experimental study on assessing interaction of quay walls and random waves using artificial neural network

Quay walls are sheltering structures used for protecting coastal regions against wave-induced forces. because of the random nature of the wave behavior, the application of physical models for the study of wave-structure interaction can be quite efficient. the aim of this study was to investigate the behavior of quay walls under random waves through experimental methods. the study used walls with vertical geometrical form, which were exposed to sea random waves under the jonswap spectrum. surface level and wall strain values were measured using built-in sensors. a neural network model was developed using the feed-forward method with the backpropagation algorithm to analyze the time series of water surface level and strain. high coefficients of determination during the training and verification phases were observed, indicating good network performance. self-correlation analysis of the time series showed that the data exhibited first-degree markov characteristics. this finding was taken into consideration and increased the coefficients of determination in the neural network model.