An ANN-based approach for estimating load capacity and strain in FRP confined concrete columns
This study introduces an innovative hybrid predictive model utilizing artificial neural network (ANN) techniques to accurately forecast the load-carrying capacity (Pcc) and confined strain (εcc) of Polyvinyl Chloride – Carbon Fiber Reinforced Plastic (PVC-CFRP) confined concrete columns under axial...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-09-01
|
| Series: | Egyptian Informatics Journal |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S1110866525001318 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | This study introduces an innovative hybrid predictive model utilizing artificial neural network (ANN) techniques to accurately forecast the load-carrying capacity (Pcc) and confined strain (εcc) of Polyvinyl Chloride – Carbon Fiber Reinforced Plastic (PVC-CFRP) confined concrete columns under axial compression. The use of PVC-CFRP in civil engineering improves durability, strength and stiffness of structural components, and accurate prediction of these properties is needed for design and safety evaluations. Incorporating Random Forest for feature selection and Neural Network trained with the Beetle Antenna Search (BAS) algorithm, the proposed model is more precise and reliable in predicting the system response. Empirical validation was done by training the model on a dataset of 268 data points and the model achieved a test R squared (R2) of 0.971 with lower prediction errors (Root Mean Square Error (RMSE) of 25.05684, Mean Absolute Error (MAE) of 13.18642, Mean Absolute Percentage Error (MAPE) of 0.0178) than existing models. The level of accuracy in the study is high, indicating the robustness of the model and the possibility of using it in its practical engineering context. In addition, the research presents the development of a user interface platform for the easy application of the predictive model, enabling its usability by professionals in the field. The main novelty of this work is the way it tries to bridge the gap between the advancements in machine learning techniques and practical applications in engineering by giving an example of a future innovation in structural engineering analytics. In addition, this model has better predictive accuracy, yet also improves interpretability and usability, which are crucial for improving current design and assessment practice in civil engineering. |
|---|---|
| ISSN: | 1110-8665 |