This paper documents the application of a finite element model for impedance eduction by an inverse method in a flow duct facility. The duct geometry is square, and the frequency range allows only propagation of plane waves in unlined sections of the duct. Acoustic lining test panels are inserted on the top wall, with the bottom wall and side walls unlined. The propagation model is two-dimensional, allowing no variation of the acoustic field between the side walls. Impedance eduction from measured acoustic transfer functions on the bottom wall of the duct is based on an iterative process in which impedance of the sample lining, duct termination impedance, and mean flow Mach number are set and the acoustic field is predicted. A search algorithm varying 5 degrees of freedom (real and imaginary parts of the two impedances and Mach number) is used to best fit the predicted acoustic field to the measured acoustic field. With Mach number as a search parameter, substantially improved data fitting is achieved as compared with the use of an a priori assumption of effective Mach number presumed to account for real flow effects. Published NASA Langley Research Center grazing incidence tube data are used in benchmark examples. Data from the Spirit AeroSystems flow duct facility are used to emphasize the quality of results achieved.