Civil and Mechanical Engineering |
|
|
|
|
A hybrid AR-EMD-SVR model for the short-term prediction of nonlinear and non-stationary ship motion |
Wen-yang Duan, Li-min Huang, Yang Han, Ya-hui Zhang, Shuo Huang |
Department of Shipbuilding Engineering, Harbin Engineering University, Harbin 150001, China; Key Laboratory of Renewable Energy, Chinese Academy Science, Guangzhou 510640, China |
|
|
Abstract Accurate and reliable short-term prediction of ship motions offers improvements in both safety and control quality in ship motion sensitive maritime operations. Inspired by the satisfactory nonlinear learning capability of a support vector regression (SVR) model and the strong non-stationary processing ability of empirical mode decomposition (EMD), this paper develops a hybrid autoregressive (AR)-EMD-SVR model for the short-term forecast of nonlinear and non-stationary ship motion. The proposed hybrid model is designed by coupling the SVR model with an AR-EMD technique, which employs an AR model in ends extension. In addition to the AR-EMD-SVR model, the linear AR model, non-linear SVR model, and hybrid EMD-AR model are also studied for comparison by using ship motion time series obtained from model testing in a towing tank. Prediction results suggest that the non-stationary difficulty in the SVR model is overcome by using the AR-EMD technique, and better predictions are obtained by the proposed AR-EMD-SVR model than other models.
|
Received: 02 March 2015
Published: 03 July 2015
|
|
|
Viewed |
|
|
|
Full text
|
|
|
|
|
Abstract
|
|
|
|
|
Cited |
|
|
|
|
|
Shared |
|
|
|
|
|
Discussed |
|
|
|
|