Prediction of Peak Discharge for Dam Overtopping Failure Based on Physical Experimental Analysis

The increasing frequency of high-intensity rainfall events has increased the risk of dam and levee failures caused by overtopping, which can lead to severe downstream flooding. Accurate estimation of peak discharge during dam failure is therefore essential for flood hazard assessment and emergency planning. This study developed a dimensionless empirical formula for estimating peak discharge during overtopping-induced failure of homogeneous sandy embankment dams. Controlled laboratory experiments were conducted using scaled dam models in an artificial channel. A total of 155 experimental cases were analyzed, including a 128-case main experimental matrix and 27 additional cases with larger reservoir volumes. The effects of reservoir volume, downstream slope, and median particle size on breach outflow were examined. Outflow discharge was estimated from the temporal variation of upstream reservoir water level and storage balance. Dimensional analysis was applied to derive dimensionless variables, and a multivariable power-law regression model was used to develop the peak discharge formula. The proposed formula reasonably reproduced the observed dimensionless peak discharge for the present experimental dataset and was further evaluated using results from previous studies. The results indicate that the dimensionless formulation improves physical interpretability and can support preliminary peak discharge estimation within comparable overtopping-dominated failure conditions. However, further validation is required for various dam materials, geometries, and failure mechanisms.