In this study, the laser cutting characteristics were analyzed according to the shape of the back side of the specimen, and the laser cutting characteristics were compared according to the thickness of the edge (10 mm, 20 mm, and 30 mm). A Yb-YAG laser was used in this study, and the cutting target was STS304 with a thickness of 50 mm, and the cutting process was analyzed using a high-speed camera. In the experiment, it was found through image analysis that the cutting performance was excellent at 30 mm thickness of the edge. In order to analyze this reason, a thermal conduction analysis (numerical simulation) was performed, and it was confirmed that the thicker thickness of the edge caused a preheating effect during laser cutting due to a large amount of heat accumulation. This effect can be used as a reference for the initial processing state while cutting thick metals as it is a characteristic that has not been revealed before.

To understand the relationship between microstructure and corrosion, in this study, underwater bead-on-plate laser welding was compared with the in-air laser welding of 10-mm-thick 304 stainless steel plates at different laser powers (2, 4, and 6 kW). Welding was performed via local dry underwater laser welding (UWLW) using a custom-designed nozzle and a fiber laser at a water depth of 70 mm. The best weld quality was obtained in both underwater and in-air environments using 2 kW of laser power. To understand the relationship between the microstructure and corrosion resistance of 304ss in underwater laser welding (UWLW), this study was conducted using a custom-designed nozzle. The grain boundary analysis revealed that the specimen prepared by UWLW had high-angle grain boundaries content approximately 1.5 times higher than that of the specimen produced by in-air laser welding, and the fraction of the coincidence site lattice (CSL) boundaries was increased remarkably. High residual stress and microchromium precipitation were observed in the UWLW specimen, and the corrosion rate of the same at 2 kW laser power was considerably similar to that of the in-air laser weld specimen.

While dismantling nuclear power plants, the reactor vessel internal is cut underwater using mechanical and thermal cutting. In laser thermal cutting, assist gas must be used to remove melted metal; consequently, a large number of radioactive aerosols can be generated. To reduce the generation of aerosols, the assist gas pressure should be lowered. However, below the pressure limit, the molten metal is not well-removed from the cut surface and fails to cut. In this study, an assist gas visualization experiment was performed to find a condition for the gas to flow well inside the cut surface, even at low pressures. The top kerf width, nozzle type, distance between nozzle and specimen, and assist gas pressure were selected as process parameters, and in the case of large top kerf width condition, assist gas was able to penetrate deeply. In the actual laser-cutting experiment, the laser beam focus position was set to −20 mm and −30 mm. In the case of −30 mm, the top kerf width was widened due to the characteristics of the laser beam profile, and cutting was successful even though the assist gas pressure was lowered by 20%.

Despite the dismantling structures that are submerged to significant depths of water during the decommissioning of nuclear power plants, there is limited research on deep-water laser cutting processes. A self-designed pressurized chamber was used in this study and successfully conducted the world’s first laser cutting experiment in a simulated 10 m water depth environment. laser cutting was performed in a 10 m underwater environment, and the cutting efficiency was compared to that observed in a 1 m underwater environment. Therefore, A 100 mm thickness of 304 stainless steel was successfully cut underwater, and the highest cutting speed of 100 mm/min was achieved. The result indicates that, as the cutting speed increased during underwater laser cutting, both the heat input and the mass flow rate of the assist gas decreased, resulting in a narrower rear kerf width and an ineffective evacuation of the molten metal.

Electron beam welding (EBW) has been attracted due to its high energy density, high depth-to-width ratio, low thermal strain, compared with arc and laser welding. In addition, it ensures high quality weld joint in structural metals in a wide range of thickness from 0.025 mm to 300 mm. Therefore, EBW is widely used in industries such as automotive, aviation, nuclear power plant, etc. Although a lot of studies on EBW for several decades have been presented, EB based repair welding with supplement of filler metal has less reported. In this work, EB repair welding for 5083 aluminum alloys was investigated because the aluminum alloys are becoming more important as a lightweight material in mobility industries. Material characterization and mechanical properties of the repair-welded specimens were evaluated, compared with characteristics of base metal and bead welding without repair. The optimized experimental conditions showed tensile strength of 294 MPa, which is a similar to base metal. Our study indicates that EB repair welding can be considered for the candidate of repair welding of aluminum alloys.

Underwater laser cutting is a promising remote segmentation method for highly activated reactor internals because it enables non-contact cutting while maintaining water shielding. Although previous studies have demonstrated the feasibility of thick-section underwater laser cutting, process monitoring remains challenging because bubbles, optical scattering, local gas-cavity behavior, water cooling, and complex melt ejection can alter the relationship between the cutting state and externally measured signals. Therefore, the monitoring characteristics of different sensing modalities during underwater thick-section cutting need to be clarified. This study synchronized a camera, a broadband photodiode, and a triaxial accelerometer during underwater laser cutting of 100 mm thick STS 304 stainless steel. Camera observations were used as the reference for identifying process states. During partial penetration, a broad lateral plume developed, followed by transient lower-lateral ejection and pronounced downward discharge at full penetration. The photodiode mean decreased by 88.8% and 93.0% from partial to full penetration in Experiments 1 and 3, respectively, and remained low afterward. Accelerometer responses around full penetration were less consistent, whereas in Experiment 2, late-stage abnormal behavior was accompanied by a 128.8% increase in the mean PSD in the 75–245 Hz band. These results indicate distinct monitoring characteristics of the three sensing modalities.