RT Journal Article ID 059b4e8169e43929 A1 Benech, J. A1 Freton, P. A1 Gonzalez, J. J. A1 Franceries, X. T1 EXPERIMENTAL DETERMINATION OF THERMAL PLASMA TEMPERATURES JF High Temperature Material Processes: An International Quarterly of High-Technology Plasma Processes JO HTM YR 2009 FD 2009-08-26 VO 13 IS 1 SP 25 OP 44 K1 3D K1 non axisymmetry K1 plasma characterization K1 Boltzmann diagram K1 Abel inversion K1 tomography AB This work is related to a theoretical study on the experimental determination of temperatures on thermal plasma medium. Three different techniques have been chosen and applied to a specific configuration: a 3D model of a transferred arc in argon; this arc is deflected by a magnetic field providing a non axisymmetric shape. Two of the chosen methods: the Boltzmann diagram and the Abel inversion are commonly used in thermal plasma domain. These methods are based on constraining assumptions on the symmetry of the studied object: plasma homogeneity along the considered direction for Boltzmann diagram and cylindrical symmetry for Abel inversion. The third considered method is named tomography and is very developed in medical imaging domain. Tomography is a technique which does not assume symmetry of the studied object, and consequently, it naturally appears to be adapted for 3D emissivity reconstruction. In this paper, tomographic reconstruction of the plasma temperature is performed with a MART (Multiplicative Algebraic Reconstruction technique) algorithm, with four projections located every 45°. However, in the whole paper, only two perpendicular directions of acquisition are simulated. Finally, the temperature profiles reconstructed by the means of each method are compared. It has been demonstrated that Boltzmann diagram and Abel inversion can lead to reconstructed profiles with important errors, as well as quantitatively than qualitatively. On the contrary, tomography with the MART algorithm gives very satisfactory results; the non axisymmetric shape of the plasma is totally respected and a relative error of about 2% - about 300K for a temperature around 15500K- is committed on the reconstructed temperature values. Therefore, the ability of tomographic reconstruction in 3D plasma characterization has been demonstrated and moreover, it could be used for 3D model validation. PB Begell House LK https://www.dl.begellhouse.com/journals/57d172397126f956,5eb5b830375bbc0c,059b4e8169e43929.html