Predicting the Discharge Coefficient of Rectangular Broad-Crested Porous Weirs Using the SVR Model

Esmaili, Seyed Alireza; Heydari, Majid; Gohari, Saeed

doi:10.22048/wss.2026.591151.1034

Predicting the Discharge Coefficient of Rectangular Broad-Crested Porous Weirs Using the SVR Model

Articles in Press

Document Type : Research Article

Authors

Seyed Alireza Esmaili ¹

Majid Heydari ²

Saeed Gohari ²

¹ Ph.D. Candidate, Department of Water Science and Engineering, Water Structures, Bu-Ali Sina University, Hamedan, Iran

² Department of Water Science and Engineering, Faculty of Agriculture, Bu-Ali Sina University, Hamedan, Iran

10.22048/wss.2026.591151.1034

Abstract

Rectangular broad‑crested gabion weirs are widely recognized as flow‑control structures used for erosion mitigation, upstream channel‑bed stabilization, and energy dissipation in hydraulic and river engineering. However, predicting their discharge coefficient (C_d) remains challenging due to the complex interplay of structural geometry, substrate grain‑size distribution, and porosity. This study aimed to predict the discharge coefficient of rectangular broad‑crested gabion weirs using a support vector regression (SVR) model. To this end, a laboratory experimental dataset comprising gabion weirs of various geometries tested under free‑flow conditions was compiled and employed. The SVR model was implemented in Python and trained and validated across several train:test partitions. Evaluation of the model using multiple statistical metrics indicated that allocating 80% of the data to training and 20% to testing markedly enhanced predictive performance. Specifically, the NSE and RMSE attained values of 0.91 and 0.128, respectively, thereby confirming the high accuracy of the SVR model in predicting the discharge coefficient of gabion weirs under free‑flow conditions.

Keywords

Gabion Weir

Erosion

SVR

Free Flow

Python

Subjects