Experimental investigation of a diffuser for use in skydiving vertical wind tunnel

Highlights

• A diffuser for a vertical wind tunnel is investigated in suction and blower modes.

• Velocity profile, pressure loss, and turbulence intensity have been studied.

• Pressure loss coefficient in the suction mode flow is about 10% less than the blowing mode.

• Effect of non-uniform velocity is remarkable in pressure loss measurement.

• No severe separation occurred in the flow inside the diffuser for both modes.

Abstract

Diffusers are vital components of vertical wind tunnels. In this experimental study, the velocity distribution, pressure recovery coefficient, turbulence intensity, and flow separation of a 1:13 model of a diffuser with an octagonal inlet and square outlet for use in the skydiving vertical wind tunnel have been investigated. The experiments were carried out both in suction and blowing flow modes. Results showed no significant pressure loss and separation in the diffuser under study. Also, the pressure loss results were compared with the results obtained using semi-empirical correlations for a diffuser with a circular inlet and square outlet, which is geometrically the closest to the diffuser under consideration in the present work. It was observed that to correctly predict the pressure loss in the model diffuser using the semi-empirical correlations, the diffuser big-angle plane of 10.66° should be used in these correlations instead of the equivalent cone angle. Besides, if the effect of the non-uniformity of the velocity in the diffuser is not considered in the determination of the pressure loss, a significant error will incur in determining the pressure loss coefficient.

Introduction

Diffusers are important components in fluid systems such as wind tunnels. The main role of a diffuser is to reduce the dynamic pressure and convert it to static pressure, resulting in a reduced loss in the downstream flow. Losses in diffusers occur in the form of friction and expansion losses; the former is caused by diffuser length and the latter by variations in the pressure gradient [1], [2].

Eckert et al. [3] offered simple methods for the aerodynamic design and prediction of the decrease in the pressure loss in various components of a subsonic wind tunnel. They suggested semi-empirical correlations for predicting the pressure loss coefficient for diffusers having outlet and inlet configurations of circle-circle, square-square, and circle-square.

Sanborn and Kline [4] showed that separation is defined considering two parameters, namely shape factor, and pressure gradient; each parameter predicting the separation point for external flows with almost equal accuracy.

Von Den Hoff and Tetervin [5] used the shape factor technique in a study as a method for determining the boundary layer separation. They demonstrated that when the boundary layer flow is laminar, the shape factor for flow separation inception is 3.5, whereas in the case of a turbulent boundary layer flow the shape factor value is 2.4. Reneau et al. [6] studied four types of flow regimes generated in two-dimensional and straight diffusers and explored their performances in two modes of skewness and kurtosis. Waitman et al. [7] studied the effects of the flow boundary layer thickness at the inlet of two-dimensional plane-wall diffusers and demonstrated that a thicker boundary layer affects the flow regime, and results in diminished diffuser performance.

Furuya et al. [8] used the boundary layer suction at the diffuser’s inlet as a technique for preventing flow separation, which results in high efficiency in wide-angle conical and two-dimensional diffusers. Rao and Raju [9] employed splitter vanes for flow control in wide-angle conical diffusers with an angle of 38° and an area ratio of 15. These planes divided the diffuser into several sections, thereby decreasing its angle, and ultimately making the airflow consistent at the diffuser’s outlet. Senoo et al. [10] used vortex-generating blades to increase diffuser efficiency. Their experiments on diffusers with divergence angles of 8°, 12°, 20°, and 30° and an area ratio of $4$ showed that vortex-generating blades prevent flow separation up to a divergence angle of 16°, where the pressure recovery coefficient is equal to the pressure recovery coefficient of an 8° diffuser.

Ye et al. [11] studied the effects of divergent angle on the flow behavior in low-speed wind accelerating ducts. Herbst et al. [12] investigated the effect of Reynolds number on flow separation in diffusers using numerical simulation and LES model and showed that increasing the Reynolds number reduces the boundary layer flow separation in a diffuser. In a study by Yang et al. [13], the flow control mechanism and θ design principal in a conical diffuser with an implicit large-eddy simulation were investigated. A diffuser design based on their method can suppress massive flow separation in the expansion section and downstream region, promoting pressure recovery performance in the diffuser. Li et al. [14] simulated a three-dimensional diffuser with a square geometry at different divergence angles to obtain a better understanding of the flow diffusion across the geometry, velocity distribution at the outlet, and reverse flow. Their simulation shows the velocity is substantially reduced to nearly zero at corners of the diffuser with a square cross-section.

Studies have shown that the pressure loss and flow separation in wind tunnel diffusers depends on their type of design and shape. The present study aims to investigate a diffuser for use in skydiving vertical wind tunnel application (refer to Fig. 1).

In the conventional design of diffusers, some of the design parameters can be determined using semi-empirical correlations. However, semi-empirical correlations have been presented for some diffusers such as diffusers with a circular inlet and square outlet [3]. Moreover, since the configuration of the actual diffuser understudy for the skydiving vertical wind tunnel application is that of octagonal inlet and square outlet design, the use of semi-empirical correlations of Ref [3] does not yield accurate results for the design of this type of wind tunnel diffuser. Therefore, the design parameters of the diffuser understudy with the specified configuration must be investigated in more detail. For this purpose, the velocity profile, pressure recovery coefficient, turbulence intensity, and flow separation of a 1:13 model of the diffuser have been investigated in this experimental study. The experiments were carried out both in suction and blowing flow modes using wind tunnels of a comparable scale. Besides, the pressure loss results obtained from the diffuser model were compared with those obtained from the semi-empirical correlations for a diffuser with a circular inlet and square outlet, which is the nearest configuration to the diffuser in the present work.