Ou study focuses on the effects of phase transformations on crack propagation. We want to understand the changes of fracture toughness during welding. In this work, fracture toughness is expressed by J-integral. There are many experimental methods to obtain the critical toughness JIC but they are impractical for our investigation during phase transformation. That is the reason why we have proposed a method coupling mechanical tests, digital image correlation and finite element simulation. The fracture tests are implemented on pre-cracked single edge notched plate sample which is easy for machining and heat conduct during phase transformation. The tests are conducted at different temperatures until rupture. Digital image correlation gives us the displacement information on every sample. Each test is then simulated by finite element where the fracture toughness is evaluated by the method G-Theta at the crack propagation starting moment found by potential drop method and digital image correlation technical. Two materials have been studied, 15Cr-5Ni martensitic precipitation hardening stainless steel and 16MND5 ferritic low carbon steel. For these two materials, different test temperatures were chosen before, during and after phase transformation for testing and failure characterization of the mechanical behavior. Investigation result shows that metallurgical phase transformation has an influence on fracture toughness and further crack propagation. For 15-5PH, the result of J1C shows that the as received 15-5PH has higher fracture toughness than the one at 200°C. The toughness is also higher than the original material after one cycle heat treatment probably due to some residual austenite. Meanwhile, pure austenite 15-5PH at 200°C has higher fracture toughness than pure martensitic 15-5PH at 200°C. For 16MND5, the result also proves that the phase transformation affects fracture toughness. The as received material has bigger J1C than the situation where it was heated to 600°C. On the other hand, the material at 600°C just before isothermal bainite transformation after the austenitization during cooling process also has higher fracture toughness than the one at 600°C before austenitization. These two conclusions are consistent well with the result of 15-5PH. But the final situation of 16MND5 after one cycle heat treatment has a slightly smaller J1C than the receiving situation. It means that one cycle heat treatment hasn't an significant influence on 16MND5fracture toughness. Conclusions show that one should pay attention to the heating period before austenitization of the substrate material when people do the welding as the higher temperature will bring the lower fracture toughness during this process. While during cooling period, the fracture toughness doesn't change a lot during, before or after the cooling induced phase transformation. Even for 15-5PH, it has a better fracture toughness after the martensite transformation than before.

Cette thèse porte sur l’influence des transformations de phases solide-solide sur la propagation de fissure. On souhaite ainsi mieux comprendre les variations de ténacité en cours de soudage par exemple, ou bien pendant la réparation d’une fissure. Dans ce travail, la ténacité est obtenue à partir de l’intégrale J. Il existe de nombreuses méthodes expérimentales permettant d’obtenir la ténacité critique JIC mais qui sont difficilement applicables pour des essais se déroulant pendant une transformation de phase. C’est pourquoi nous avons proposé une méthode couplant essai mécanique et mesure par corrélation d’images avec de la simulation par éléments finis. Les essais sont réalisés sur de simples éprouvettes plates pré fissurées, faciles à usiner et simple à chauffer par induction. Les essais sont conduits pour différentes températures et jusqu’à rupture. En sus des mesures d’efforts et déplacements de traverse, la corrélation d’images nous fourni également les champs de déplacement sur chaque face de l’éprouvette. Chaque essai est ensuite simulé par éléments finis où la ténacité critique est calculée par la méthode G-Theta au maximum de la charge supportée par l’éprouvette. Les simulations précédentes intègrent les conditions aux limites obtenues par corrélation et le comportement mécanique considéré est celui que nous avons identifié sur des essais de caractérisation. Deux nuances de matériau ont été étudiées avec cette méthode ; l’acier inoxydale 15-5PH ainsi que l’acier ferritique 16MND5. Pour ces deux matériaux, différentes températures d’essai ont été choisies avant, pendant et après la transformation pour effectuer les essais de rupture ainsi que de caractérisation du comportement mécanique. Les résultats de cette étude montrent que la transformation de phase peut avoir un impact non négligeable sur la ténacité. Ainsi, pour le 15-5PH, le taux d’austénite résiduel est un facteur important et les essais pendant la transformation martensitiques montrent que la ténacité critique peut être inférieure pendant celle ci à celle du matériau purement austénitique. Dans le cas du 16MND5, la ténacité est beaucoup plus faible à 600°C (et bainitique) qu’à température ambiante ce qui est assez logique. Par contre, lors du refroidissement, depuis 600° (austénitique) jusqu’à la température ambiante (bainitique), nous avons obtenu une ténacité critique relativement constante. En conclusion, cette étude apporte une solution quant à la mesure de la ténacité critique de matériau pendant des transformations de phases, ce que ne permettent pas forcément les essais normalisés. Pour le 15-5PH, la ténacité critique semble évoluer pendant la transformation martensitique et est assez dépendante du taux d’austénite résiduelle. Il semble par contre que pour le 16MND5, la ténacité critique soit assez peu dépendante de la fraction volumique d’austénite et la valeur obtenue varie peu au cours du refroidissement du matériau depuis 600°C. / Ou study focuses on the effects of phase transformations on crack propagation. We want to understand the changes of fracture toughness during welding. In this work, fracture toughness is expressed by J-integral. There are many experimental methods to obtain the critical toughness JIC but they are impractical for our investigation during phase transformation. That is the reason why we have proposed a method coupling mechanical tests, digital image correlation and finite element simulation. The fracture tests are implemented on pre-cracked single edge notched plate sample which is easy for machining and heat conduct during phase transformation. The tests are conducted at different temperatures until rupture. Digital image correlation gives us the displacement information on every sample. Each test is then simulated by finite element where the fracture toughness is evaluated by the method G-Theta at the crack propagation starting moment found by potential drop method and digital image correlation technical. Two materials have been studied, 15Cr-5Ni martensitic precipitation hardening stainless steel and 16MND5 ferritic low carbon steel. For these two materials, different test temperatures were chosen before, during and after phase transformation for testing and failure characterization of the mechanical behavior. Investigation result shows that metallurgical phase transformation has an influence on fracture toughness and further crack propagation. For 15-5PH, the result of J1C shows that the as received 15-5PH has higher fracture toughness than the one at 200°C. The toughness is also higher than the original material after one cycle heat treatment probably due to some residual austenite. Meanwhile, pure austenite 15-5PH at 200°C has higher fracture toughness than pure martensitic 15-5PH at 200°C. For 16MND5, the result also proves that the phase transformation affects fracture toughness. The as received material has bigger J1C than the situation where it was heated to 600°C. On the other hand, the material at 600°C just before isothermal bainite transformation after the austenitization during cooling process also has higher fracture toughness than the one at 600°C before austenitization. These two conclusions are consistent well with the result of 15-5PH. But the final situation of 16MND5 after one cycle heat treatment has a slightly smaller J1C than the receiving situation. It means that one cycle heat treatment hasn't an significant influence on 16MND5fracture toughness. Conclusions show that one should pay attention to the heating period before austenitization of the substrate material when people do the welding as the higher temperature will bring the lower fracture toughness during this process. While during cooling period, the fracture toughness doesn't change a lot during, before or after the cooling induced phase transformation. Even for 15-5PH, it has a better fracture toughness after the martensite transformation than before.