The discharge of treated municipal wastewater into coastal ecosystems poses a significant environmental challenge in semi-arid regions. This study evaluated the phytoremediation potential of vetiver grass (Chrysopogon zizanioides (L.) Nash) for the uptake of heavy metals from the effluent of the Bushehr Municipal Wastewater Treatment Plant. A randomized complete block design with four replications and ten irrigation treatments was conducted at Persian Gulf University. The treatments included different ratios of treated municipal wastewater and freshwater, wastewater irrigation combined with natural organic fertilizer, and pH adjustment. Heavy metal uptake was strongly influenced by irrigation regime and rhizosphere management. The highest uptake efficiencies for Pb (99
This study aims to improve the hydraulic stability of bridge piers and reduce local scour by examining the influence of pier base geometry on key hydrodynamic parameters. Despite extensive research on flow patterns around bridge piers, a reliable integrated physical index that enables comprehensive evaluation of hydraulic stability and systematic comparison of different pier geometries is still lacking. To address this gap, a novel index named BPHSI (Bridge Pier Hydraulic Stability Index) was introduced to integrate the most influential parameters affecting scour. Three-dimensional numerical simulations of flow around six pier models were conducted using FLOW-3D. The models were categorized into three geometric families—SMH, DUH, and COH—and analyzed under two geometric configurations: curved (a) and angular (b). Key hydrodynamic variables, including pressure, turbulence intensity, turbulence dissipation rate, shear velocity, and bed elevation changes, were extracted and evaluated. Results showed that in the SMH (a) model the turbulence dissipation rate increased from about 0.0070 upstream to approximately 0.0095 near the pier, while angular configurations produced stronger flow fluctuations. The BPHSI values in the upstream region ranged from 34.69 to 37.05, with the COH model showing the greatest sensitivity to geometric variation. Overall, curved pier fronts exhibited more stable hydrodynamic behavior and lower scour potential than angular forms, indicating that the BPHSI index can serve as a practical quantitative tool for bridge pier design optimization.
This study investigates the effectiveness of various shear reinforcement schemes in flat slabs with edge columns and openings through experimental testing and numerical analysis. Four specimens with dimensions of 1500×1000×120 mm, each containing one opening, were experimentally tested to evaluate different shear reinforcement schemes (bent bars, shear bands, and shear studs). The experimental program assessed maximum load capacity, deflection, steel strain, energy absorption, and ductility. Numerical analysis using ANSYS v.21 validated the experimental results and extended the investigation to six additional models. The study comprised three groups: solid specimens with different reinforcement schemes, specimens with one opening and varied reinforcement, which were compared to the practical slabs, and specimens with two openings, focusing on shear stud performance. Experimental results demonstrated that shear reinforcement significantly enhanced structural performance, with shear studs showing superior effectiveness by improving first cracking load by 22
Steel plates are widely used in protective and industrial structures where resistance against blast loading is a critical design requirement. Understanding the influence of key structural parameters on the blast-induced response of steel plates is therefore essential for improving their performance and safety. In this study, the blast response of square steel plates is numerically investigated with particular emphasis on the effects of plate thickness and material strength. A series of three-dimensional finite element models is developed using the LS-DYNA software, in which the interaction between the blast wave and the structure is simulated through an Arbitrary Lagrangian–Eulerian (ALE) formulation coupled with fluid–structure interaction (FSI). Square steel plates with different thicknesses and steel grades are analyzed under various blast intensities defined in terms of scaled distance. The numerical models are first validated against available experimental results, demonstrating good agreement in terms of central deflection and deformation patterns. Subsequently, a comprehensive parametric study is conducted to evaluate the influence of thickness, material strength, and scaled distance on key response parameters, including mid-point displacement, effective plastic strain, and energy components. Unlike previous studies that primarily focused on individual parameters or specific blast scenarios, the present work provides a systematic comparative evaluation of the combined influence of plate thickness, steel grade, and blast intensity within a unified validated numerical framework. The study further establishes engineering-oriented relationships between deformation, effective plastic strain, energy response, and the relative importance of the investigated design parameters for blast-resistant steel plate design. The results indicate that plate thickness plays a dominant role in controlling the blast-induced response, whereas the influence of material strength is comparatively less pronounced, particularly under severe blast loading conditions. Furthermore, the sensitivity of structural response parameters to the considered variables is quantified to identify the most influential design factors. The findings of this study provide useful insights into the blast-resistant behavior of square steel plates and can assist engineers in selecting appropriate plate configurations for steel structures subjected to explosive loading.
This study investigated a composite activation strategy for a system composed of steel slag (SS), hemihydrate flue gas desulfurization gypsum (HFGD gypsum), and ground granulated blast-furnace slag (GGBFS), aiming to evaluate the feasibility of using SS as an activator to replace ordinary Portland cement (OPC). By analyzing the performance development and microstructural evolution of specimens at different curing ages, the reaction mechanism and the factors governing performance development were elucidated. The results showed that all systems produced similar hydration phases, while the synergistic interaction between SS and HFGD gypsum mainly regulated the quantity, rather than the type, of hydration products. Among the mixtures investigated, SW4, containing 10
In this research, a new unified AI-based optimization approach for ductile concrete, which is made up of recycled glass fiber and industrial slag, has been developed in order to increase its seismic robustness. The current work differs from the previous literature on this topic because, unlike the previous works, where each AI algorithm was used for one prediction problem, the proposed framework combines reinforcement learning based mixture optimization, physics informed GNNs, transformers for fatigue prediction, and self-supervised crack detection through hybrid neural networks. The optimized formulation has demonstrated an improvement in the ductility index from 4.5 to 7.8 (Ductility increased by 73
Scour of piers is one of the primary issues in bridge engineering, and so the scour depths need to be estimated accurately to ensure the integrity and safety of bridges. There are numerous traditional equations available to predict the equilibrium scour depth for different flow, geometry, and bed roughness conditions. However, none of these give a more accurate estimation of scour depth for a wide range of input parameters. To overcome these drawbacks, the present study proposes a systematic comparison of eleven computational models comprising five hybrid architectures with six well-known empirical formulas. In order to train the model, 201 datasets of clear water scouring were collected from various published literature. The Buckingham π theory is then used to formulate a robust analytical model that reduces the physical parameter to a functional relationship (dse/y) = f (b/y, d50/y, Fr, u/uc, σg). This dimensional analysis identifies the normalized equilibrium scour depth (dse/y) as a function of flow intensity (u/uc), flow shallowness (b/y), relative sediment size (d50/y), sediment gradation (σg), and the Froude number (Fr). The unique contribution of this work is to generate a governing mathematical equation via Gene Expression Programming (GEP), which was subsequently integrated as a physics-informed constraint within the PINN framework to ensure physical consistency. Selection of input parameters plays an important role in accuracy; optimal feature selection was selected by subset regression and permutation importance to measure the dependency between features. Sequential and recursive elimination techniques are also utilized to isolate the most influential parameters for predicting accurate normalized equilibrium scour depth (de/y). The study showed that the developed GEP_PINN model achieved superior predictive performance with accuracy of a Mean Absolute Percentage Error (MAPE) below 12.0
Non-thermal plasma is an effective Advanced Oxidation Process (AOP) that can be enhanced through catalytic integration with semiconductor materials. This study investigates the synergistic plasma-catalytic removal of Methylene Blue (MB) dye from aqueous solutions using zinc oxide (ZnO) nanoparticles as a catalyst. The effects of operational parameters, including treatment time (16–64 min), initial MB concentration (3–27 mg/L), ZnO concentration (31–269 mg/L), applied voltage (13–16 kV), and pH (0.5–12.4), were evaluated. The MB removal efficiency was optimized using Response Surface Methodology (RSM) based on a Central Composite Design (CCD) implemented in Design Expert software (Version 13). The optimal multi-objective conditions for the plasma–ZnO process were determined as an applied voltage of 15 kV, initial MB concentration of 20 mg/L, treatment time of 31 min, pH of 9, and ZnO concentration of 200 mg/L, resulting in a predicted MB removal efficiency of 73.9
In coastal regions, freshwater scarcity has increased interest in the use of alternative curing media, such as seawater and raw wastewater, although their influence on concrete performance remains insufficiently understood. This study evaluated the 28-day mechanical performance, durability-related indicators, and qualitative microstructural morphology of concrete cured in potable water, seawater, and raw wastewater using mixtures incorporating natural zeolite (Z) and nano-silica (NS) as supplementary cementitious materials. All concrete mixtures were prepared using potable water as the mixing water, whereas seawater and raw wastewater were used only as curing media. Compressive strength, splitting tensile strength, water absorption, maximum water penetration depth, and qualitative SEM observations were evaluated. Compared with potable water curing, seawater and raw wastewater reduced the compressive strength of the reference concrete by approximately 5.7
Reinforced concrete (RC) flat slabs are widely being applied to almost every building structure due to their distinct advantages. Therefore, it is important to precisely foresee the governing slab failure modes. This research was targeted at predicting the main failure mode of RC slab-column connections subjected to unbalanced moment and various vertical shear forces. Thus, the failure modes of the connections were derived by comparing the unbalanced moment capacity at punching shear failure and unbalanced moment strength at the flexural mechanism. The unbalanced moments of shear and flexural modes were respectively controlled by the codes guidelines and yield line theory (YLT). The procedure was validated by the results of experimental tests carried out at authentic research in the literature. Afterward, 200 case studies were done on the connections under moment transfer at 20
Hyperparameter tuning remains a critical challenge in neural-network applications for hydraulic engineering, particularly when datasets are small and derived from laboratory experiments. This study develops a benchmark-driven hyperparameter optimization framework for predicting wave-induced equilibrium scour depth beneath submarine pipelines using three physically relevant input variables: embedment ratio (e/D), Shields parameter (θ), and Keulegan–Carpenter number (KC). An exhaustive Grid Search over 2,304 ANN configurations was used as a fixed-split reference benchmark to evaluate Random Search and Bayesian optimization with Tree-structured Parzen Estimator (BS-TPE) under identical validation conditions. Exploratory data analysis, multicollinearity diagnostics, outlier screening, and the Gamma Test were used to support input selection and improve dataset transparency. The Grid Search reference configuration achieved a validation MSE of 8.32 × 10⁻⁴, while BS-TPE approached a near-reference configuration after 167 evaluations, corresponding to approximately 7.3
The present research examines the feasibility of utilizing red mud in the production of interlocking bricks as an alternate to overcome industrial by-products utilization. The mix designs of the various proportions (parts by weight) of red mud (40–50), fly ash (15–25), Ground Granulated Blast Furnace Slag (10–15), cement (8–15), and M-sand (33–43), are systematically explored. Compressive strength, water absorption, bond strength, freeze-thaw, moisture resistance, efflorescence resistance, and heavy-metal leaching were tested in the research. Masonry prism is tested to ascertain the load bearing capacity. Microstructural characterization is carried out through Scanning Electron Microscopy (SEM) with Energy-Dispersive X-ray (EDX) spectroscopy. The best mix (Mix 7) indicates compressive strength of 7.9 MPa, water absorption of 12.3
Concrete compressive strength prediction is a crucial aspect of construction engineering, directly impacting the structural integrity and durability of built structures. Traditional empirical methods often struggle to capture the intricate interplay of factors influencing concrete behavior, necessitating the exploration of advanced computational techniques. In this work presents a new hybrid model that combines the Improved LinkNet Model with Neural ODE to forecast concrete compressive strength. The hybrid model uses input parameters of cement blast furnace slag fly ash water superplasticiser coarse aggregate fine aggregate and age because these factors together determine concrete compressive strength. The proposed model uses the Improved LinkNet Model and Neural ODE techniques to solve traditional empirical methods because they lack complete understanding of concrete behavior. Through the fusion of these methods, the hybrid model endeavors to deliver accurate and reliable predictions of concrete compressive strength across varying conditions and compositions. By capturing both spatial and temporal dependencies within the data, the model offers a versatile framework for concrete engineering and construction practices, contributing to the advancement of predictive modeling in the field and enhancing the safety, efficiency, and sustainability of infrastructure projects worldwide. The analysis on the MAE metric offers valuable insights into the predictive performance of the proposed model for both short-term (7 days) and long-term (28 days).
Glass powder can reduce clinker demand in concrete, but compressive strength depends on coupled effects of replacement level, water-to-binder balance, binder content, aggregate proportions, admixture dosage, and curing age. This study developed a leakage-safe grouped machine-learning workflow for predicting the compressive strength of glass powder-blended concrete from mixture composition and curing age. The cleaned dataset contained 1244 observations and 275 unique mix groups. The workflow combined grouped train-calibration-test splitting, domain-guided interpretable feature engineering, benchmarking of eight regression models, split conformal prediction, SHapley Additive exPlanations (SHAP)-based interpretation, and a Tkinter decision-support interface. Under grouped cross-validation, Extreme Gradient Boosting (XGBoost) was the best single model, with a mean root mean squared error of 6.378 MPa, a mean absolute error of 4.519 MPa, and a mean coefficient of determination of 0.799. Engineered interpretable features improved grouped CatBoost performance from a root mean squared error of 7.402 to 6.397 MPa. On the main seed-2026 final test set, the proposed stacking framework achieved R^2 = 0.845 , RMSE = 6.710 MPa, and MAE = 5.112 MPa, while providing calibrated 95 PICP_95=0.952 and MPIW_95=28.504 MPa. However, grouped cross-validation, fold-wise Wilcoxon testing, and repeated grouped-split sensitivity analysis showed that the stack was statistically comparable to XGBoost rather than clearly superior. Therefore, the proposed framework is presented as a competitive uncertainty-aware decision-support workflow rather than as a universally dominant point-prediction model. However, fold-wise Wilcoxon testing showed no statistically significant difference between the proposed stack and XGBoost, so the stacking framework is interpreted as a competitive uncertainty-aware alternative rather than a statistically superior predictor. The framework provides a practical tool for initial mix screening, batch evaluation, and uncertainty-aware decision support.
Poor geotechnical properties of dune sand (DS) pose severe structural challenges in arid regions, while the accumulation of municipal and agricultural residues necessitates sustainable management strategies. This study evaluates the valorization of waste glass powder (GP) and organic crushed date pits (DP) as eco-friendly stabilizers to optimize the mechanical performance of dune sand DS. A comprehensive experimental campaign was conducted including standard Proctor compaction, direct shear, and California Bearing Ratio (CBR) tests. Standard Proctor, direct shear, and California Bearing Ratio (CBR) tests were conducted on DS blended with GP (2
In this study, the stepped spillway with positive-sloping-steps is introduced and experimentally evaluated. This spillway is consisted of a series of consecutive positive-sloping-steps, their fronts is perpendicular to the flow direction and their floor is sloping along flow direction. Energy dissipation, flow aeration, and hydraulic parameters of hydraulic jumps at downstream of physical models from classical and suggested stepped spillways are experimentally measured and evaluated. Results indicated that due to forming of a series of consecutive sloping hydraulic jumps over the spillway steps, skimming flow regime only shapes on stepped spillway with positive-sloping-steps. Under relative critical depth range of 0.51 < yc/h < 1.16, relative step height of h/l = 0.54, and similar steps number of N = 23, the residual energy at the toe of suggested stepped spillway is averagely obtained 38.6
This study experimentally and numerically investigates the seismic performance of a Ductile Corrugated Pipe Damper (DCPD) with different corrugated wave-pitch configurations at two seismic performance levels: Immediate Occupancy (IO) and Life Safety (LS). Three damper specimens were subjected to axial cyclic loading with inner/outer wave-pitch combinations of 10/10 mm, 24/10 mm, and 10/24 mm, respectively. The experimental results demonstrated that reducing the wave pitch in both corrugated layers enhanced the seismic performance of the dampers by improving stiffness, strength, energy dissipation capacity, and buckling resistance. Among the investigated configurations, the specimen with fine wave pitches in both layers (10/10 mm) exhibited the most favorable overall seismic performance, achieving an energy absorption capacity exceeding 20,500 kN·mm and an equivalent viscous damping ratio of 40.5
Hydraulic jump conjugate depth ratio is a fundamental parameter in the design and analysis of hydraulic structures, including stilling basins, spillways, and energy dissipation systems. Accurate prediction of this parameter is therefore essential for ensuring hydraulic safety and engineering efficiency. This study develops and evaluates ten machine learning (ML) algorithms for predicting the conjugate depth ratio in horizontal rectangular channels using dimensionless variables derived from the Buckingham π theorem. The investigated models include KNN, Ada, RF, GB, ET, CB, XGB, LAS, LAR, and SVR. Hyperparameter tuning was performed using Bayesian optimization, while model performance was comprehensively evaluated using statistical metrics, graphical analyses, the objective function (OBJ), and the uncertainty coefficient (U95), and further compared with four widely used empirical equations. Among the ten evaluated models, SVR achieved the best overall performance, followed by LAR, XGB, and ET, whereas RF and CB ranked among the lowest-performing models. Using four input variables, SVR achieved MAE = 0.112, RMSE = 0.137, NSE = 0.999, MAPE = 1.43
Engineered Cementitious Composites (ECC) have become a specialised class of fibre-reinforced composites, characterised by strain-hardening behaviour, pronounced tensile ductility, and multiple cracking under load. Despite growing interest in ECC for structural and retrofit applications, the effect of binder and supplementary cementitious material (SCM) selection on early-phase strength progression and residual mechanical performance under thermal exposure remains inadequately addressed in the literature. The present study undertakes a comparative investigation of ECC mixtures prepared using two binder systems, namely Ordinary Portland Cement (OPC) and High Alumina Cement (HAC), and two SCMs, namely Ground Granulated Blast Furnace Slag (GGBS) and fly ash. Polyvinyl alcohol (PVA) fibres at a 1.5
The incorporation of waste materials in advanced cementitious composites presents a promising approach for promoting sustainable construction. This study investigates the mechanical and durability performance of Engineered Cementitious Composites (ECC) incorporating construction and demolition (C D) waste, industrial foundry waste (IFW), and ceramic waste (CER) as partial and full replacements for manufactured sand (M-sand) used as fine aggregate. ECC mixtures were prepared with replacement levels of 10
