Evaluation of Growth of Subsurface Crack by Fatigue: Simulation of Penetration of Internal Flaw to Surface of the Structure
Masanori Kikuchi,Shinji Yamada,Ryousuke Serizawa,Li Yulong
DOI: https://doi.org/10.1115/pvp2013-97580
2014-01-01
Abstract:In nuclear power plant, there is a proximity rule to evaluate subsurface crack, which exists near free surface of the structure. If the distance between this subsurface crack to free surface exceeds some limit, this subsurface crack is transformed to surface crack, and residual fatigue life is evaluated. Subsurface crack has many parameters, such as, crack length, crack depth, distance between crack front to free surface, and aspect ratio of subsurface crack. As a result, proximity rule is proposed by many organizations, and each rule is different from each other largely. It is necessary to verify which is more preferable, but to determine it experimentally is very difficult due to existence of many parameters. Numerical simulation is needed for this purpose. This problem is simulated using S-version FEM. Using S-FEM, subsurface is modeled independently from global structure, and crack growth is easily simulated. In maintenance code of nuclear power plant, initial defects are modeled as elliptical cracks in a normal plane to tension loading direction, and growth rate is estimated in this plane. But by using S-FEM, real defect shape is modeled realistically, and crack growth by fatigue is simulated. Usually, such small defects are subjected to multi-axial loading, and crack growth behaviors are very complicated. Finally, detect shape becomes elliptical or circular crack in a plane normal to tension loading direction in the structure. Fatigue cycles for these growing processes are calculated, and conservativeness of this maintenance code is discussed. Then subsurface crack growth is simulated. Inner subsurface crack grows toward free surface, penetrate to free surface and grows as a surface crack. These processes are simulated smoothly by S-FEM. Parametric studies are conducted for this problem, and proximity rules are verified with numerical results.