Neural Network calibration method for VARANS models to simulate wave-coastal structures interaction: Fid-Bau Portal

Neural Network calibration method for VARANS models to simulate wave-coastal structures interaction

Díaz-Carrasco, Pilar / Molines, Jorge / Gómez-Martín, M. Esther / Medina, Josep R.

Abstract This study develops a calibration method for the porous media to properly model the interaction between waves and coastal structures using VARANS models. The proposed method estimates the porosity, n p, and the optimum values of the Forchheimer coefficients, $α$ and $β$ , that best represent the wave-structure interaction for a complete set of laboratory tests. Physical tests were conducted in a 2D wave flume for a homogeneous mound breakwater under regular wave conditions. Numerical tests were carried out using the IH-2VOF model to simulate the corresponding physical tests and incident wave conditions (H I, T). The numerical tests covered a wide range of Forchheimer coefficients found in the literature, $α$ and $β$ , and the porosity, n p, with a total of 555 numerical tests. The results of 375 numerical tests using IH-2VOF were used to train a Neural Network (NN) model with five input variables (H I, T, n p , $α$ and $β$ ) and one output variable $(K_{R}^{2})$ . The NN model explained more than 90% (R² > 0.90 and RMSE <5%) of the variance of the squared coefficient of reflection, $K_{R}^{2}$ . This NN model was used to estimate the $K_{R}^{2}$ in a wide range of n p , $α$ and $β$ , and the error ( $ε_{a}$ ) between the physical measurements with regular waves and the NN estimations of $K_{R}^{2}$ was calculated. The results of $ε_{a}$ as function of n p , $α$ and $β$ showed that for a given porosity, n p, it was difficult to obtain a pair of $α$ and $β$ values that gave a common low error if few physical tests are used for calibration. Then to calibrate properly a VARANS model it seems necessary to check the results obtained for each combination of $α$ and $β$ with many laboratory {H I, T} tests. The minimum root-mean-square error of $K_{R}^{2}$ ( $ε_{r m s})$ was calculated to find the optimum values of porosity and Forchheimer coefficients: n p = 0.44, $α$ = 200 and $β$ = 2.825 for the tested structure. Blind tests were conducted with the remaining 180 numerical tests using IH-2VOF to validate the proposed method for VARANS models. In this study, eight or more physical tests were required to find adequate values of n p , $α$ and $β$ for VARANS models related to the best performance of wave-porous structure interaction.

Highlights • A Neural Network model to predict $K_{R}^{2}$ on mound breakwaters much faster than any numerical VARANS model. • The physical measurement of the porosity, n p, is not reliable and it should be calibrated. • The selection of few physical tests (H I , T) to calibrate Forchheimer coefficients, $α$ and $β$ , is not sufficient to obtain the best performance of wave-porous structure interaction. • The proposed method based on a Neural Network model is a robust, accurate and computational efficient tool to calibrate {np, $α$ and $β}$ in VARANS models. • The proposed calibration method obtains the optimum combination of {np, $α$ and $β}$ related to the best performance of wave-porous structure interaction.