variavel0=<b>E. R. Filletti</b> - filletti@sc.usp.br USP
<b>P. SELEGHIM JR.</b> - seleghim@sc.usp.br USP
<b>Abstract.</b> Thermal tomography of materials and flow processes is of fundamental importance to the optimization of industrial applications, particularly in the present context of continuum demand for increasing performances, minimization of energy consumption, pollutant emission reduction and so on. In fact, the number of scientific and technical dedicated is undoubtedly growing, although many phenomenological questions are still open. One of them concerns the possibility of interrogating the studied medium with the help of an acoustic field produced on its boundaries and of reconstructing the internal temperature field from the corresponding boundary responses, i.e. time-of-flight and/or intensities distributions. The method is based on the fact that, for a medium with known physical properties, the local velocity of sound depends on the internal temperature distribution in a way that, mathematically, it is possibly possible to reconstruct a one-to-one relation between both distributions. Even though possible, the relation between boundary measurements and the internal propagation velocity distribution (which maps the temperature distribution) is inverse and, consequently, the problem of reconstruction the last from the former is intrinsically ill conditioned. Concretely, this may so that the reconstructed procedure is extremely sensitive to experimental errors in the boundary measurements and, to obtain an acceptable reconstruction, it would be necessary to measure with an unrealistic degree of accuracy. The precise determination of the excitation and measurement conditions in which one incurs in such sensitivity problems constitutes the main objective of this work. This is done from numerical simulations of the modified Helmholtz equation, representing a two-dimensional domain within which different sensing conditions are reproduced. The sensitivity was assessed by varying the position of a test inclusion (a region in which the sound propagation velocity is different) and by computing the corresponding variation in sound intensities at the boundary.
<b>Keywords.</b> thermal tomography, thermography, acoustic sensing, inverse problem, sensitivity.