KMITL

Permanent URI for this communityhttps://dspace.kmitl.ac.th/handle/123456789/1

Browse

Search Results

Now showing 1 - 6 of 6

Some of the metrics are blocked by your
consent settings
From Experiment to Prediction: Machine Learning Solutions for Concrete Strength Assessment with Steel Clamps
(2026-02-01)
Saingam, Panumas
;
Chatveera, Burachat
;
Sua-Iam, Gritsada
;
Chaimahawan, Preeda
;
Suthumma, Chisanuphong
This study examines the confined compressive strength (Fcc) of circular, square, and rectangular column geometries under varying confinement conditions. Results indicate that circular columns have the highest Fcc values, exceeding those of square and rectangular shapes. Increased confinement through clamps significantly enhances compressive strength. Five machine learning models, Linear Regression, Decision Tree, Random Forest, AdaBoost, and Gradient Boosting, were used to predict Fcc based on geometric and confinement parameters. Linear Regression and Decision Tree models achieved moderate predictive performance, with R2 values of 0.84 and 0.83, respectively, and relatively higher error measures (RMSE, MAE, and MAPE), indicating limited ability to capture complex nonlinear relationships in the data. In contrast, ensemble-based methods demonstrated superior performance. The Random Forest model improved the coefficient of determination to 0.90 while substantially reducing all error metrics, reflecting enhanced generalization through bagging. The boosting-based approaches yielded the best results, with AdaBoost achieving the highest R2 value of 0.99 and the lowest RMSE, MAE, and MAPE among all models, followed closely by Gradient Boosting with an R2 of 0.98. These results confirm that ensemble learning techniques, particularly boosting algorithms, yield more accurate and robust predictions than single learners for the problem studied. Data visualization techniques, including Regression Error Characteristic curves (REC) and SHapley Additive exPlanations (SHAP) value analysis, highlighted model performance and feature importance, emphasizing the roles of confinement and geometry in compressive strength. This research demonstrates the potential of machine learning to optimize structural engineering design and suggests further exploration of alternative shapes and confinement strategies to enhance structural integrity.
Some of the metrics are blocked by your
consent settings
Aircraft Wing Design Against Bird Strike Using Metaheuristics
(2025-05-01)
Timhede, Vanessa
;
Timhede, Silvia
;
Winyangkul, Seksan
;
Sleesongsom, Suwin
Bird strikes pose a significant threat to aviation safety, particularly affecting the wing structures of aircraft. This research aims to design and analyze the impact of bird strikes on wing structures using response surface method and metaheuristics (MHs), which are used to explore various risk minimization and damage mitigation techniques. The optimization problem is the minimization of the maximum von Mises stress of aircraft wing structure against bird strike that is subject to displacement and stress constraints. The design variables include skin and rib thickness, as well as sweep angle. Difficulty due to embedded bird strike simulation and optimization design can be alleviated using a response surface method (RSM). The regression technique in the RSM of the data can reach our goal of model fitting with a higher R2 until 0.9951 and 0.9919 are obtained for the displacement and von Mises stress model, respectively. The response surface function of the displacement and von Mises stress are related to skin thickness, while sweep angles rather than rib thickness have a greater impact on both design variables. The optimized design of the design variables is performed using MHs, which are TLBO, JADE, and PBIL. The comparative result of MHs can conclude that the PBIL outperformed others in all descriptive statistics. The optimized design results revealed that the optimum solution can release better energy due to bird strike with the highest limit of skin thickness, moderate rib thickness, and less than half of the sweep angle. The results are in accordance with the response surface function analysis. In conclusion, the optimized design of the aircraft wing structure against bird strike can be accomplished with our proposed technique.
Some of the metrics are blocked by your
consent settings
Machine Learning and Regression Models for Evaluating Ultimate Performance of Cotton Rope-Confined Recycled Aggregate Concrete
(2025-01-01)
Rodsin, Kittipoom
;
Ejaz, Ali
;
Wang, Huaping
;
Saingam, Panumas
;
Joyklad, Panuwat
This study investigates the use of cotton ropes (CRs) as a sustainable and cost-effective substitute for synthetic fiber-reinforced polymers for concrete confinement, offering significant environmental benefits such as lower CO<inf>2</inf> emissions and reduced energy consumption. The work evaluates the effectiveness of CR strips for confining concrete, including scenarios with recycled concrete aggregates (ReCA). Compressive strength improvements varied among specimens, with Specimen I-3F showing a 140.52% increase and Specimen II-3F achieving a 46.67% improvement. Strip configurations for Type I recycled aggregate concrete (RAC) outperformed full wraps on Type II RAC, exemplified by Specimen I-3S’s 84.51% improvement. Ultimate strain enhancements ranged from 915% to 4490.91%, driven by the significant rupture strain of cotton rope confinement. For Type I RAC, complete wrapping significantly outperformed strip configurations by 56%, 50%, and 32% in ultimate strength improvement for 1, 2, and 3 layers, respectively. The confinement ratio, varying from 0.10 to 0.70, greatly influenced the compressive behavior, with compressive strength normalized by unconfined strength increasing consistently with the confinement ratio. A minimum confinement ratio of roughly 0.40 is required to achieve an increasing second part in the compressive behavior. The initial parabolic branch was modeled using Popovics’ formulation, revealing an elastic modulus approximately 20% lower than ACI 318-19 predictions. The second branch was described using a linear approximation, and nonlinear regression analysis produced expressions for key points on the idealized compressive curve, enhancing model accuracy for CR-confined RAC. The (Formula presented.) values for the nonlinear regression analysis performed on experimental results were greater than 0.90. This study highlights the effectiveness of neural network expressions to predict the compressive strength of CR-confined concrete. A strength reduction (ratio of full wrap and strip wrap height CRs) factor of 0.67 was proposed and used for strip-wrapped specimens. It was seen that the neural network models also predicted the compressive strength of partially wrapped specimens with reasonable accuracy using the strength reduction factor.
Some of the metrics are blocked by your
consent settings
Incorporation of Steel Fibers to Enhance Performance of Sustainable Concrete Made with Waste Brick Aggregates: Experimental and Regression-Based Approaches
(2023-11-01)
Yooprasertchai, Ekkachai
;
Bahrami, Alireza
;
Saingam, Panumas
;
Hussain, Qudeer
;
Ejaz, Ali
Each year, an enormous amount of construction waste is produced worldwide. The reuse of construction waste in construction works is a sustainable solution. The present research work utilized recycled brick aggregates in the production of concrete. The resulting concrete exhibited substandard splitting tensile, flexural, and compressive properties. Steel fibers were used to improve these substandard properties of recycled brick aggregate concrete. The volume fractions of 1%, 2%, and 3% for steel fibers were mixed in concrete, whereas recycled brick aggregates were obtained from solid fired-clay bricks, hollow fired-clay bricks, and cement–clay interlocking bricks. The compressive strength was enhanced by up to 35.53% and 66.67% for natural and recycled brick aggregate concrete, respectively. Strengthened flexural specimens demonstrated up to 8765.69% increase in the energy dissipation. Specimens strengthened with steel fibers showed substantially improved splitting tensile, flexural, and compressive responses. Separate equations were proposed to predict the peak compressive strength, strain at peak compressive strength, elastic modulus, and post-peak modulus of recycled brick aggregate concrete. The proposed regression equations were utilized in combination with an existing compressive stress–strain model. A close agreement was observed between experimental and predicted compressive stress–strain curves of recycled brick aggregate concrete.
Some of the metrics are blocked by your
consent settings
Low-Cost Fiber Chopped Strand Mat Composites for Compressive Stress and Strain Enhancement of Concrete Made with Brick Waste Aggregates
(2022-11-01)
Joyklad, Panuwat
;
Saingam, Panumas
;
Ali, Nazam
;
Ejaz, Ali
;
Hussain, Qudeer
Given the excessive demolition of structures each year, the issues related to the generated structural waste are striking. Bricks being a major constituent in the construction industry, also hold a significant proportion of the construction waste generated annually. The reuse of this brick waste in new constructions is an optimal solution considering cost-effectiveness and sustainability. However, the problems related to the substandard peak stress and ultimate strain of concrete constructed with recycled brick aggregates (CRAs) limit its use in non-structural applications. The present study intends to improve the unsatisfactory mechanical characteristics of CRAs by utilizing low-cost glass fiber chopped strand mat (FCSM) sheets. The efficacy of FCSM sheets was assessed by wrapping them around CRA specimens constructed with different concrete strengths. A remarkable increase in the peak compressive stress and the ultimate strain of the CRA specimens were observed. For low, medium, and high strength CRAs, the ultimate strain improved by up to 320%, 308%, and 294%, respectively, as compared to the respective control specimens. Several existing analytical models were utilized to predict the peak compressive stress and ultimate strain of the CRAs strengthened using FCSM sheets. None of the considered models reproduced experimental results accurately. Therefore, equations were formulated using regression predicting the peak stress and ultimate strain of the CRAs confined with FCSM sheets. The predicted values were found to correlate well with the experimental values.
Some of the metrics are blocked by your
consent settings
Structural Behavior of LC-GFRP Confined Waste Aggregate Concrete Square Columns with Sharp and Round Corners
(2022-09-01)
Parichatprecha, Rattapoohm
;
Rodsin, Kittipoom
;
Chaiyasarn, Krisada
;
Ali, Nazam
;
Suthasupradit, Songsak
Reusing construction brick waste to fabricate new concrete is an economical and sustainable solution for the ever-increasing quantity of construction waste. However, the substandard mechanical properties of the concrete made using recycled crushed brick aggregates (RBAC) have limited its use mainly to non-structural applications. Several studies have shown that the axial compressive performance of the concrete is a function of the lateral confining pressure. Therefore, this study proposes to use low-cost glass fiber-reinforced polymer (LC-GFRP) wraps to improve the substandard compressive strength and ductility of RBAC. Thirty-two rectilinear RBAC specimens were constructed in this study and tested in two groups. The specimens in Group 1 were tested without the provision of a corner radius, whereas a corner radius of 26 mm was provided in the Group 2 specimens. Specimens in both groups demonstrated improved compressive behavior. However, the premature failure of LC-GFRP wraps near the sharp corners in Group 1 specimens undermined its efficacy. On the contrary, the stress concentrations were neutralized in almost all Group 2 specimens with a 26 mm corner radius, except low-strength specimen with six layers of LC-GFRP. As a result, Group 2 specimens demonstrated a more significant improvement in peak compressive strength and ultimate strain than Group 1 specimens. An analytical investigation was carried out to assess the efficiency of existing compressive stress–strain models to predict the peak compressive stress and ultimate of LC-GFRP-confined RBAC. Existing FRP models were found unreliable in predicting the key parameters in the stress–strain curves of LC-GFRP-confined RBAC. Equations were proposed by using nonlinear regression analysis, and the predicted values of the key parameters were found in good agreement with the corresponding experimental values.

KMITL

Browse

Filters

Settings

Sort By

Results per page

Search Results