La première partie de ce travail présente une nouvelle méthode qui permet de détermination rapidement de la température et de la concentration en fer d'un plasma d'arc MAG utilisé dans la technologie de soudage par plasma. Le plasma est un mélange [Ar-CO2] en présence de fer, avec un rapport molaire entre Ar et CO2 constant de 82%-18%. La seconde partie s'intéresse au rayonnement d'un plasma d'argon utilisé dans la technologie soudage GTAW et pouvant conduire à des dangers photobiologique. Dans le chapitre 1, le contexte et la motivation de ce travail sont présentés. Les travaux déjà effectués et publiés dans ces deux axes scientifiques sont passés en revue. Dans le chapitre 2, les compositions à l'équilibre sont calculées par la méthode de la minimisation de l'énergie libre de Gibbs. Les densités et fonctions de partitions obtenues pour chaque espèces présente dans le plasma sont ensuite utilisées pour déterminer les pertes radiatives des plasmas [Ar-CO2]-Fe via la méthode du coefficient d'émission net (CEN). Cette estimation des pertes ne peut se faire sans le calcul préalable du coefficient d'absorption spectral que nous avons réalisé finement par la méthode " raie par raie ". Tous les mécanismes radiatifs prédominants dans le plasma ont été pris en compte: continua atomique et moléculaire, raies atomiques et moléculaires. Cette partie constitue la base de cette étude sur laquelle se fonde notre nouvelle méthode de diagnostic destinée à déterminer à la fois la température et la concentration en fer d'un plasma d'arc de soudage. Le chapitre 3 est dédié à l'étude expérimentale d'un plasma d'arc MAG constituée d'une analyse spectroscopique permettant de remonter au profil de température et tester l'hypothèse de l'Équilibre Thermodynamique Local. La température d'excitation est obtenue par la méthode de Boltzmann tandis que la mesure d'élargissement de Stark pour les raies de fer et d'argon permet de remonter à la température et la densité des électrons. / This PhD thesis introduces a method that allows the fast determination of temperature and iron concentration for MAG-P Arc. The MAG-P Arc is in fact [Ar-CO2]-Fe mixtures, with a constant molar ratio between Ar and CO2 [82%Ar-18%CO2]. In a second time, this thesis presents a study of the optical radiation associated to photobiological hazards for argon GTAW Arc. In chapter 1, the background and motivation of this work is introduced. The previous works published in this field are reviewed. In chapter 2, the equilibrium compositions are calculated firstly using the minimization of Gibbs free energy. Then the radiative properties of [Ar-CO2]-Fe plasmas are obtained in the frame of the net emission coefficient (NEC) approach, using the accurate "line by line" method. All significant radiative contribution mechanisms are taken into account in the calculation. This study will constitute a groundwork to build the diagnostic method that allows determination of temperature and iron concentration profiles in welding arc. In chapter 3, spectroscopic investigation of the LTE hypothesis across the MAG-P Arc is made. Excitation temperature is obtained with Boltzmann plot method while iron and argon lines Stark broadening measurements are used to get electron temperature and electron density. LTE hypothesis validity across the arc is discussed considering the agreement between the two temperatures, the electron density and iron content. Results show supporting evidence for the main part of the plasma, along radial and axial directions. Discrepancies occur only at the fringe of the arc, where the two temperatures differ by more than 2000 K. In chapter 4, a method allowing a fast determination of space- and time-resolved plasma temperature and iron concentration in MAG arcs during the high-current phase is introduced. This method consists in measuring the plasma spectral radiation of the arc with iron vapours using a high-speed camera filtered by narrow band filters in the spectral intervals of 570-590 nm and 606-627 nm respectively; calculating theoretically the dependence of the absolute emissivity e570-590 nm and relative emissivity e570-590 nm/e607-627 nm versus the plasma temperature and the iron concentration. This method has also been validated for a layer of plasma by adopting other existing diagnostics such as Stark broadening, which demonstrates the effectiveness of this new method. In chapter 5, a theoretical investigation of the UV (180-400 nm), UVA (315-400 nm) and blue light (300-700 nm) radiation associated with the photobiological hazards to workers for argon GTAW arcs is presented. The radiative properties of argon plasma are calculated for the three spectral regions, and a two-dimensional model of a GTAW arc is then developed to determine the local emissions in the arc, the total radiation escaping from the arc and corresponding effective irradiances. This study clearly supports the importance of undertaking an effective protection strategy for workers, particularly for skin and eyes, in the welding environment. Finally, a general conculsion is given in chapter 6.

The path explored in this thesis is testing a plasma temperature measurement approach that accounts for interference in a spectrum. The Atomic Emission Spectroscopy (AES) technique used is called Laser Induced Breakdown Spectroscopy (LIBS) and involves focusing a laser pulse to a high irradiance onto a sample to induced a plasma. Spectrally analyzing the plasma light provides a "finger print" or spectrum of the sample. Unfortunately, spectral line broadening is a type of interference encountered in a LIBS spectrum because it blends possible ionic or atomic transitions that occur in plasma. To make use of the information or transitions not resolved in a LIBS spectrum, a plasma temperature method is developed. The basic theory of a LIBS plasma, broadening mechanisms, thermal equilibrium and distribution laws, and plasma temperature methods are discussed as background support for the plasma temperature method tested in this thesis. In summary, the plasma temperature method analyzes the Full Width at Half the Maximum (FWHM) of each spectral line for transitions provided from a database and uses them for temperature measurements. The first implementation of the temperature method was for simulated spectra and the results are compared to other conventional temperature measurement techniques. The temporal evolution of experimental spectra are also taken as a function of time to observe if the newly developed temperature technique can perform temporal measurements. Lastly, the temperature method is tested for a simulated, single element spectrum when considering interferences from all the elements provided in an atomic database. From stimulated and experimental spectra analysis to a global database consideration, the advantages and disadvantages of the temperature method are discussed.