The trend for utilizing sustainable products is becoming increasingly important in the construction industry. Precast concrete sandwich systems containing expanded polystyrene (EPS) is gradually replacing the conventional blockwork systems due to its lightweight, enhanced insulation properties and rapid installation. Accordingly, there is a need to optimize the design and production of this system to provide superior insulation, durability and rapid installation while ensuring adequate bonding, strength and mechanical properties to sufficiently fit its purpose. Research has shown that the system’s technical properties are highly dependent on the mix design and the production and installation methods. This study involved the design and production of a sustainable lightweight precast concrete sandwich panel for non-load bearing partition wall systems. An experimental program was developed to identify the optimum mixture proportions for the EPS concrete core. Practical procedures for system production and installation were proposed. Physical and mechanical properties for the proposed EPS concrete system were determined based on experimental testing and calculations. A relationship between the plastic density of fresh EPS concrete and its compressive strength was established to ensure quality control before concrete is cast. Results revealed a high degree of correlation of 0.97 between the core density and compressive strength. Technical comparison revealed that the proposed system is superior in comparison to the conventional lightweight blockwork system. Furthermore, the installation time for the proposed partition wall system was 3 times faster in comparison with the conventional block work system, which would lead to significant reduction in the total cost.
This paper proposes a mathematical model approach to predict the Ultrasonic Pulse Velocity (UPV) of concrete. This paper presents the formulation, calibration, evaluation, and validation of the proposed UPV model. An experimental program was developed to calibrate and evaluate the proposed model. Furthermore, the model was validated using a separate group of measurements in the experimental program, to test its accuracy and generalization capability. Analysis of the results reveals that the proposed model provides a good fit to the experimental data and does not contain outliers or discerning pattern. The corresponding standard error and determination coefficient are 45 m/s and 0.92, respectively. Model validation demonstrates high predictability and generalization capability. The corresponding standard error and determination coefficient are 74 m/s and 0.91, respectively.
8 December 2025 in Engineering, Technology & Applied Science Research
This paper presents an improved machine learning approach to predict the compressive strength of concrete from nondestructive Rebound Hammer (RH) measurements and water-to-cement (W/C) ratio. Several regression models, including Linear Regression (LR), Decision Tree (DT), Support Vector Regression (SVR), Random Forest (RF), and Gradient Boosting (GB) were initially applied and experimented on a comprehensive experimental dataset. GB achieved good baseline accuracy with R² = 0.86 and MAE = 6.34 MPa. To improve performance, hyperparameter tuning using grid search was adopted, and the optimized GB model improved accuracy with R² = 0.8704 and MAE = 5.90 MPa. Other models such as XGBoost, deep neural networks, and an ensemble averaging model were then explored. Among these, XGBoost had the best overall performance of R² = 0.8735 and MAE = 5.85 MPa, with the tuned GB coming close. A second-order polynomial regression model was further derived from the XGBoost predictions to provide a reference equation. This polynomial equation presents a simple and comprehensible method for field engineers to estimate compressive strength from RH and W/C values alone, without the need for computers. To support practical deployment, a user-friendly application was developed using Streamlit, which enabled users to estimate concrete strength in a real-time interface. This app uses the XGBoost model and allows for fast, portable, and accurate predictions in the field. Overall, this work demonstrates the value of combining domain knowledge with modern data-driven techniques to improve the accuracy, interpretability, and usability of nondestructive testing in concrete evaluation. The proposed models and tools offer practical benefits for real-time, reliable estimation that bridge the gap between conventional field tests and intelligent predictive analytics.