A Pioneering Approach to Predicting the Shear Strength of RC Beams by Employing Artificial Intelligence Techniques

Civil Engineering and Architecture, Journal Year: 2023, Volume and Issue: 12(1), P. 218 - 230

Published: Dec. 7, 2023

Shear failures exhibit a brittle nature, often resulting in catastrophic collapse without sufficient advance warning or the capacity to redistribute internal stresses.Consequently, shear pose greater risk and require more attention from structural engineers.It is crucial incorporate preventive measures design avoid abrupt failures.The work presented this article attempts predict strength of reinforced concrete beams as complex engineering problem need for extra computational resources by employing capabilities Artificial Intelligence (AI) techniques.In recent decades, significant amounts research have been done on AI methods such artificial neural networks (ANNs), fuzzy logic genetic algorithms RC beams.In research, adaptive neuro-fuzzy inference system (ANFIS) ANNs are developed beams.The required data form major factors affecting lacking stirrups compressive concrete, beam depth, effective width, span-to-depth ratio, proportion longitudinal steel yield considered study.Also, context investigation, comparison was conducted between techniques ANFIS.The outcomes demonstrated that both exhibited favourable predictive capabilities.Nevertheless, ANFIS architecture proposed, which incorporates hybrid learning algorithm, outperformed multilayer feedforward ANN utilizes backpropagation algorithm.The findings indicated suitable technique predicting intricate relationships input output parameters, thus making it valuable tool beams.

Language: Английский

---

The Efficiency of Using Machine Learning Techniques in Fiber-Reinforced-Polymer Applications in Structural Engineering

Sustainability, Journal Year: 2023, Volume and Issue: 16(1), P. 11 - 11

Published: Dec. 19, 2023

Sustainable solutions in the building construction industry have emerged as a new method for retrofitting applications last two decades. Fiber-reinforced polymers (FRPs) garnered much attention among researchers improving reinforced concrete (RC) structures. The existing design guidelines FRP-strengthened RC members were developed using empirical methods that are based on specific databases, limiting accuracy of predicted results. Therefore, use innovative and efficient prediction tools to predict behavior has become essential. During few years, efforts been progressively focused machine learning (ML) feasible effective technique solving various structural engineering problems. Its capability complex nonlinear systems while considering wide range parameters offers distinctive opportunity make more predictable accurate. This paper aims evaluate current state ML algorithms strengthened with FRP enable determine capabilities well find research gaps carry out bridge revealed knowledge practice gaps. Scopus databases searched predefined standards. search ninety-six articles published between 2016 2023. Consequently, these analyzed field retrofitting, including flexural shear strengthening beams, slabs, confinement compressive strength columns, bond strength. results reveal 32% reviewed studies application techniques 6.5% 22% strength, materials, 1% beam–column joints. also shown significant improvement resistance compared solutions. Supervised most favorable due their good generalization, interpretability, adaptability, predictive efficiency. In addition, selection suitable optimization is found be mainly dictated by nature problem characteristics dataset. Nonetheless, selecting appropriate model algorithm each remains challenge, given different principles methodologies.

Language: Английский

---

Hybrid Machine Learning Algorithms for Prediction of Failure Modes and Punching Resistance in Slab-Column Connections with Shear Reinforcement

Buildings, Journal Year: 2024, Volume and Issue: 14(5), P. 1247 - 1247

Published: April 28, 2024

This study presents a data-driven model for identifying failure modes (FMs) and predicting the corresponding punching shear resistance of slab-column connections with reinforcement. An experimental database that contains 328 test results is used to determine nine input variables based on mechanism. A comparison conducted between three typical machine learning (ML) approaches: random forest (RF), light gradient boosting (LightGBM), extreme (XGBoost) two hybrid optimized algorithms: grey wolf optimization (GWO) whale algorithm (WOA). It was found XGBoost classifier had highest accuracy rate, precision, recall values FM identification. In testing, WOA-XGBoost has best in resistance, R2, MAE, RMSE 0.9642, 0.087 MN, 0.126 respectively. However, calculated derived from classical analytical methods clearly demonstrates existing design codes need be improved. Additionally, Shapley additive explanations (SHAP) were applied explain model’s predictions, factors categorized according their impact resistance. By modifying these parameters, can improved while reducing unpredictable failure. With proposed algorithms, it possible slabs during preliminary stages construction.

Language: Английский

---

Hybrid Optimized Algorithms for Predicting Punching Shear Strength in Flat Slabs Considering Failure modes

KSCE Journal of Civil Engineering, Journal Year: 2024, Volume and Issue: unknown, P. 100079 - 100079

Published: Oct. 1, 2024

Language: Английский

---

Failure mode classification for hybrid FRP/steel reinforced concrete beams: a soft computing concept based on the numerical model

Innovative Infrastructure Solutions, Journal Year: 2024, Volume and Issue: 9(8)

Published: July 5, 2024

Language: Английский

---