The Prediction of Irregular Waves on Northeastern Jeju Island using the Wave and Wind Data

Abstract

In order to analyze wind wave in the field of frequency domain and assess application of a JONSWAP spectrum in IEC 61400-3 for wind offshore structures, wave and wind data was observed along the northeastern shore of Jeju island from Sep. 5, 2012 to Oct. 5, 2012 and Jan. 11, 2013 to Feb. 11, 2013. The JONSWAP spectrum is one of the models of wave spectrum that shows corresponding energy density of wave for each frequency. The JONSWAP spectrum needs some parameters such as value, peak frequency, peak period, significant wave height. Generally, value is used 3.3 or 1. However, many scientist suggested value depends on characteristic of area because of a large error. In this study, Available in study area are suggested using the observed wave data. In order to suggest calculating value for study area, the observed wave data was analysed using Fourier transform

As a result, the highest frequency is below 1m in the case of significant wave height and during the first observation, the mean of height was predicted at 0.523m and during the scend observation, it was 0.423m. Furthermore, in peak frequency, the highest frequency was 0.12Hz0.15Hz(period is nearly 6.67s8.33s), the results of from using significant wave height and peak frequency is 2.72 And the significant wave height calculated by straight linear regression equation was 1.635.

The predicted irregular waves along the northeastern shore of Jeju island using 1st regression model, SMB method, and JONSWAP spectrum wave model, and the observed wave data are compared the predicted significant wave height using SMB method, that is often used to study of offshore structure design with significant wave height using 1st regression function. In this study, each calculated significant wave heights is applied to JONSWAP spectrum and compare spectral moments of order 0 values because the spectral moment of order 0 the coefficient of determination by two kinds of transformed functionsvalues is the gross area of JONSWAP spectrum and importance parameter for calculation significant wave height.

As a result, Each coefficient of determination by two kinds of transformed functions as shown 0.58 to 0.61, the. This means that the SMB method was able to get better results than the regression function. However, in the case of RMSE (Root Mean Square Error), regression function has a lower value than SMB method (RMSE values are 0.45 and 28.47). By transformed functions, spectral moments of order 0 are 0.0473 and 0.4194, respectively and by observed significant wave height, spectral moments of order 0 are 0.0502 and 1.0054. According to the comparison of result of these values, results of the regression function are more accurate than SMB method. Finally, The significant wave height by the regression function model is transformed using the observed wind data from Jan. 11, 2013 to Feb. 11, 2013. The result of the comparison between calculated significant wave height and observed significant wave height is that the regression function shows more accurate than SMB method.