Determination of Nearshore Sandbar Crest Depth Using Neural Network Approach

Mustafa Demirci, Fatih Unes, M.Sami Akoz

Nearshore Sandbar, Crest depth, Artificial neural network, Multi-nonlinear regression, Experimental study.

For the coastal structure designs, nearshore sandbars are crucial since they are affected highly from various parameters like beach slope, the height and period of the wave and the properties of the material forming the bed. In this study, it was investigated the sediment movements in nearshore by using various bar crest depths and a physical model. Erosion profile output is used for determination of the bar crest depths. Linear and non-linear regression methods are used for obtaining the non-dimensional equations with the experimental data. These equations are then compared with the ones found in the literature. Transportation of on-off shore sediments is affected by bar crest depth which has been examined with the materials forming the beach by using various diameter of the medium as d50=0.25, 0.32, 0.45, 0.62 and 0.80 mm. In order to estimate nearshore sandbar crest depth, we have developed an approach by using neural network (ANN). For proposing the efficiency of the study, ANN and multi-nonlinear regression models are compared with each other.

10.22161/ijaers/3.12.27