Abstract
This study was carried out to investigate the regular wave run-up phenomenon on smooth slopes of perforated coastal structures constructed on sloping beaches and the various parameters that affect wave run-up. Experiments were conducted using various hydraulic and structural parameters. The relative coastal structure distance, relative depth, relative wave height, beach slope, coastal structure inclination, and surf similarity parameter were found to be positively correlated to the relative wave run-up. The wave steepness and coastal structure perforation percentage were found to be negatively correlated to the relative wave run-up. The results also show that the coastal structure perforation percentage plays a dominant role in the attenuation of short waves but a less significant role in the attenuation of long waves. The quantitative analyses were performed using multiple additive regression trees (MART) and multilayer perceptron neural networks (MLP) methods. The results indicate that the MART method׳s prediction accuracy and avoidance of over-fitting were superior to those of the MLP method. The percentage improvement in the root mean square error of the MART model over the MLP model in predicting relative wave run-up was 57.56%. The analysis results suggest that MART-based modeling is effective in predicting wave run-up.
•The wave run-up on smooth slopes of perforated coastal structures on sloping beaches is studied.•Experiments were achieved using various hydraulic and structural parameters.•The structure perforation has a noticeable effect especially for small wave steepness.•The analyses were performed using two different data mining methods.•The improvement in the RMSE of the MART model over the MLP model in predicting wave run-up was 57.56%.