Wave attenuation prediction of artificial coral reef using machine-learning integrated with hydraulic experiment

Various prevention methods have been proposed to resolve such coastal erosion problems, known as an important social issue in many countries. However, most of the methods designed and constructed involve gravity-type structures, causing various problems such as the obstruction of seawater circulation and changes in water quality and marine ecosystem. In this study, we propose an eco-friendly artificial coral reef structure that can control coastal erosion and help preserve the coastal environment. We investigated the wave height distribution and energy reduction characteristics through a two-dimensional hydraulic model experiment under various conditions, including different external force conditions (wave height, water depth, and wave period) and design conditions (dimensionless crown width and submergence ratio). Furthermore, we investigated the drift sand control and morphological characteristics of artificial coral reefs through a movable bed experiment. The results of the hydraulic experiment revealed that the dimensionless crown width and submergence ratio factors had a dominant effect on the wave height attenuation. In addition, the artificial coral reef structures outperformed broadly applied tetrapod structures in terms of coastline advancement and seabed protection under storm wave conditions. Moreover, to estimate the wave attenuation characteristics of artificial coral reefs, we introduced a prediction study of wave attenuation characteristics of artificial coral reefs to propose an explicit method composed of weights and biases using modified artificial neural network. We found that the method of calculating the wave dissipation coefficient calculated with the presented weight and bias matrix shows more accurate reliability than the calculation method based on the existing empirical formula. •Development of artificial coral reef method to cope with the coastal erosion.•Hydraulic characteristics and erosion reduction performance of artificial coral reef method.•Prediction of wave attenuation characteristics using artificial neural network.•Easy method for calculating wave attenuation coefficients using weights and biases.