Prediction of Short-term Wind and Wave Conditions Using Adaptive Network-based Fuzzy Inference System (ANFIS) for Marine Operations

Abstract

The paper focuses on prediction of short-term environmental conditions by using an improved ANFIS method for marine operations. The hindcast data used here consists of ten-year long one-hourly time series of mean wind speed Uw, significant wave height Hs and peak spectral wave period Tp at the North Sea Center. Before applying ANFIS models, a non-stationary decomposition technique is applied to the initial time series in order to extract the non-stationary character of the series and obtain the corresponding stationary part. Then both time series (initial and stationary one) are employed to establish ANFIS models for the prediction of future values of wind and wave parameters, respectively. The performance of forecasting models is assessed by means of error measures and uncertainty quantification. Results indicate that the short-term predictions based on the stationary time series produces better forecasts in both wave and wind characteristics and has a great application potential in marine operations.