demonstrating the dilemma faced when trying to match model and wind tunnel sizes by the simple example presented in Fig.3-20. At the top of the figure a streamlined body moves through open air, causing the nearby streamlines to deform due to its presence. This disturbance in the flow is local; far from the body the streamlines will not be affected but stay parallel and straight. If this body is placed within two walls, as shown by the lower part of the figure, then the confinement forces the nearby streamlines to adjust to the gap between the model and the walls, creating larger lift and drag readings. So the first part of the dilemma is that the largest possible test section is desirable to reduce the effects of the wall in an effort to obtain results closer to open-road conditions. Example for Wind Tunnel Correction:

Wind tunnel wall corrections are used to modify the data so that it will be closer to the open air condition. Most of these wall correction methods are based on the ratio between the model frontal area A and the wind-tunnel test-section (or open jet) cross-section area C. For example, one of the simplest formulas for the blockage correction in a closed test section is :

?12?2??V???2?c?1A???1????12??4C??V????2?m

12?V?Here the correction is applied to the dynamic pressure 2 and the subscript c stands for corrected

and m for measured, respectively. This correction can be applied to any aerodynamic coefficient and, for example, when applied to the lift coefficient we can write that

????1?CLc?CLm?2??1A????1??????4C??

To demonstrate the principle of this equation let us assume a large blockage ratio of A/C=0.075（7.5%）and assume that we have measured a lift coefficient of

CLc1 = 0.300 in the wind tunnel. The corrected lift

coefficient value (estimated for the road), based on the correction, is:

CLc?0.300?

?1??1??0.075??4?2?0.289

In general, open-jet test sections are less sensitive to blockage corrections, and the magnitude of these corrections can be as low as 1/4 of the equivalent closed test-section corrections. Simulation of Moving Ground:

The need to simulate a moving ground (or road) in the wind tunnel considerably complicates wind tunnel testing. Before listing the various solution to the problem, let us prove first that there is a problem. This can be demonstrated by observing Fig.2-23. As you can see, there is a difference in the shape of the boundary layers between the on-the-road and in-the-wind-tunnel conditions. Now recall the fact that in the boundary layer the airspeed near the surface of a stationary object slows down to zero. A closer look at the velocity profile between the car and the road reveals a velocity deficiency near the vehicle’s surface, as

- 19 -

shown within the circle.

Fig.2-23. Generic shape of the boundary layer for a vehicle moving on the road(A), and for a vehicle

mounted in a wind tunnel with fixed walls and floor(B).

In the wind tunnel, the tunnel floor is a stationary object relative to the air flow, and even without a car in the test section a boundary layer exists on the floor. This is shown by the second circular insert which sample the velocity profile ahead of the model in the wind tunnel. When a car is placed in the wind tunnel, the velocity profile under the vehicle (right-hand circle) is the result of the two boundary layers, one formed on the ground and one on the vehicle’s lower surface.

The main questions are: How thick are those boundary layers, and how large an effect do they have on the aerodynamic result? Fig.3-24 presents measured boundary layer thickness values in the full-scale GM tunnel. This data indicates that even when applying boundary layer suction ahead of the model, boundary

- 20 -

106.. This is possibly a result of a rapid change in the location of the separation lines on the wheel. When the Re number increases across this rang, the separation point (or line) behind the wheel, shown in the inset in this figure, moves backward, and the size of the separated flow region is reduced (from the top and the two side views). It is quite remarkable that such a small Reynolds-number increment has this strong effect on the lift and drag of the wheel-----the lift by more than 50%. Conclusions Wind Tunnel Methods:

While it may seems if wind tunnels can never exactly simulate actual road condition (due to ground erect, wheel rotation, Reynolds number, etc.), the wind tunnel is, in facet, the primary tool used to study automotive and race car aerodynamics.

In regard to the accuracy of the data, many wind tunnel operators proudly advertize the microscopic accuracy their balance system, but in reality some of the problem mentioned earlier refute this claim of high fidelity. And measuring lift coefficients with more than 2 digits behind the decimal point may not be necessary. The bottom line is :

1、 Understanding the aerodynamic problem is more important than having too sensitive equipment. Since

vehicle improvement requires only incremental data (to judge if an idea is good), productive vehicle improvements can be achieved with minimum resource (good results have been achieved with 14% blockage in a wind tunnel with no moving ground belt)

2、 Whatever works satisfactorily in the wind tunnel, will usually work on the track (or on the road) 3、 A design optimized in a small-scale wind tunnel will be too conservative on the actual road, and the

vehicle can be further improved.

- 26 -

本文来源：https://www.bwwdw.com/article/kr5v.html