Artificial neural network modeling of suspended load inside surf zone using wavelet transform

Anzy Lee, Hyun-Doug Yoon, Kyung-Duck Suh

Tuesday 30 june 2015

17:42 - 17:45h at Mississippi (level 1)

Themes: (T) Sediment management and morphodynamics, (ST) Sediment transport mechanisms and modelling, Poster pitches

Parallel session: Poster pitches: 7A. Sediment - Erosion

Sediment suspension inside the surf zone occurs aperiodically after wave breaking so that it is difficult to accurately predict its timing and magnitude. Moreover, it scarcely can be investigated by physical or analytical way because of its complexity of mechanism. For these reasons, data-driven models have received great attention as an alternative method to physics-based models. In this study, an artificial neural network (ANN) model is developed to predict the sediment suspension load at a specific time using the components of different scales that have been extracted from the wavelet analysis of wave surface elevation data. Considering that the period of each wave in a wave group can be detected by the wavelet transform of the surface elevation data, a time series of each component from the analysis with a scale is chosen to be an input data of the ANN model if it passes through the local maximum which can be noted by in a time- scale domain. The developed model is used to predict the time-dependent sediment concentration data collected during the CROSSTEX (CROss-Shore Sediment Transport EXperiment) at Oregon State University. The peak wave period was 4 s for the erosive case and 6.8 s for the accretive case.