Eine Plattform für die Wissenschaft: Bauingenieurwesen, Architektur und Urbanistik
Eine Plattform für die Wissenschaft: Bauingenieurwesen, Architektur und Urbanistik
Modeling vibration and crack behavior of reinforced concrete beams: developing artificial neural network predictive models
A numerical method is presented for modeling the crack in the opening mode, and the vibration of cracked RC beams with composite sheets is analyzed with the aid of FEM. Through the modification of the second surface moment in two complete and cracked sections are examined in the Euler–Bernoulli beam analysis. Using continuity conditions at the crack location, the equations of two microelements of the cracked element are related, and a torsion spring simulates the crack in this study. Several factors determine stiffness, including reinforcement place, where composite sheet is placed, and how deep the crack is. As a result of applying composite sheets, steel reinforcements, and cracks to the equations, the stiffness and mass matrices are modified. Vibration analysis is implemented using these improved matrices to determine the natural frequency (NF) of beams. To ensure the results are correct and accurate, Abaqus performs a comprehensive analysis. According to the presented method, reinforced concrete structures that are crack-resistant can be analyzed using the obtained results. A machine learning algorithm is proposed to predict NF, and the results were very promising. The model had a mean absolute error (MAE) of 0.91%. The model uses hyperparameter tuning to optimize the specifications of the artificial neural network. A comparison of the computational costs of the modeling and predictive model is presented.
Modeling vibration and crack behavior of reinforced concrete beams: developing artificial neural network predictive models
A numerical method is presented for modeling the crack in the opening mode, and the vibration of cracked RC beams with composite sheets is analyzed with the aid of FEM. Through the modification of the second surface moment in two complete and cracked sections are examined in the Euler–Bernoulli beam analysis. Using continuity conditions at the crack location, the equations of two microelements of the cracked element are related, and a torsion spring simulates the crack in this study. Several factors determine stiffness, including reinforcement place, where composite sheet is placed, and how deep the crack is. As a result of applying composite sheets, steel reinforcements, and cracks to the equations, the stiffness and mass matrices are modified. Vibration analysis is implemented using these improved matrices to determine the natural frequency (NF) of beams. To ensure the results are correct and accurate, Abaqus performs a comprehensive analysis. According to the presented method, reinforced concrete structures that are crack-resistant can be analyzed using the obtained results. A machine learning algorithm is proposed to predict NF, and the results were very promising. The model had a mean absolute error (MAE) of 0.91%. The model uses hyperparameter tuning to optimize the specifications of the artificial neural network. A comparison of the computational costs of the modeling and predictive model is presented.
Modeling vibration and crack behavior of reinforced concrete beams: developing artificial neural network predictive models
Asian J Civ Eng
Khateeb, Ahmed H. (Autor:in) / Abdulwahed, Larah R. (Autor:in) / Mohammed, Aymen R. (Autor:in)
Asian Journal of Civil Engineering ; 25 ; 129-139
01.01.2024
11 pages
Aufsatz (Zeitschrift)
Elektronische Ressource
Englisch
Artificial neural network model for fibre reinforced polymer laminated reinforced concrete beams
| British Library Online Contents | 2010
Artificial neural network model for fibre reinforced polymer laminated reinforced concrete beams
| Online Contents | 2009
Shear Resistance Prediction of Post-fire Reinforced Concrete Beams Using Artificial Neural Network
| Springer Verlag | 2019