Work

Support vector machines and gene expression programming for discharge coefficient prediction in triangular labyrinth weirs

Year: 2026
Type: article
Abstract: Labyrinth weirs significantly enhance flow capacity through extended crest length, making them vital for water management infrastructure. This study presents an experimental and computational investigation of triangular labyrinth weirs with varying sidewall angles (15°, 25°, 30°, 40°) and cycle numbers (3, 4, 5). Using 195 experimental data points, predictive models for the discharge coefficient ( C d ) were developed using Support Vector Regression (SVR) and Gene Expression Programming (GEP). The SVR model ( R 2 = 0.96, MAPE = 1.14%) achieved superior prediction accuracy by incorporating four dimensionless parameters ( h / p , α , w / p , and N ), whereas the GEP model produced an explicit predictive equation (MAPE = 5.68%) suitable for practical engineering applications. Unlike previous studies that mainly investigated single-cycle triangular labyrinth weirs, the present study employs a comprehensive multi-cycle experimental dataset comprising three, four, and five cycle configurations and explicitly incorporates the cycle number ( N ) as an input variable. The comparison demonstrates that SVR is more suitable for accurate hydraulic prediction, whereas GEP provides a simple and interpretable equation for rapid field estimation where computational simplicity is preferred.
Language: English
Open Access status: hybrid
Estimated APC paid: $1,390
OpenAlex ID: W7233769426
FWCI:
Cites: 35
Cited by:
Related to: 10