|
Survey the suitable approach to predict the local scour depth around a bridge pier |
|---|---|
| รหัสดีโอไอ | |
| Creator | Waqed H. Hassan |
| Title | Survey the suitable approach to predict the local scour depth around a bridge pier |
| Contributor | Waqed H. Hassan, Halah K. Jalal |
| Publisher | Asia-Pacific Journal of Science and Technology |
| Publication Year | 2564 |
| Journal Title | Asia-Pacific Journal of Science and Technology |
| Journal Vol. | 26 |
| Journal No. | 2 |
| Page no. | 14-Jan |
| Keyword | Scour depth prediction, Gene expression programming (GEP), Artificial neural network (ANN) |
| URL Website | https://www.tci-thaijo.org/index.php/APST |
| Website title | https://so01.tci-thaijo.org/index.php/APST/article/view/245177/167330 |
| ISSN | 2539-6293 |
| Abstract | Local scouring around the piers of bridges is one of the main reasons for bridge failure. The main purpose of this study was to survey the suitable empirical equation to predict local scour depth around pier bridges, by applying various approach techniques to achieve more effective predictions. These approaches include gene expression programming (GEP), artificial neural networks (ANN) and statistic non-linear regression (NLR) methods. The empirical equations derived were based on shape of pier, intensity of flow, flow depth ratio, pier width ratio and attack angle. The data set, a total of 729 data points obtained from numerical simulations using Flow-3D, were divided into training and validation datasets (test). A functional relationship was created using GEP, its performance compared to ANN and NLR. Identification of the best techniques to predict scour depth, was achieved using three statistical parameters: R2, RMSE and MAE. The equation obtained using GEP, performed better than the conventional regression NLR model, but slightly poorer than that of ANN (R2 = 0.89, RMSE=0.152 and MAE= 0.118). Even though ANN performed better than GEP (R2 = 0.93, RMSE=0.129 and MAE=0.088), the latter is preferred because of its ability to provide compressed and explicit arithmetic expressions. The GEP model equation has been verified with laboratory data, predicting a good result. |
