This study analyzes a three-stage roll flattening process to improve the flatness of 1.5 GPa grade AHSS sheets. Unlike conventional leveler rolls, which mainly relieve residual stress through longitudinal tension-compression, the second roll has a sloped pattern to induce transverse deformation and redistribute local residual stresses. A twisted sheet was processed under different roll gap settings (1.3 mm, 1.1 mm, 0.9 mm, and 0.7 mm), and experimental measurements were compared with Abaqus Explicit simulations. At a 1.1 mm gap, the RMSE between experiment and simulation is 0.22 mm, showing the highest agreement. Both twist and crossbow defects are reduced by over 80%, achieving optimal flattening. At 1.3 mm, the simulation overestimates the second roll’s effect, causing excessive localized deformation. Reducing the gap to 0.9 mm or 0.7 mm increases discrepancies due to roll fixation differences. Experiments allow more central bending, amplifying crossbow, while simulations assume rigid rolls, underestimating curvature. Adjusting the second roll’s geometry to enhance transverse tension-compression and setting the gap to 1.1 mm effectively reduces defects. This method improves flatness while minimizing the number of rolls needed in high-strength steel sheet production.

This study focuses on an equivalent model of Polyvinyl Butyral (PVB) laminated glass to simulate the Head Injury Criterion (HIC) when a pedestrian collides with a TRAM. To simulate the collision behavior that occurs when a pedestrian’s head collides with PVB laminated glass, a comparison was made between the results of the widely used PLC model for PVB laminated glass modeling and an actual dynamic head impact test. The material properties of the tempered glass and PVB film used in the PLC and equivalent models were obtained via four-point bending tests and tensile tests, respectively. The proposed equivalent model is developed by assigning the thickness, material properties, and positional information of each layer in the multilayer PLC model to the integration points of the shell element. The results of the equivalent analysis model were found to accurately simulate the collision behavior when compared with the results of both the dynamic head impact test and the PLC model. Moreover, the analysis cost improved to approximately 15% of that of the traditional PLC model.

This study establishes experimentally grounded circumferential thermal criteria for heat-shrinkable crosslinked polyethylene (PEX) joint casings by coupling DSC-defined thermal activation with through-thickness thermal lag measured under trench-constrained irradiation. The activation temperature was identified as 140 °C from DSC, while an upper bound of the allowable outer-surface temperature was set to avoid thermal damage during installation. Full-scale temperature mapping revealed persistent circumferential non-uniformity caused by geometric line-of-sight limitations and inter-module gap regions, where the outer-surface temperature remained approximately 10–15 °C lower than directly irradiated locations, and the inner surface exhibited a delayed response due to the low thermal conductivity of PEX. Based on these observations, a two-stage heating sequence—an initial high-power stage followed by a reduced-power soaking stage—was experimentally derived to satisfy dual constraints: achieving inner-surface activation (≥140 °C) while maintaining the outer surface below the conservative outer-surface upper bound (~280 °C) and reducing circumferential temperature differences without surface overheating. Comparative joint tests confirmed that the proposed thermal criteria and sequence promote stable interfacial bonding and cohesive failure in the mastic layer, yielding higher repeatability and smaller strength scatter than conventional manual torch heating. The proposed framework provides experimentally grounded thermal criteria and a transferable procedure for designing heating conditions for heat-shrinkable polymer casing systems under constrained field environments.