This study aims to understand crack opening-closing behavior on pre-fatigued welded joints repaired by High Frequency Mechanical Impact (HFMI). Detailed rat-hole specimen models inserting different depths of rectangle slit in the weld toes to simulate initial cracks were used in the HFMI treatment simulation. Induced compressive residual stress, change of slit geometry, and opening-closing behavior of the slit after HFMI treatment simulation were investigated numerically. The amount of induced compressive residual stress around the slit tip is reduced when slit depth becomes larger, and slit size of about 0.4 mm remains when the slit depth is 2.0 mm. The opening-closing behavior of HFMI treated slit was examined based on change of strains along the slit face. The slit is opened from the bottom side prior to its treated surface. Additionally, the crack opening-closing behavior was investigated experimentally using pre-fatigued out-of-plane gusset welded joints repaired by impact crack closure retrofit treatment. Phased array ultrasonic testing system was used to examine the change of echo height from cracks. It could be concluded that similar behavior as the numerical investigation is observed. From both investigations, experimentally and numerically, the behavior on pre-fatigue welded joints repaired by HFMI could be studied successfully. 

In this study, to evaluate the influence of grid blast following to the peening process on the fatigue strength improvement, residual stress measurement and fatigue tests have been conducted. Examined peening processes were UIT, PPP and HP with ICR apparatus. For the residual stress measurement, plate and on-bead specimens made of JIS-SM400, SBHS400, SBHS500 and SBHS700 were used, and X-ray diffraction method was applied to the measurement. The results concluded that the differences in the compressive residual stress distribution in the vicinity of peened area with and without blast are not so significant. For the fatigue tests, out-of-plane gusset welded joint specimens made of JIS-SM490 and SBHS500 were used. Fatigue tests were conducted under stress ratio R=0.0 and 0.5. Fatigue tests showed that fatigue strength of peened specimens with blast are almost the same as that of peened specimens without blast, and those results were observed independently of steel grade and of peening methods.

This study aims to investigate influence of under- and over-treatment on residual stress state induced by HFMI numerically. Finite element simulations were performed using a flat plate model considering S355 where feed rate and the number of hits were varied. To reduce computational time, the mass scaling method was adopted to the simulations. In addition, in order to survey influence of surface removal on residual stress state, electropolishing was conducted after the HFMI simulation. Additional simulations were performed on bead on plate model considering JIS-SM400 in order to investigate applicability of isotropic hardening model for residual stress estimation. From the results, the mass scaling method can result in reducing computational time more than 90% with reasonable good estimation of the residual stresses.The investigations regarding under- and over-treatment reveal that high feed rate mainly influences residual stress state on the treated surface and the number of hits is independent of amount of induced residual stress. Surface removal after the simulation can result in slightly improving the accuracy of the estimated residual stresses. The simulations to bead on plate model give reasonable results in a depth of around 0.2 mm even when the residual stresses due to welding is disregarded.