Development of Hybrid Models for Short Term Ocean Wave Height Forecasting

Abstract

The significant wave height prediction plays a very crucial role in wave power generation. Apart from this, the hourly efficient prediction of SWH can significantly help to improve the decisions in maritime and offshore activities. But, the highly random and chaotic nature of ocean waves makes the significant wave height prediction task difficult and challenging. In this thesis, we have developed a series of wave hybrid models for hourly significant wave height prediction. Our developed hybrid models use a signal decomposition method along with a regression model. We have used the ?-Support Vector Regression (?-SVR), Least Squares Support Vector Regression (LS-SVR), Long ShortTerm Memory (LSTM) and Largemargin Distribution Machine based Regression (LDMR) model for the regression task. For signal decomposition methods, we have considered the Wavelet Decomposition (WD), Empirical Mode Decomposition (EMD) and Variational Mode Decomposition (VMD) method. Apart from this, we have also used the Particle Swarm Optimization (PSO) method to tune the parameters of the used regression model in our wave hybrid models. Till now, the VMD method and LDMR model have not been used in any wave hybrid model. We have evaluated the performance of our developed wave hybrid models on timeseries significant wave heights, collected from four different buoys, and located at different geographical locations using the different evaluation criteria. After the brief analysis of the obtained numerical results, we conclude that the LDMR based wave hybrid model outperforms the other regression model based hybrid model. Also, the VMD based wave hybrid models can obtain better performance than other decomposition based hybrid models. Further, we perform two way ANOVA analysis on obtained numerical results which statistically infer that the use of a particular decomposition method and a particular regression model affect the prediction accuracy significantly but, their effects are independent