| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
||
|---|---|---|
Although diffusers are widely used, their flow characteristics are still not fully understood. The flow through a diffuser inevitably depends on its geometry, defined by the area ratio, wall expansion angle, cross-sectional shape and wall contour. Other parameters such as conditions at entry and exit and boundary layer control devices, if any, also affect the diffuser performance. With an arbitrary combination of these parameters, the flow through a diffuser becomes too difficult to predict in detail. The issue is further complicated by the occasional presence of boundary layer separation caused by the adverse pressure gradients necessarily present in diffusers. It was in fact this separation in a conical diffuser that inspired Prandtl and in 1904 the concept of boundary layers emerged.
These features, coupled with the general problems of predicting behavior of flow turbulent flow, make the diffuser one of the least understood fluid flow devices. Almost all the knowledge acquired about diffusers is empirical.
The uniformity of flow in a wind tunnel can be greatly improved if a contraction section or nozzle of large area ratio is proved immediately upstream of the working section. If a large diffuser area ratio is required, so as to provide a large contraction ratio, a long conventional diffuser, with an equivalent cone angle of 5°, is undesirable economically. For instance, if a contraction ratio of 12 is required, a 5° diffuser with a length of 20 times the working-section diameter would be needed. A short diffuser, with a large area ratio and, consequently, a large equivalent cone angle, is then installed as shown in (Fig. 2). This is a “wide-angle” diffuser, defined as a diffuser in which the cross-sectional area increases so rapidly that separation can be avoided only by using boundary layer control. A wide-angle diffuser should be regarded as a means of reducing the length of a diffuser of given area ratio, rather than a device to effect pressure recovery (the dynamic pressure ahead of the settling chamber is so small compared to that in the test section that pressure gains or losses are almost negligible). Uniformity and steadiness of the flow at the diffuser exit are of prime concern since this affects the performance of vital components downstream. In particular, intermittent separation in the wide-angle diffuser would almost certainly result in an unacceptable level of flow unsteadiness in the working section.
One of the most popular means of preventing separation in wide-angle diffusers is by introducing screens of perforated metal, or more usually, woven wire gauze. For the smaller (blower) tunnels, the cheaper plastic (nylon or polyester) mesh screens may be strong enough. A screen is usually specified by the “mesh,” defined as the number of openings per lineal inch (1/L), and by the wire diameter (d). Given these two specifications, the screen open-area ratio b, equal to (1 – d/L)˛ for square meshes, can be determined.
Gauze screens make the velocity profile more uniform and thus reduce the boundary layer thickness. The ability of a boundary layer to withstand a given adverse pressure gradient is increased when the boundary layer thickness is decreased. Therefore, a screen, besides removing the direct effects of boundary layer growth and incipient separations, gives the boundary layer a new lease of life so that a further pressure rise can be negotiated without separation. However, the corner flow in a wide-angle diffuser of rectangular cross-section is still liable to separate because of secondary flow effects, but, providing that the separation region remains confined to the corner region, this just “fills-up” the corners and does not present a major problem, especially since the pressure recovery is not of primary importance. Screens also reduce the turbulence intensity level in the whole flowfield and refract the incident flow, towards the local normal, downstream of the screen. These last two factors, however, do not necessarily take the boundary layer further away from separation, although further downstream of the screen, the boundary layer growth rate could be reduced due to lower entrainment from the core flow which now has a lower turbulence level. Note that the tunnel boundary layers often have high free-stream turbulence – even when the flow is not a fully developed duct flow which it sometimes is.
The effectiveness of a conventional woven wire screen depends largely on its pressure drop coefficient (K), which is defined as the ratio of the static pressure drop across the screen ( |
In a wide-angle diffuser, it is better to use several screens, each with a relatively low value of K (say between 1 and 2) and preferably placed at the points where the flow would otherwise separate, than to use a single screen with a large K. This is because increasing K at one station in the diffuser has little effect on the skin friction at a station much further downstream. The most efficient diffuser, needing the fewest screens, is in general one with curved walls, which distributes the extent of the adverse pressure gradient. However, straight-wall diffusers, with discontinuities of wall slope at entry and exit, are easier to build and perform satisfactorily if the equivalent cone angle is not too large. In inviscid flow, a stagnation point occurs at a concave corner and an infinite velocity at a convex corner, and it is only the presence of the boundary layer that smoothes out the flow. Even so, separation is liable to occur at the corner and persist for a long distance downstream. The positioning of screens is also of vital importance, particularly in straight-wall diffusers.
Diffuser behavior is greatly affected by the flow conditions at entry. A relatively thin boundary layer with steady flow conditions at entry would, no doubt, delay separation and improve pressure recovery considerably. Conversely, non-uniform or inclined entry flow can lead to early separation. It has been found that a swirling flow in the wide-angle diffuser (such as that produced by a single-inlet centrifugal blower) decreased the diffuser efficiency, but it can be used to advantage in preventing separation. The effect of inlet conditions on the outlet velocity profile is less in screened diffusers than in any other type of diffuser because of the strong effect of gauze screens on the whole flowfield.
Optimum Design Charts for Wide-Angle Diffusers:The four most important parameters in a wide-angle diffuser are: -
Design rules for wide-angle diffusers using these four parameters are derived by Mehta and Bradshaw "Design rules for small low speed wind tunnels", Aero. Journal (Royal Aeronautical Society), Vol. 73, p. 443 (1979) (the remarks about wide-angle diffusers apply to any size of tunnel). Fig. 7 gives boundaries for the number of screens as a function of the diffuser angle and area ratio. The total pressure drop coefficient required for a given area ratio is given by the boundary in Fig. 8 [Ksum > (A-1)/1.14].
A diffuser comfiguration satifying both these curves should perform successfully provided that certain other design factors are kept in mind:
|
|
p) to the dynamic pressure of the uniform parallel flow approaching the screen (q = ˝ 
U˛). Thus, K = 
, the inlet flow angle. The refractive index of the screen (µ), defined as in optics, is about 0.9 
(1 + K), for small angles of incidence. Allowance for its effect on lateral component velocity modifies simple formulae for the attenuation factor for velocity non-uniformity. The simple formulae suggest that a single screen with K = 2 eliminates small non-uniformities.