RESEARCH ARTICLE A sensitivity analysis of the point reference global correlation (PRGC) technique for spatio-temporal correlations in turbulent flows F. Kerherve ´ Æ J. Fitzpatrick Æ Ludovic Chatellier Received: 18 December 2006 / Revised: 28 September 2007 / Accepted: 30 September 2007 / Published online: 13 November 2007 Ó Springer-Verlag 2007 Abstract The point reference global correlation (PRGC) technique which combines single and global measurements as proposed by Chatellier and Fitzpatrick (Exp Fluids 38(5):563–757, 2005) is of significant interest for the analysis of the turbulent statistics for noise source modeling in jet flows as it allows the 2D spatio-temporal correlation functions to be obtained over a region of the flow. This enables the statistical characteristics including inhomoge- neous and anisotropic features to be determined. The sensitivity of the technique is examined in some detail for the specific case of laser doppler velocimetry (LDV) and particle image velocimetry (PIV). Simulated data are used to enable a parametric study of the accuracy of the PRGC technique to be determined as a function of the various measurement parameters. The sample frequencies and the number of samples of both the LDV and PIV signals are shown to be critical to errors associated with the estimated spatio-temporal correlations and that low data rates can lead to significant errors in the estimates. Measurements per- formed in single stream and co-axial jet flows at Mach 0.24 using PIV and LDV systems are reported and the 2D space– time correlation functions for these flows are determined using the PRGC technique. The results are discussed in the context of noise source modeling for jet flows. 1 Introduction The turbulence statistics of length and time scales are important for aeroacoustics as these quantities are directly related to the scales at which turbulent kinetic energy generates noise. For aeroacoustic source modeling, the acoustic analogy approach is based on modeling the tur- bulence correlation tensor which represents the conversion mechanisms which produce acoustic energy. While sig- nificant progress has been made using CFD tools for these modeling space–time properties, it still remains necessary to measure these properties using experiments. Two-point measurements combining intrusive or optical techniques have been extensively used to extract the typical turbulence statistics required for modeling the correlation tensor as reported by, for example, Chu (1956) and Harper-Bourne (2003). However, these measurements allow only one component to be estimated in one direction for each series of measurements. Obtaining a 3D model using such measurements is not realistic and remains a significant challenge. Although particle image velocimetry (PIV) measure- ments can be used to extract the spatial correlation over a 2D area, the relatively low repetition rate means that temporal information such as time scales and convection speeds cannot be obtained easily. Using low repetition rate global measurement techniques to extract such quantities require restrictive assumptions based on the Taylor frozen- flow approximation (Batchelor 1953). This approximation is only applicable to turbulent jets for well defined coherent F. Kerherve ´(&)  J. Fitzpatrick Mechanical and Manufacturing Engineering Department, Trinity College-Dublin, Dublin, Ireland e-mail: franck.kerherve@yahoo.fr J. Fitzpatrick e-mail: john.fitzpatrick@tcd.ie L. Chatellier Laboratoire d’Etudes Ae ´rodynamiques, UMR-CNRS 6609, Poitiers, France e-mail: ludovic.chatellier@lea.univ-poitiers.fr 123 Exp Fluids (2008) 44:577–589 DOI 10.1007/s00348-007-0417-9