The complex cross-sectional shape of oversized beam blanks and the size effect of ultra-large-section beam blanks create severe issues related to the surface and internal quality of the castings. To ensure quality and control in the production of ultra-large-section beam blanks, a numerical and physical model of molten steel flow in the three-port submerged entrance nozzle (SEN) mould, with section dimensions of 1300 × 510 × 140 mm, was established. This model was created using numerical simulations and NSGA-II genetic algorithm optimisation, and the impact of the casting speed and SEN immersion depth on the mould’s flow behaviour was investigated. The results showed that a deeper SEN immersion depth resulted in, a greater impact depth of the molten steel, and the surface flow velocity decreased. Both the impact depth and the surface flow velocity of the molten steel increased with increasing casting speed. The physical simulation and numerical simulation of the molten steel flow form and flow velocity distribution in the mould were in good agreement with each other, thus verifying the accuracy of the numerical simulation. The process parameters derived from this study were all within an appropriate range, which can help to improve the quality of continuously cast beam blanks. This also provides guidance for selecting optimal parameters for actual continuous casting production processes.

Lightweight, high-strength materials formed by bonding aluminum alloy (Al) and carbon-fiber-reinforced thermoplastic (CFRTP) composite materials are widely used in impact-exposed structures, leading to structural damage and reduced service life. In this study, transverse impact experiments were conducted on Al/CFRTP adhesive joints with varying impact energies (10 J, 20 J, 30 J, 40 J) on different impact surfaces (Al and CFRTP). The energy absorption rate, damage morphology, residual strength, strain evolution on the impact surface, and failure mechanisms of the joint cross-section were analyzed. The results show that using CFRTP as the impact surface can minimize damage and maintain residual tensile strength. Regardless of whether Al or CFRTP is used as the impact surface, the energy absorption rate remains around 80%, with the final impact outcome being determined by both impact damage and mechanical interlocking. This study provides a theoretical basis for the design of high-impact-resistant Al/CFRTP adhesive joints.

Flow maldistribution in a single-inlet, closed-end cooling manifold arises from the interaction between pressure recovery in the main header and the distributed hydraulic resistance of the branches. This maldistribution can cause non-uniform transverse cooling of hot-rolled steel strips, thereby affecting product quality and strip-shape stability. This study investigates the underlying flow physics of a U-shaped cooling manifold using an experimentally validated computational fluid dynamics model. Continuous mass extraction through successive branches causes progressive decay of the axial momentum in the main header and a corresponding recovery of static pressure, thereby providing progressively greater hydraulic driving pressure to downstream branches. Near the closed end, however, recirculation and enhanced shear dissipation locally weaken the pressure-recovery process and generate localized spatial variations in terminal branch flow rates. The results further show that the outlet-to-inlet area ratio alone does not uniquely determine the flow-distribution behavior: varying the main-header diameter primarily modifies the axial velocity level and pressure-recovery gradient, whereas varying the throttling-orifice diameter mainly changes branch resistance and discharge capacity. These findings indicate that spatial coordination between header pressure recovery and branch resistance provides a practical physical basis for controlling flow maldistribution. Based on this physical understanding, a segmented orifice-resistance distribution reduces the flow non-uniformity coefficient from 0.828% to 0.756% and the total system pressure drop from 2867.51 to 2310.8 Pa, corresponding to reductions of 8.70% and 19.41%, respectively. The findings provide a physics-informed interpretation and a practical resistance-redistribution strategy for controlling flow maldistribution in closed-end multi-branch cooling manifolds.