33 巻 (2015) 1 号 p. 42-48
GTA (Gas Tungsten Arc) welding has been used in manufacturing field and studied as a higher level welding process. To understand the arc phenomena that are complicated, it is significant to measure the temperature distribution of arc. Although Fowler-Milne method is one of the popular measurement methods, it is applied only under the assumption of LTE (Local Thermodynamic Equilibrium). In this study, the reliability of Fowler-Milne method is investigated by comparison of results between Fowler-Milne method and numerical simulation for six shielding gas conditions (Ar, He, N2, Ar+50%He, Ar+20%N2 and Ar+5%H2). The area in which the agreement between both ways is more than 95% is considered as that in which the LTE assumption is valid. The validity of this evaluation is discussed from the viewpoint of the collision frequency of species in the arc. The present study shows that LTE is valid for the area with the collision frequency of 2.0×1034 m-3s-1 for pure Ar arc, and the area with the collision frequency of more than 1.0×1035 m-3s-1 for other four conditions (N2, Ar+50%He, Ar+20%N2 and Ar+5%H2). However, for pure He arc, the collision frequency is much lower than the other conditions. Therefore, He arc is not in LTE for the whole area, which suggests that another method is required to measure the accurate temperature distribution for pure He arc.