A Comparison of Soft Computing Methods for the Prediction of Wave Height Parameters

Abstract

In the previous studies on the prediction of wave height parameters, only the significant wave height has been considered as the unknown parameter to be predicted. However, the other wave height parameters, which may be required for the design of coastal structures depending on their importance level, have been neglected. Therefore, in this study, novel soft computing methods were used to predict all wave height parameters required for the design of coastal structures. To this end, wave data were derived from a buoy located in Southwest Black Sea Coast. Then, Multi-layer Perceptron Neural Network (MLPNN) and Adaptive-Neuro Fuzzy Inference System (ANFIS) models were developed to predict wave height parameters. Various input combinations were selected to create seven different sub-models. These sub-models were applied using developed MLPNN and ANFIS models. Accuracy of sub-models were evaluated for each wave height parameters in terms of performance evaluation criteria. The results showed that the wave height parameters predicted by the MLPNN and ANFIS methods are similar and both methods yield results acceptable for design purposes. However, for maximum wave height, H_max, ANFIS sub-model yields slightly better results.