Xiangming Zhou

This study combines flood risk assessment and concrete durability research, presenting a standard-integrated, four-layer ”Hierarchical Fuzzy Inference System” linking environmental monitoring data to material-level degradation of ultra-high-performance concrete (UHPC). Layer 1 converts ambient SO2 and NO2 concentrations into a chemical hazard score and precipitation acidity; Layer 2 combines precipitation intensity and exposure duration to estimate acid deposition flux; Layer 3 represents UPHC ix parameter strength loss; and Layer 4 synthesises these outputs into a defuzzified degradation rate. Membership thresholds are anchored to ASTM C1202, ISO 9223, EN 13577, and EN 1991-1-6. The framework is illustrated for UHPC infrastructure elements on the A4 corridor in Hammersmith, London, using local roadside monitoring and North Kensington urban background records. The degradation rate increases with permeability class, from approximately 0.0024 mm/year for UHPC with negligible chloride-ion penetrability (< 100 C) to above 0.010 mm/year for the low class (1000-2000 C). With adopted coupling assumptions, adding rain deposition and hydrodynamic flood score to atmospheric exposure raises the degradation rate by approximately 33% at moderate multi-hazard severity (≈ 0.55), with flood score evaluated at an upper-bound velocity. The decision surface rises non-linearly as permeability and multi-hazard severity increase together. The transparent IF-THEN rule base allows assessment managers to trace each outcome back to physical parameters, supporting a shift from fixed-interval to risk-based inspection and maintenance. Because the framework is specified by its data requirements (not site particulars), it is transferable to other urban infrastructure contexts through re-anchoring of its data layers

This study proposed and validated an amino acid-assisted aqueous carbonation route for regulating carbonation-derived products and improving the subsequent performance of limestone calcined clay cement (LC3). A cement fraction corresponding to 25 wt.% of the final binder was carbonated as an aqueous slurry containing either glutamic acid or glycine and then incorporated into LC3. The effects of the two amino acids on CO2 uptake, CaCO3 polymorph formation, hydration kinetics, phase assemblage, pore structure, and compressive strength were systematically evaluated. Both amino acids promoted vaterite formation and modified the carbonation-derived phase assemblage. Incorporation of carbonated slurries substantially altered the hydration kinetics and later-age microstructural development of LC3. Despite suppressed early-age reaction in the amino acid-assisted carbonated systems, hydration recovered substantially between 3 and 28 days. At 90 days, LC3 incorporating glutamic-acid-assisted carbonated slurry (CLC3-glu) and glycine-assisted carbonated slurry (CLC3-gly) reached compressive strengths of 42.50 and 47.52 MPa, respectively, representing increases of 28.75% and 43.96% compared with the reference LC3. CLC3-glu exhibited pronounced later-age pore refinement, whereas CLC3-gly showed less pronounced pore refinement than CLC3-glu but continued structural and morphological evolution of the hydration products. These results demonstrate that amino acid-assisted aqueous carbonation regulated CaCO3 polymorph formation and improved later-age strength, despite lower CO2 uptake and reduced 3-day hydration.

While limestone calcined clay cement (LC3) can reduce CO₂ emissions by up to 50%, low early-age strength in low-clinker formulations limits its structural adoption. This study investigates practical strategies to accelerate early strength development in low-clinker LC3 mortars (clinker fraction ~42.5%) using sodium carbonate (1–3%) and a commercial chemical accelerator (1.5–3%), evaluated individually and in combination. Eleven mortar mixes were tested for compressive and flexural strength at 3, 7, and 28 days per BS EN 196-1, alongside ATR-FTIR hydration characterization and dry versus wet mixing comparisons. Results show that 2% sodium carbonate is the optimal single dosage, increasing 7-day compressive strength by ~29% (from 6.8 to 8.8 MPa) and flexural strength by 22% (from 2.46 to 3.00 MPa). Combining 1% sodium carbonate with 1.5% accelerator produced a pronounced synergistic effect, boosting 7-day compressive strength by 38% (9.4 MPa) and flexural strength by 34% (3.30 MPa). FTIR spectra confirmed enhanced silicate polymerization and carboaluminate precipitation. These findings provide construction practitioners with a scalable, low-cost chemical activation method to facilitate early demoulding and formwork stripping in sustainable low-carbon construction.

Rapid urbanisation in developing economies exposes building occupants to elevated ambient pollutant concentrations, while limited monitoring infrastructure prevents timely ventilation and envelope interventions. This study develops a deep learning framework for next-hour forecasting of four criteria pollutants (particulate matter (PM10), nitrogen dioxide, carbon monoxide, and sulphur dioxide) to serve as a dynamic early warning system for building-level air quality management in construction-affected urban zones. Hourly observations recorded between 2016 and 2018 (26,268 records) at a Ministry of Environment monitoring station in the Greater Amman Municipality (GAM), Jordan, were processed through a processing pipeline in which three imputation techniques were benchmarked, with moving-average imputation selected for its fidelity to pollutant fluctuation dynamics. Three recurrent architectures (long short-term memory (LSTM), bidirectional LSTM, and gated recurrent unit (GRU)) were benchmarked under identical hyper parameters. No single architecture dominated across all criteria: the bidirectional model achieved the strongest fit (correlation coefficient R=0.935), the GRU generalised best on the test subset (RMSE=76.70) with the shortest training time, and the standard LSTM performed best on the validation set (RMSE=70.82). With testing errors differing by approximately 7% across models, the findings indicate that architecture selection for deployment can be governed by computational constraints rather than accuracy loss. Forecast outputs were mapped to Jordanian exceedance thresholds (JS 1140:2006) to trigger anticipatory interventions, including ventilation modulation and building envelope sealing, thereby providing the lead time required to limit outdoor-to-indoor pollutant transfer.

Basalt fibre-reinforced polymer (BFRP) dowel bars offer a corrosion-resistant alternative to steel dowels for jointed concrete infrastructure, particularly in aggressive marine environments. However, the durability of full-scale BFRP dowels remains insufficiently understood because their shear-dominated behavior differs from that of conventional small-diameter FRP reinforcement. This study investigates the durability of 25-mm epoxy-based BFRP dowel bars exposed separately to artificial seawater and alkaline solution at 60 °C under accelerated continuous-immersion conditions. Mechanical performance was evaluated through flexural strength, transverse shear strength, and interlaminar shear strength (ILSS), while chemical and microstructural changes were examined using Fourier-transform infrared spectroscopy (FTIR) and scanning electron microscopy coupled with energy-dispersive spectroscopy (SEM/EDS). At the common 2- and 4-week exposure durations, seawater caused greater flexural-strength losses (10.0% and 11.5%) than alkaline exposure (1.3% and 8.5%), whereas alkaline exposure caused greater reductions in transverse shear strength (7.3% and 8.3% vs. 2.5% and 5.5%) and ILSS (7.0% and 10.0% vs. 1.3% and 2.4%). FTIR and SEM/EDS observations provided complementary evidence of matrix modification and fibre–matrix interfacial deterioration. Overall, the investigated BFRP dowel system showed limited mechanical degradation under the accelerated conditions, indicating promising material-level durability performance within the investigated exposure durations.