Earth and Environmental Sciences

Due to the porous nature of clay brick waste, achieving efficient resource utilization presents significant challenges. This study employed porous clay brick waste-derived recycled sand and recycled powder as crack inducers to develop eco-friendly high-ductility cementitious composites (HDCC). The findings reveal that recycled powder contained a substantial amount of SiO 2 crystalline phases and some amorphous components, exhibiting good filling effect and moderate pozzolanic activity. Substituting recycled powder for cement and slag inhibited hydration reactions and reduced hydration product formation, particularly at high replacement levels, leading to increased flaws and micro-cracks in the HDCC. The HDCC strength decreased with increasing recycled powder content, yet remained relatively high even at a 70% replacement level. Moderate incorporation of recycled powder enhanced tensile strength and strain while increasing crack numbers, whereas high-volume addition diminished tensile performance of HDCC. Porous recycled sand incorporation provided effective internal curing and improved HDCC strength and ductility; meanwhile, incorporating recycled sand introduced numerous uniformly distributed micro-pores, promoting uniform crack initiation and propagation within the HDCC. Even at 100% replacement of silica sand, recycled sand-incorporated HDCC exhibited better tensile performance compared to the reference HDCC. Combining recycled sand with recycled powder could produce fully recycled HDCC that met diverse strength and ductility requirements. The tensile strengths of HDCC-Reference, 100SS-C30S30, and 100SS-C70S70 were 11.62 MPa, 12.18 MPa and 8.49 MPa, with corresponding tensile strains of 5.67%, 6.07% and 6.53%.

ABSTRACT The reuse of steel reinforcing bars (rebars) recovered from demolitions can reduce resource demand, energy consumption, and recycling costs while promoting circular economy practices. However, systematic knowledge on their surface condition and repassivation capacity remains limited. This study assessed the surface corrosion condition of 94 rebars (376 sections), collected from various reinforced concrete elements in six demolition sites in the São Paulo metropolitan area. Rebar surface corrosion condition was visually classified, a reusability index (RI) was determined. A quantitative screening was conducted on representative segments using micro-XRF mapping, mass variation after mechanical cleaning, caliper-based cross-section assessment, and complementary 3D scanning to support visual classification. The repassivation potential was comparatively evaluated through a five-group electrochemical program under alkaline (pH ∼12.5) and carbonated (pH ∼8.0) conditions in saturated Ca(OH) 2 solution used as a simplified concrete pore solution analogue. Surface assessment revealed that 72% of section-level observations exhibited negligible corrosion. Exploratory statistical analysis identified structural elements as the most relevant factor within the sampled dataset: slabs showed the lowest reusability index, whereas columns showed the highest. By contrast, the effects of diameter, building position, and demolition site could not be robustly isolated. Electrochemical monitoring (Ecorr and Icorr) showed that, under alkaline conditions, the combined mechanical and chemical cleaning procedure promoted repassivation behavior comparable to new rebars, while other treatments only reduced corrosion activity. After carbonation, all groups shifted to more unfavorable electrochemical conditions, with no practical differences between methods under the solution boundary conditions of this study. These results support the use of surface-based screening to identify reused rebars with more favorable conditions for low-risk, non-structural applications with shorter service lives, while broader datasets, a complete mechanical properties screening, and long-term durability validation remain necessary before large-scale implementation.

The exchange of building models between architectural design and structural analysis remains a persistent challenge in practice. Although Industry Foundation Classes (IFC) models provide detailed geometric descriptions, they often lack the analytical abstractions required for reliable FEM modelling, such as structural grids, reference axes, and consistent connectivity. Consequently, the transformation from BIM geometry to analysis-ready structural models frequently requires substantial manual interpretation and correction. This paper presents a learning-based preprocessing framework for inferring analytical structural grids and substructures from IFC-derived geometry. The current implementation operates on two-dimensional point sets derived from column locations and processes structural layouts using Dynamic Graph Convolutional Neural Networks (DGCNNs) for substructure grouping and grid-index classification. Predicted grid representations are combined with deterministic post-processing for analytical grid fitting and assignment correction. The approach is evaluated on synthetic datasets and demonstrated on realistic IFC-derived building models. Results show consistent recovery of structural grid systems for a range of column-governed layouts and support automated generation of analysis-ready FEM models with reduced manual preprocessing effort.

Membrane action can substantially enhance the resistance and deformation capacity of restrained reinforced concrete (RC) elements, offering critical additional capacity for elements subjected to impulsive loading. However, its practical utilisation remains limited, largely due to the absence of a consistent, simplified approach capable of describing the full structural response. This study conducted experimental investigation of membrane action in six longitudinally and rotationally restrained RC beams subjected to static three-point bending. The specimens varied in slenderness, reinforcement ratio, and fracture strain of the reinforcement. All beams exhibited combined flexural and membrane action, with an enhanced peak resistance attributable to compressive membrane action (CMA). Clear differences in reinforcement fracture and cracking development were observed in the post-peak response across the specimens, reflecting the influence of geometry, reinforcement properties and boundary conditions. The results showed that CMA contributed significantly to the energy absorption capacity of specimens with lower slenderness or lower reinforcement ratios, while more slender beams with reinforcement with higher fracture strain benefited primarily from tensile membrane action (TMA). Reinforcement fracture during the CMA phase was observed in some specimens, underscoring the need for predictive methods that explicitly account for this phenomenon. Two simplified approaches for predicting the full load–deflection response were also presented: a finite element (FE) model and an analytical approach extending an existing framework. The FE model showed very good agreement with the experimental results, accurately capturing peak resistance and reinforcement fracture. The analytical model provided satisfactory predictions, particularly during the CMA phase.