Prediction of Wave Reflection from Berm Breakwaters, Part A: Presenting a New Formula

Wave reflection is an important hydrodynamic parameter in the design procedure of berm breakwaters. The existing formulae for predicting the wave reflection from berm breakwater are mainly based on regression model of available experimental data. Recently, applications of soft computing approaches and data mining techniques in tackling coastal engineering related problems received considerable attention. In this paper, accuracy of existing berm reflection formulae was evaluated by use of statistical measures and M۵′ model tree algorithm is employed to predict the wave reflection with high precision. M۵′ model trees are trained and tested with the available experimental data. Both hydrodynamic and structural factors including wave steepness, berm permeability and structure slope have been considered in developing the prediction models. Performance of developed models is tested against the experimental data by using statistical error parameters. The results show that the proposed formulae by M۵′ model tree algorithm yields in more accurate prediction of wave reflection from berm breakwaters than existing formulae.

کلیدواژه ها:

Average structure slope ، Berm breakwater ، Berm permeability ، M۵’ model tree ، Wave reflection