INVESTIGATION OF THE MIXING LAYER IN A SLIGHTLY UNDEREXPANDED SUPERSONIC JET BY PARTICLE IMAGE VELOCIMETRY

Abstract

A slightly underexpanded supersonic jet at a Mach number M j of 1.10 is studied experimentally. Schlieren visualizations and particle image velocimetry are applied in order to characterise the shock-cell structure and turbulence in the mixing layer, which are the two elements at the origin of the shock-associated noise emitted by such a jet. It is found in particular that the velocity gradients typical of the shock-cell structure still exist in the subsonic part of the mixing layer. From the evaluation of some turbulence properties (turbulence level, momentum thickness and spatial correlation), it is shown that the jet behaves very similarly to jets at high subsonic Mach numbers. It is believed that such data could shed light on the shock-associated noise source. INTRODUCTION The commercial aircraft powered by turbofan engines exhaust slightly underexpanded supersonic jets at cruise conditions, characterised by the presence of a shock-cell pattern in the jet plume. The interaction in the mixing layer between the turbulence and the shock-cell system is responsible for the so-called shock-associated noise component of jet noise, which is in addition to the ever present turbulent mixing noise. Shock-associated noise is made up of two distinct parts : a tonal one, referred to as screech, and a broadband one (the broadband shock-associated noise). While extensive static pressure measurements have been performed to characterise the shock-cell structure (e.g., Norum & Seiner (1982)), detailed accounts of the turbulence in imperfectly expanded jets are scarce. Seiner & Norum (1980) measured turbulence levels and spectra using a hot film probe. Panda & Seasholtz (1999) obtained the coherent part of the density fluctuation in choked jets using the Rayleigh scattering technique and related this to the screeching process. Several studies applied particle image velocimetry (PIV) to these flows. Alkislar et al. (2003) separated the random from the coherent turbulent motion in the mixing layer of a screeching rectangular jet using stereoscopic PIV, and pinpointed the relation between coherent vortices and screech generation. Bridges & Wernet (2008) applied high-speed PIV to screeching and non-screeching supersonic jets, mainly focusing on turbulence spectra. The objective of the present experimental study is to focus on some properties of the turbulence and the shockcell structure in the mixing layer of a slightly underexpanded jet, with the ultimate goal of clarifying the broadband shock-associated noise generation process. Particle image velocimetry is thus applied to a jet at an ideally expanded Mach number M j = 1.10. To begin with, the strength of the shock-cell structure in the mixing layer is estimated. Then, a study of the turbulence in this jet is reported. It addresses the turbulence levels, the mixing layer thickness, the location of the sonic line and the spatial correlations. EXPERIMENTAL METHODS The facility employed in the present work has already been used to study single-stream supersonic jets (Andre et al. (2013b)) as well as co-axial jets (Andre et al. (2011)). The configuration considered here is the latter one, with the outer stream set at a Mach number of 0.05 to seed the surroundings of the inner, supersonic jet, during the PIV measurements. The supersonic jet flow originates from a continuously operating compressor mounted upstream of an air drier. It exhausts through a round, contoured and convergent nozzle of diameter D = 38.7 mm. Since the underexpanded jets exiting typical turbofan engines of civil aircraft do not seem to emit screech, it appears relevant to eliminate it in the smallscale study. For that purpose, a screech-suppressing nozzle is employed: shallow notches are cut into its lip. As indicated in a previous study (Andre et al. (2013a)), this nozzle non-intrusively suppresses screech. The reservoir temperature Tt is measured upstream of the exit. Here, the jets are unheated and Tt ≈ 30C. The nozzle pressure ratio (NPR), defined as the ratio between jet stagnation pressure and ambient pressure, is set by measuring the wall static pressure fifteen nozzle diameters upstream of the exit. A conventional Z-type schlieren system is used to visualise the flow. It consists of a light-emitting diode, two 203.2 mm-diameter f/8 parabolic mirrors, a straight knifeedge set perpendicular to the flow direction and a highspeed Phantom V12 CMOS camera. Particle image velocimetry has also been applied to measure the velocity in a plane containing the jet axis and a notch. Illumination is provided by a pulsed double-cavity Nd:YLF Quantronix Darwin Duo laser. The sheet thickness is 1.7 mm± 0.3 mm. The supersonic jet is seeded with olive oil by means of custom-designed Laskin nozzle generators. The mean particle size is known to be around 1 μm. The secondary flow is seeded by smoke. Both seeding devices are mounted far enough upstream of the exit so that the particle concentration in each flow is approximately uniform.