Methods for surface roughness elimination from thermal-wave frequency scans in thermally inhomogeneous solids

Two approaches for eliminating surface roughness in the thermal-wave freque ncy response of inhomogeneous solids are developed. The first approach is b ased on the theoretical formulation of roughness as an effective homogeneou s overlayer and is adequate for eliminating low roughness levels from exper imental data. The second approach models roughness as random spatial white noise resulting in a linear superposition of logarithmic-Gaussian distribut ions representing roughness scales in the spatial frequency spectrum and in the modulation frequency domain. Two scales of roughness on the surface of hardened AISI 8620 steel with the same hardness depth profiles are found a nd the experimental data are reconstructed to retrieve similar inhomogeneou s thermal diffusivity depth profiles. (C) 2001 American Institute of Physic s.