No abstract

No abstract

No abstract

Phased array test techniques are becoming increasingly popular in aeroacoustics. This is due, in part, to the use of optimized array designs (e.g., spiral shapes) which can operate over usefully wide frequency ranges with acceptable resolution and sidelobe rejection characteristics with practical numbers of microphones. Sometimes constraints such as limited space for microphone installation prevent the use of optimum array designs, raising the prospect of high sidelobe levels. Several modifications to classical beamforming which may reduce sidelobe levels for imperfect arrays (and possibly improve the performance of optimized arrays) are available. The effectiveness of these techniques is explored by applying three of them to four array designs using simulated wind tunnel data. The three bearnforrning techniques are robust adaptive beamforming, cross spectral matrix element weighting, and the CLEAN algorithm. The array designs are a cross, an array of four crosses, a filled square, and a spiral array. NOMENCLATURE N = Number of microphones y n = Location of microphone n, n=l...N M = Number of point sources x m -location of source m, m = 1...M w(r) = Narrowband array data vector S m (t) = Narrowband source strength C(x) = Array steering vector for point x P(t) -Scattering matrix associated with turbulence in boundary layer W(t] = Gaussian noise vector ( ) = Time average, also assumed to give expected value A = Cross spectral matrix K = Normalization factor for C(x) Pnn> = yn~y* G = Decay constant for boundary/shear layer effect b b f = Beamform map function for classical beamforming b csmw = Beamform map function for beamforming with cross spectral matrix weighting b cl = Beamform map function for beamforming with the CLEAN algorithm b 0 = The highest level in a beamform map b psl = The peak sidelobe level in a beamform map R(n,n'} = Vector spacing redundancy index used in cross spectral matrix weighting v nn , = Weight for CSM element (n,n'} B = Result of weighting the elements of A w = Microphone weighting vector in minimum variance or adaptive beamforming, used like C in classical beamforming a = Loop gain parameter in CLEAN algorithm INTRODUCTION

No abstract