Experimental Application of an Active Control Loop on Backward-Facing Step Flow.

Abstract

Experimental results are presented on the active control of a backward-facing step flow revealing some features observed in combustion chambers. These experiments were performed under nonreactive conditions, and the effect of the pressure fluctuations induced by the unsteady heat release was simulated by an external acoustic excitation produced by two loudspeakers placed far downstream from the step. This excitation signal was delivered by a generator or by a hot-wire probe placed in the shear layer. The latter configuration reproduces some aspects of the actual coupling that have been observed with combustion instabilities. This preexcited flow was then controlled using six synthetic jets located on the vertical side of the step. The control loop uses an recursive least-mean-square autoadaptive algorithm applied to a signal from either a microphone or a hot-wire probe placed downstream of the step edge. The influence of the external acoustic excitation on the flow organization and the effect of the control loop are described. Without excitation, the control algorithm is not able to reduce the natural unsteady phenomena. However, it does damp the shear-layer pressure or velocity fluctuations intensified by an external excitation and bring the flow close to its natural feature. The efficiency of the control depends on the frequency of the external excitation, the signal used in the control loop, and the phase shift between the velocity and the pressure signals in the shear layer. NomenclatureA e = input voltage for the external excitation, V Fr = efficiency parameter f d = natural roll-up frequency of the shear layer, Hz f c = external excitation frequency, Hz G energy = gain parameter H m− f ( f ) = transfer function between pressure and velocity signals, Pa · s/m h = step height, m P excitation ( f ) = autospectrum of the pressure signal measured with the external excitation, Pa P total ( f ) = autospectrum of the pressure signal measured with the flow submitted to an external excitation, Pa RMS 1 = rms of the velocity or pressure signal measured without flow, Pa or m/s RMS 2 = rms of the velocity or pressure signal measured with the flow, with the external excitation switched on and with the control loop, Pa or m/s RMS 3 = rms of the velocity or pressure signal measured with the flow with both the external excitation and the control loop switched off, Pa or m/s RMS 4 = rms of the velocity or pressure signal measured in the shear, Pa or m/s Sr h = Strouhal number calculated from the step height Sr θ = Strouhal number calculated from the momentum thickness U e = freestream velocity, m/s V ( f ) = autospectrum of the velocity signal measured in the shear layer, m/s

Transverse combustion instabilities: Acoustic, fluid mechanic, and flame processes

“…With this background, the rest of this section reviews studies that have analyzed the forced response of flows to transverse (or non-axisymmetric) forcing. Studies of shear layer instability and forced response include forcing the flow from the upstream plenum [224,225], by mechanical actuation [226,227], or transversely deflecting the shear layer [228][229][230][231][232][233][234]. For example, Fiedler and Mensing [233] forced the shear layer transversely by exciting a waveguide that opened at a slit at the separation point of the mixing layer.…”

Reducing the pressure drag of a D-shaped bluff body using linear feedback control

“…For a BFS whose separating boundary layer has 3-D flow fluctuations at sufficiently high Reynolds number, the wake flow exhibits global instability and is an "oscillator" flow; most previous studies agree on the existence of a shedding/step mode and a lower frequency flapping or bubble pumping mode [13][14][15]. Previous numerical [15,16] and experimental [8,[17][18][19] feedback control studies have been performed; with the exception of [15], the sensor signal has been either based on wake velocity measurements or a measure of the reattachment length.…”

Effects of flow control on flow past bluff bodies

“…The step mode based on the step height h, and defined by Strouhal number, S rh = f h U ⁄ characterizes the second instability mode. He observed these values to be around Srθ = 0.012 and Srh = 0.185 respectively, these values match with the values obtained lately by Marrot et al (2005), Sth =0.14 and Dejoan et al (2005), Sth =0.2. In Vukasinovic et al ( 2010), active flow control in a backwards facing step under turbulent conditions was studied, it was reported that a slight waviness of energy in the boundary layer results in a low scale periodic vortex train.…”

“…When studying boundary layer performance in low-pressure turbines with AFC, Sondergaad et al (2002) and , observed that unsteady suction was more effective that steady one, and AFC jets located perpendicular to the surface were more effective than tangential ones. Marrot et al (2005) used a loudspeaker located downstream of the step to activate the boundary layer, the maximum effect was observed when the loudspeaker frequency was equal to the boundary layer instability mode, at these conditions a small reduction of the recirculation length was obtained. The pre-excited flow was controlled using two rows of synthetic jets located on the vertical side of the step.…”

“…A similar method, consisting of two arrays of jets located on the vertical side of the step of which the upper row steadily absorbs and the lower row steadily injects, was analyzed numerically in 2D by Creusé et al (2009), in fact different AFC techniques including pulsating flow were evaluated. In Dahan et al (2012), and Marrot et al (2005), open loop and closed loop control were studied. For open loop control, the optimum actuation frequency, as already found by many other researchers, coincided with the fundamental frequency, which was Sth =0.14, the optimum amplitude being A=0.2, this was the same amplitude used by Dahan et al (2012).…”