PSI - Issue 7

2

NADOT Yves / Structural Integrity Procedia 00 (2017) 000–000

Y. Nadot et al. / Procedia Structural Integrity 7 (2017) 530–535 Peer-review under responsibility of the Scientific Committee of the 3rd International Symposium on Fatigue Design and Material Defects. 531

Keywords: High cycle fatigue, Surface artificial defect, Armco iron, Grain size, Dimensionless Kitagawa diagram

1. Introduction The fatigue design of metallic cast parts is strongly linked to the casting process. The designer needs to reach a compromise between the fatigue resistance of the component and the allowable defect size due to the process. In order to perform this optimization, a criterion that takes into account defect influence on the fatigue limit is necessary. Many approaches have been proposed in order to assess the influence of a defect on the fatigue life. An overview of that problem can be found in Murakami’s work [1]. A defect can be considered against fatigue using four parameters: defect type (inclusion, pore, shrinkage, ... ), defect morphology (spherical, hemispherical, complex, ... ), defect position (internal, sub-surface or surface), defect size (related or not to loading direction). In the present study, a defect is defined as a hemispherical pore on surface and the defect size is a variable whose influence on the fatigue strength is studied. Murakami and Endo [2] proposed an empirical approach where the defect size is measured in relation with loading direction using the ‘ area ’ defect parameter. The influence of the defect size on the fatigue limit can be characterized by two regimes: for small defects, the defect does not affect the fatigue resistance of material, whereas for a large defect, the larger the defect the stronger decrease is observed. The designer needs more than a simple criterion so that a general methodology, such as the Through Process Modeling (TPM ) proposed by Li et al. [3] would be helpful for a general optimization of fatigue life including casting process simulation. For cast parts, defects are typically from few microns to millimeters so that the microstructure surrounding the defect is of the same order that the defect size because cast microstructure is generally rough. Some authors, as Vallellano et al. [4], tend to propose a microstructural model with short crack arrest approach. According to the authors, this model is limited to uniaxial tension and small mean stress domain. Related to the morphology of the defect, the full 3D analysis of Buffiere et al. [5] aims at understanding the role of complex geometry of the defect in relation with the microstructure. The analysis proposed by Nicoletto et al. [6] adds an important information to the 3D damage analysis obtained by XR tomography: the variation of stress concentration factor as a function of complex shrinkage 3D geometry and loading. In this framework, a criterion based on stresses at defect scale becomes realistic. 2. Material and experimental set-up The purpose of this study is to determine the influence of defect size, in relation to microstructure characteristic length, upon fatigue limit. This is why the material chosen for the experimental study is the Armco iron. The microstructure of this pure iron is simple: it is composed of a single phase and the grain morphology is equiaxed. So its microstructure at the mesoscopic scale can roughly be characterised by one characteristic length, the grain size. To compare defect size to grain size, it is extremely important to control and measure the grain size of the material in a precise way. Among several definitions of the grain size proposed in the literature, the average grain diameter, noted D g , was chosen in order to compare the latter with the defect surface diameter. D g corresponds to the diameter of the disk whose area equals to the average grain area. The Armco iron bar is manufactured by rolling in the longitudinal direction. Rolling does not influence the average grain diameter in every direction. The average grain diameter obtained for the Armco iron bar is 30 µ m. The smallest manufactured defect diameter obtained by EDM for this study at the surface is equal to 130 µ m. Thus, it is difficult to introduce defect smaller than grain in the Armco iron raw bar. To study the influence of defect size to grain size relationship upon fatigue limit, it is necessary to increase the grain size. The grain diameter must be higher than 130 µ m and lower than 500 µ m, in order to keep at least about a dozen of grains within the cross section. Otherwise experimental fatigue limits obtained would have been too dependent on the local grain orientations and the scatter of experimental results would have increased. Inspired by Pipan’s work [7], the process to control the grain size of the commercially pure iron is in two stages. First, cold compressive plastic strain is applied in a hydraulic press. The strain level applied at this stage will determine the size of the grains after the recrystallization. The second stage is a temperature ramp from room temperature to 850 C at a rate of 2 C/min; then, the temperature is to be kept constant during 150 h and finally decreases to room temperature at a rate of 2 C/min. At this stage new crystals growing at the expense of the old ones are formed. The final grain size obtained depends critically on the level of