Author(s): Seok-Bong Lee, Kyung-Duck Suh

Linked Author(s): Seok Bong Lee

Keywords: Coastal structures, artificial neural network (ANN), wave run-up, wave forecasting, wave overtopping

Abstract: Artificial neural network (ANN) is a statistical learning algorithm which is inspired by biological neural network. ANN has been widely used in various forecasting areas. Therefore, it can be applied to forecast wave run-up scale. The wave run-up scale which was introduced by Na et al. (2011), describes the degree of wave run-up on a breakwater. In a previous research, Na et al. (2011) developed a wave run-up forecasting model by performing multiple linear regressions between several offshore wave data and photographed wave run-up scale data. In this research, we use an artificial neural network model instead of multiple linear regression models. The target is tha the breakwaters of this research are the same as those of the previous research, i. e. four breakwaters (Aninjin, Ayajin, Jungja and Kyungjung breakwater) located on the east coast of Korea. The ANN model is trained by using almost 70% of total data, and the remaining 30% was used to test the model. Good results are observed except at the Kyungjung breakwater, where Na et al. 's (2011) regression model also gave a bad result, probably because the wave run-up scale is smaller than at other breakwaters

DOI:

Year: 2017