Strength prediction and failure mode classification for SRC shear beams using GA-BP ANN method
For the steel reinforced concrete (SRC) beam, accurately predicting its shear behavior can be quite challenging. Considering the advantages of machine-learning (ML) approaches, the back-propagation (BP) artificial neural network (ANN) method combined with genetic algorithm (GA) was employed to the p...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-07-01
|
| Series: | Case Studies in Construction Materials |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S221450952500052X |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832592791115399168 |
|---|---|
| author | Gangfeng Yao Bingyi Li |
| author_facet | Gangfeng Yao Bingyi Li |
| author_sort | Gangfeng Yao |
| collection | DOAJ |
| description | For the steel reinforced concrete (SRC) beam, accurately predicting its shear behavior can be quite challenging. Considering the advantages of machine-learning (ML) approaches, the back-propagation (BP) artificial neural network (ANN) method combined with genetic algorithm (GA) was employed to the prediction of strength and failure mode of SRC shear beams. The parameters considered in this study are shear span-to-effective depth ratio, axial concrete compressive strength, ratio of stirrup area, ratio of steel-web area, tension rebar ratio, yield strengths of various steels and a coefficient introduced to consider the concrete-confined effect. To gain the best ANN model, the optimum input combination, hidden-nodes number, weights and biases were firstly researched and determined, based on a newly-built database of 130 experimental specimens. Then, the importance of input variables were analyzed and the most recommended models were provided. Finally, a comparative study was also conducted between the ANN models and literature methods. Results indicate that the shear span-to-effective depth ratio, axial concrete compressive strength and ratio of steel-web area are the three most important variables to shear strength. As for the failure mode, the shear span-to-effective depth ratio is the most influential factor. Besides, the coefficient account for concrete-confined effect enhances the predictive accuracies of ANN models for shear capacity. Compared to the literature methods, the ANN models show a much better performance both in predicting shear strength and failure mode. |
| format | Article |
| id | doaj-art-6afebd3c20814a708e82c4b0bccd436a |
| institution | Kabale University |
| issn | 2214-5095 |
| language | English |
| publishDate | 2025-07-01 |
| publisher | Elsevier |
| record_format | Article |
| series | Case Studies in Construction Materials |
| spelling | doaj-art-6afebd3c20814a708e82c4b0bccd436a2025-01-21T04:13:07ZengElsevierCase Studies in Construction Materials2214-50952025-07-0122e04253Strength prediction and failure mode classification for SRC shear beams using GA-BP ANN methodGangfeng Yao0Bingyi Li1School of Civil Engineering, Suzhou University of Science and Technology, Suzhou 215011, China; Key Laboratory of Structural Engineering of Jiangsu Province, Suzhou University of Science and Technology, Suzhou 215011, ChinaSchool of Civil Engineering, Suzhou University of Science and Technology, Suzhou 215011, China; School of Transportation, Southeast University, Nanjing 211189, China; Corresponding author at: School of Civil Engineering, Suzhou University of Science and Technology, Suzhou 215011, China.For the steel reinforced concrete (SRC) beam, accurately predicting its shear behavior can be quite challenging. Considering the advantages of machine-learning (ML) approaches, the back-propagation (BP) artificial neural network (ANN) method combined with genetic algorithm (GA) was employed to the prediction of strength and failure mode of SRC shear beams. The parameters considered in this study are shear span-to-effective depth ratio, axial concrete compressive strength, ratio of stirrup area, ratio of steel-web area, tension rebar ratio, yield strengths of various steels and a coefficient introduced to consider the concrete-confined effect. To gain the best ANN model, the optimum input combination, hidden-nodes number, weights and biases were firstly researched and determined, based on a newly-built database of 130 experimental specimens. Then, the importance of input variables were analyzed and the most recommended models were provided. Finally, a comparative study was also conducted between the ANN models and literature methods. Results indicate that the shear span-to-effective depth ratio, axial concrete compressive strength and ratio of steel-web area are the three most important variables to shear strength. As for the failure mode, the shear span-to-effective depth ratio is the most influential factor. Besides, the coefficient account for concrete-confined effect enhances the predictive accuracies of ANN models for shear capacity. Compared to the literature methods, the ANN models show a much better performance both in predicting shear strength and failure mode.http://www.sciencedirect.com/science/article/pii/S221450952500052XSRC beamArtificial neural networkGenetic algorithmShear strengthFailure mode |
| spellingShingle | Gangfeng Yao Bingyi Li Strength prediction and failure mode classification for SRC shear beams using GA-BP ANN method Case Studies in Construction Materials SRC beam Artificial neural network Genetic algorithm Shear strength Failure mode |
| title | Strength prediction and failure mode classification for SRC shear beams using GA-BP ANN method |
| title_full | Strength prediction and failure mode classification for SRC shear beams using GA-BP ANN method |
| title_fullStr | Strength prediction and failure mode classification for SRC shear beams using GA-BP ANN method |
| title_full_unstemmed | Strength prediction and failure mode classification for SRC shear beams using GA-BP ANN method |
| title_short | Strength prediction and failure mode classification for SRC shear beams using GA-BP ANN method |
| title_sort | strength prediction and failure mode classification for src shear beams using ga bp ann method |
| topic | SRC beam Artificial neural network Genetic algorithm Shear strength Failure mode |
| url | http://www.sciencedirect.com/science/article/pii/S221450952500052X |
| work_keys_str_mv | AT gangfengyao strengthpredictionandfailuremodeclassificationforsrcshearbeamsusinggabpannmethod AT bingyili strengthpredictionandfailuremodeclassificationforsrcshearbeamsusinggabpannmethod |