Buy article PDF
The purchased file will be sent to you
via email after the payment is completed.
US$ 35
|
Structural Engineering and Mechanics Volume 93, Number 2, January25 2025 , pages 83-94 DOI: https://doi.org/10.12989/sem.2025.93.2.083 |
|
|
|
A novel approach combining theoretical and ANN for predicting interfacial stress in FRP-reinforced damaged RC beams |
||
Rezki Amara, Mokhtar Nebab, Hassen Ait Atmane, Zahira Sadoun, Lazreg Hadji and Riadh Bennai
|
||
| Abstract | ||
| In this research article, we present a comprehensive solution utilizing both linear elastic theory and Artificial Neural Network (ANN) methods to analyze interfacial stress in simply supported beams strengthened with bonded fiber-reinforced polymer (FRP) plates. The study investigates FRP-reinforced damaged RC beams under uniformly distributed loads. In our analytical approach, we account for adherend shear deformations in theoretical analyses by assuming a linear shear stress throughout the thickness of the adherends—a consideration often overlooked in existing solutions. Subsequently, the study validates and compares the results obtained from this approach with those available in the literature. Simultaneously, the ANN technique is employed to predict normal and shear stresses in concrete beams strengthened with FRP plates. In the architecture of the artificial neural network (ANN), the initial layer serves as the input layer with 15 inputs, followed by a hidden layer consisting of 18 neurons,and finally, two output layers. Within the hidden layer, the activation function utilized is the Transg function. This prediction relies on a dataset comprising over 3339 data points from the current analytical approach and literature sources, including analytical and finite element methods presented in this paper and others from existing literature. The ANN technique systematically explores various parameters, such as material characteristics, properties, and geometric details of RC beams and FRP plates. Through both the ANN method and computational analysis, the study establishes the significant influence of FRP plates on shear and normal stress. The ANN model showcases robust capabilities in handling extensive datasets and various critical parameters. | ||
| Key Words | ||
| analytical approach; composite plates; interfacial stress; machine learning; RC beams; regression problems | ||
| Address | ||
| Rezki Amara: Laboratory of Structures, Geotechnics and Risks, Department of Civil Engineering, Hassiba Beénbouali University of Chlef, Algeria Mokhtar Nebab: Laboratory of Structures, Geotechnics and Risks, Department of Civil Engineering, Hassiba Beénbouali University of Chlef, Algeria; Department of Civil Engineering, Faculty of Technology, University of M'Hamed BOUGARA Boumerdes, Algeria Hassen Ait Atmane: Laboratory of Structures, Geotechnics and Risks, Department of Civil Engineering, Hassiba Beénbouali University of Chlef, Algeria Zahira Sadoun: Laboratory of Structures, Geotechnics and Risks, Department of Civil Engineering, Hassiba Beénbouali University of Chlef, Algeria Lazreg Hadji: Department of Civil Engineering, University of Tiaret, 14000, Algeria Riadh Bennai: Laboratory of Structures, Geotechnics and Risks, Department of Civil Engineering, Hassiba Beénbouali University of Chlef, Algeria | ||