Introduction to Microramp VGs

Owing to the detrimental effects of large shock-induced boundary layer separations discussed in the preceding chapters, it is often desireable to control/reduce SBLI separation size. In mixed-compression inlets [79], this reduction is most often achieved by bleeding away the lower- momentum portions of the incoming boundary layer. While effective, the bled mass can repre- sent a significant portion of the total engine flux (up to 20%) [80], and inlet frontal area must be increased in order to recover the lost throughput. This increase in size is accompanied by additional drag penalties, reducing vehicle range. Therefore, any suppression of shock-induced separation through means other than wall bleed is highly desirable for the purposes of im- proving vehicle efficiency. Sub-boundary layer thickness VGs were developed in response to

this need. The shapes and sizes of these devices vary, with some of the most common config- urations being the microramp [81], tapered/untapered microvanes [81], the robust-vane [78], and triangular/rectangular split-vanes [82]. The details regarding the flow features of many of these device types and their effectiveness in suppressing SBLI separation are captured in sev- eral recent reviews [83–85]. Of these VGs, one of the most extensively utilized/examined is the microramp-type device.

Several studies have been performed on the viability of shock-induced separation reduc- tion using microramp VGs in planar configurations, following the sizing proposed by Ander- son et al. [81]. Babinsky et al. [86] tested the effect of multiple device heights (ranging from h = 0.3δ0 −1.0δ0) in single and multiple VG arrangements on the separation induced by an

impinging/reflected oblique SBLI along the wind tunnel floor at Mach 2.5. Using laser Doppler anemometry, they effectively demonstrated the ejection of a low-momentum fluid region from the boundary layer by the upwash generated by a counter-rotating primary vortex pair. These vortices were also observed to mix high-momentum fluid into the near-wall flow, effectively energizing the boundary layer. The rate of low-momentum fluid ejection/high-momentum en- trainment was found to scale with device height. Ultimately, little effect of streamwise device location or device height was observed, with all devices decreasing the separation length scale and reducing the upstream influence of the interaction at the centerline (i.e. directly behind the devices). While the tallest device (h = 1.0δ0) was observed to have the strongest effect,

it also incurred the largest drag penalty. The smallest device height of h = 0.3δ0 was found

to be almost as beneficial but with reduced drag, suggesting smaller microramps are advan- tageous practically. However, smaller devices must be placed far closer to the interaction due to the flowfield downstream of the VG scaling with device height, as mentioned previously. Using RANS computations, Ghoshet al.[87] simulated the experiments of Babinskyet al.[86]. They observed that the primary effect of the microramp VGs on the SBLI was to deform the separation, rather than to significantly reduce its size. While the flow directly behind the micro- ramps did exhibit reduced upstream influence from the uncontrolled interaction, the upstream

influence grew in the regions of the flow that were not energized; these competing effects led to minimal integral improvements. Giepmanet al. [88] performed two-component particle im- age velocimetry (PIV) experiments for multiple microramp heights and placements upstream of a planar impinging/reflected SBLI at Mach 2.0. In agreement with Ghosh et al. [87], they found that while the total (spanwise-integrated) separated area decreased with VG control, the uncontrolled regions experienced local separation lengths larger than a clean configuration without microramps.

Experiments and simulations have also focused on the effects of microramp devices on planar compression ramp SBLI. Manisankar et al. [89] placed arrays of VGs upstream of an α = 24° compression ramp on a flat plate at Mach 2.05. They observed a corrugated separation line, in agreement with the local decreases in upstream influence noted behind the devices in the im- pinging/reflected case [86–88]. Wall pressure measurements indicated minimal differences in the upstream influence/separation regions from the uncontrolled interaction [89]. Yanet al.[90] per- formed LES of a single microramp VG upstream of a compression ramp, and noted a significant reduction in separation length across the interaction. They also revealed that the momentum deficit behind the device was not a viscous wake, but rather the lower momentum portion of the incoming boundary layer which was “scooped” away from the wall by the microramp. It was further shown that turbulent vortex rings were generated by the Kelvin-Helmholtz instability in the shear layer surrounding this momentum deficit, confirming earlier experimental findings by Sun et al.[91]. These rings enhanced turbulent mixing, further contributing to the momentum recharge of the boundary layer and strongly impacting the separation shock front.

5.2

Motivation

It is interesting to note that in each of the aforementioned VG studies, the interaction of interest occurs on a flat (planar) body. In designing the inlet of high speed air-breathing vehicles, an axisymmetric geometry (e.g. Busemann type) is often favored due to its superior total pressure recovery and relative lack of detrimental 3-D/corner effects. The preceding chapter showed that

axisymmetric SBLIs are subject to different confinement effects from what have been observed in the rectangular literature. Specificially, centerline shock/shock interactions have been seen to lead to a more complex outer (inviscid) shock structure. Further, geometric area contraction was found to impose a greater blockage on the core flow than in a planar geometry; the separated flow of the H-I configuration has been shown to be highly sensitive to these mechanisms. Owing to its common usage in planar configurations, the aim of this chapter is to present and determine the effectiveness of a logical extension of the microramp VG geometry of Andersonet al.[81] in an axisymmetric configuration. The VG flowfield is first characterized through a thorough analysis of its surface flow and off-body features (outer shock structure). An experimental investigation is then performed into single VG-controlled and array-controlled H-I compression ramp SBLIs.

5.3

General Experimental Description

The H-I SBLI case with α = 20° and h/R = 0.246 was selected for the axisymmetric VG investigations [64, 65]. The model geometry is described in Sec. 2.2.2, the ramp parameters are provided in Table 5.1. A general description of the features of the baseline (uncontrolled) H-I compression ramp interaction can be found in Chapter 4. Of particular importance is an ascertained compression strength (back pressure ratio) of pb/p∞= 4.0, which is well in excess

of the 2-D weak solution pressure rise for an α = 20° deflection. Details regarding the VG geometry and placement can be found in the subsequent section. The experimental methods utilized to study the VG flowfield and VG-controlled SBLIs included unsteady SSVs, high- frequency WSP transducer measurements, pitot boundary layer measurements, PLS imaging, and mean PSP imaging. More information on each of these methods can be found in Secs. 2.3.1-2.3.5. The WSP transducer data presented in this chapter were hardware low-pass filtered using a cutoff of f = 50 kHz and were subsequently software low-pass filtered at f = 25 kHz for the computation of RMS pressure values.

Table 5.1: H-I compression ramp configuration selected for the microramp VG study. True, PLS-discerned back pressure ratio and percentage of Korkegi limit for 2-D incipient separation also shown.

α h (mm) h/R h/δ0 pb/p∞ (actual, see Chapter 4) % 2-D KL

20° 9.24 0.246 2.64 4.0 139