Asian Journal of Civil Engineering最新文献

An isolation system must withstand the vertical forces generated by the weight of the superstructure and seismic loads. In addition, to facilitate movement at the isolation level and prevent potential damage, restricting the displacement of isolators should always be considered. This research focuses on an RC (reinforced concrete) dual wall-frame structure and aims to differentiate the behavioral characteristics of lead rubber bearings under columns and shear walls using non-linear analyses. Suggestions are provided to improve the design of isolators in both near and far-field regions. The results indicate that the maximum pressure exerted on all isolators during an earthquake is greater than the results obtained from linear analysis. Furthermore, observations show that the lateral displacement of the isolation system surpasses design code limits. Therefore, to address these challenges, design modifications are proposed: increasing isolator diameters by 4–37% based on their position beneath the structure and seismic region, amplifying design displacements by 20% in far-field and 50% in near-field regions, and incorporating a larger lead core (approximately (frac{1}{ 3}) to (frac{2}{ 7}) of the overall isolator diameter) to control the shear strain and lateral displacement of the isolation system. These results underscore the necessity of nonlinear analysis in isolator design for RC dual systems to ensure compliance with safety and serviceability requirements under diverse seismic conditions.

Reinforced concrete deep beams (RCDBs) exhibit complex nonlinear shear behaviour, making accurate strength prediction challenging for traditional models. This study presents a data-driven approach that combines k-fold cross-validated stepwise regression with a graphical user interface (GUI) for predicting RCDB shear strength. Using a dataset of 789 experimental cases, both linear and polynomial stepwise models were developed. The polynomial model (SPR) outperformed prior models, achieving an R² of 0.964, with effective depth and beam width identified as key influencing factors. Comparative analysis using Taylor diagrams, violin plots, and error thresholds confirmed SPR’s superior predictive accuracy, with 100% of predictions falling within 13% error. The interactive GUI enables users to adjust input parameters and visualize results, bridging analytical rigor with engineering usability. Finally, the proposed model provides a practical and accurate tool for RCDB shear prediction.

The Modified Adaptive Weight Approach (MAWA) represents a straightforward method commonly applied to solve time–cost optimization problems. Such problems are generally treated as multi-objective optimization problems, for which metaheuristic algorithms are frequently utilized. These algorithms operate on a population of potential solutions that are randomly initialized within the boundaries of the solution space. MAWA assigns uniform weight factors to all individuals in the population without accounting for their specific characteristics. However, each solution possesses unique fitness attributes relative to its position in the solution space. In this study, a multi-objective optimization model combining the Jaya with the MAWA is introduced to generate a set of Pareto-optimal solutions. Two construction project case studies, drawn from existing technical literature and comprising 18 and 29 activities, are analyzed to evaluate the effectiveness of the proposed MAWA-Jaya method. The results are benchmarked against those from previously established models that produced approximate Pareto fronts or near-optimal solutions. The findings demonstrate that the MAWA-Jaya algorithm performs efficiently in addressing time-cost trade-off problems (TCTP) within the field of construction engineering and management.

This study aims to develop precise and interpretable predictive models for estimating the compressive strength (CS) of Rice Husk Ash (RH Ash) based concrete through the application of symbolic machine learning techniques. Given the increasing emphasis on sustainable construction materials, it is essential that predictive models are both accurate and explainable. In this research, Gene Expression Programming (GEP) and Multi-Expression Programming (MEP) algorithms were employed using six critical input variables: cement, RH Ash, water, superplasticizer, age, and fine aggregate. The GEP model achieved R² values of 0.87 (training), 0.91 (testing), and 0.84 (validation), whereas the MEP model demonstrated superior performance with R² = 0.93, RMSE = 4.73 MPa, and MAE = 3.88 MPa. To enhance model transparency, SHAP (SHapley Additive exPlanations) analysis was conducted. Cement (mean SHAP value ≈ 0.60) and specimen age (≈ 0.52) were identified as the most influential predictors of CS. Water ( ≈ − 0.48) consistently exhibited a negative contribution, while RH Ash demonstrated an optimal non-linear influence (≈ 0.41), underscoring the importance of dosage control. Fine aggregate and superplasticizer exhibited lower contributions (≈ 0.28 and ≈ 0.21, respectively). The integration of symbolic machine learning and SHAP-based interpretation not only enhances predictive capability but also provides engineering insights for mix design optimization. This research will contribute to the development of performance-based design frameworks for sustainable concrete, offering a valuable tool for future research and construction industry applications involving industrial by-products such as Rice Husk Ash.

Structural health monitoring of reinforced concrete (RC) structures is critical, as their durability and safety are often compromised due to cracks caused by mechanical loads and environmental exposure. Traditional crack inspection techniques are labor-intensive, subjective, and inadequate for large-scale applications. This study addresses these limitations by leveraging deep learning, specifically Convolutional Neural Networks (CNNs), for automated crack classification and segmentation in RC beams. An experimental setup was designed involving the casting and flexural testing of RC beams with varying reinforcement configurations to simulate real-world cracking scenarios. From this experimental work, a high-resolution image dataset was created, comprising categorized images of undamaged, cracked, and severely damaged beam surfaces. Four pre-trained CNN architectures AlexNet, VGG-16, ResNet-101, and GoogleNet were fine-tuned using MATLAB’s Deep Learning Toolbox and evaluated for their performance in crack classification and segmentation. Advanced image preprocessing and segmentation techniques were employed to enhance crack visibility and model accuracy. Evaluation was conducted using a confusion matrix including precision, recall, and F1-score were recorded. Among all models, VGG-16 exhibited the highest overall performance, making it the most effective for accurate crack detection. These findings underscore the potential of CNN-based methods for real-time, scalable, and reliable structural health monitoring, paving the way for intelligent infrastructure maintenance solutions.

Such emerging demands for an environment-friendly form of construction materials led to the development of the so-called alternative construction material-Geopolymer Concrete (GPC), to overcome the current limitation of its conventional counterpart-the Portland cement concrete. However, such a developed mix design, coupled with vulnerability towards environmental aspects (including exposure to sulphates and chlorides), requires much more advanced techniques for predicting such performances. The proposed research introduces an Integrated Predictive Framework (IPF) to optimize GPC mix design. The combined power of Linear Regression (LR), Random Forest (RF), Gradient Boosting Machines (GBM), and Deep Neural Networks (DNNs) is exploited for enhancing the accuracy of predictions made on the mechanical properties of GPC, specifically on its resistance to sulphate and chloride attacks. Further model performance improvement is achieved using a new hybrid technique called Predator-Annealing Optimization (PAO), which will be used to optimize the hyperparameters. In PAO, the Marine Predators Algorithm (MPA) and Simulated Annealing Optimization (SAO) are fused to effectively navigate the search space and improve the model’s accuracy. Experimental results show that PAO-based IPF performs excellently and surpasses traditional models. Its R² value is about 0.98, while the Root Mean Square Error (RMSE) is about 0.015, and the Mean Absolute Error (MAE) is about 0.01, showing strong predictive accuracy and robustness. This work emphasizes the efficiency of integrating machine learning (ML) and hybrid optimization methods in predicting and optimizing GPC performance. It also represents a practical tool for designing sustainable and durable mixes to construct structures in sulphate and chloride environments.

The increasing demand for vertical development in urban areas has led to a surge in high-rise construction, particularly in seismically active regions. These structures are especially vulnerable to earthquake-induced vibrations due to their height and mass distribution. Hybrid seismic control systems, combining Tuned Mass Dampers (TMDs) and Base Isolation (BI), offer promising solutions to enhance structural resilience. However, the influence of TMD placement and the base isolation on different structural systems requires deeper investigation. This study focuses on determining the optimal positions of TMDs within the structure and evaluating their effectiveness using MATLAB-based analysis. Reinforced concrete (RC) and steel high-rise buildings equipped with TMDs at various locations, along with Lead Rubber Bearings (LRBs), are modeled and analysed. A 20-storey G + 19 model is developed in ETABS, and seismic response is assessed using the Response Spectrum Method in accordance with IS 1893 Part 1 and Part 6. Key performance indicators such as top-storey displacement, inter-storey drift, and base shear are examined across configurations. Results reveal that both TMD placement and base isolation significantly affect seismic performance, with hybrid systems achieving up to 65% reduction in critical response parameters. The findings offer practical guidance for optimizing seismic design of RC and steel high-rise buildings.

This study presents a nonlinear seismic fragility assessment of masonry-infilled reinforced concrete (RC) frame structures, focusing on the influence of varying infill percentages on seismic performance. Although masonry infill walls are often treated as non-structural elements, they can significantly alter the lateral stiffness, base shear capacity, and damage response of RC buildings. Numerical models were developed using equivalent diagonal strut representations for infill walls, and nonlinear static (pushover) analyses were performed to evaluate story-level displacements and strength. Results show that masonry infill reduces the fundamental period by up to 68% and lateral displacements by up to 80%, thereby enhancing seismic resistance. However, the increased stiffness also leads to higher axial and shear demands on structural elements. Fragility curves were generated to estimate the probability of damage across various limit states, indicating that buildings with higher infill ratios perform better in early damage stages but may fail abruptly once infills crack or crush. These findings highlight the dual role of infill walls in improving strength while introducing design challenges. The study offers practical insights for seismic design and retrofitting of infilled RC frames in earthquake-prone regions.

Accurate and efficient control of structural response during seismic events remains a critical challenge in structural engineering. This study proposes a novel active control algorithm based on the neural network for real time seismic response mitigation in a single-storey building frame. The proposed algorithm is developed on the assumption that the shape of the time history of the control force is similar to the shape of the time history of the earthquake. A state-space formulation of the dynamic equilibrium equations is used to compute structural responses under seismic excitation. Synthetic ground motions compatible with the response spectra for seismic Zones III, IV, and V, as defined in IS 1893:2016, are generated and used to train the neural network. The neural network is trained offline using synthetic ground motion and target responses as input, with required control force as output. Once trained, the neural network is deployed in an online simulation to generate real-time control forces aimed at achieving predefined target reductions. The performance of the proposed algorithm is also evaluated under various time-step delays to assess its stability in real-time conditions. Results show that the proposed algorithm not only achieves but consistently exceeds the predefined target reduction levels. Time-delay analysis further confirms the stability and robustness of the control strategy under implementation constraints. This approach offers a scalable pathway toward intelligent, adaptive structural control systems for seismic risk mitigation.

Concrete plays an important role in construction; however, conventional mixtures often face challenges related to strength and durability. This study investigated the impact of partially substituting ground granulated blast furnace slag (GGBFS) with varying amounts of tamarind seed polysaccharide (TSP), a natural biopolymer, on the performance of M30 grade concrete. Experiments were conducted to assess the compressive strength, elastic modulus, and flexural strength at 7, 14, and 28 days intervals. The results showed an increase in compressive strength over time, with Sample 3 reaching 26.9 MPa at 7 days, 33.2 MPa at 14 days, and 35.2 MPa at 28 days. The simulated outcomes were comparable, forecasting 37.1 MPa at 28 days for the same sample. The beam deflections under load were nearly identical in both the experimental and simulated scenarios, differing by less than 0.002 mm, thus validating the accuracy of the simulations. Machine learning models, such as Random Forest, XGBoost, and SVM, were trained on the data to predict the mechanical properties, with Random Forest demonstrating superior performance, achieving an R² of 0.99 and an MAE as low as 0.25 in strength prediction. This holistic approach, which combines experimental, computational, and AI techniques, highlights the potential of TSP and GGBFS for developing sustainable concrete mixes with enhanced mechanical properties, thus supporting more environmentally friendly construction practices.