PO054

A study on the combined wave spectrum model of wind wave and swell

Jun Tanemoto 1, Takeshi Ishihara2, Atsushi Yamaguchi2
1Wind Energy Institute of Tokyo Inc., Tokyo, Japan, 2The University of Tokyo, Tokyo, Japan

Abstract

In IEC 61400-3, JONSWAP spectrum is recommended for the load analysis of wind turbine and its support structure. However, since the fatigue load analysis is carried out under various wind speeds, variety of spectrum shapes different under each wind speeds should be taken into account. At the site facing open ocean such as Pacific Ocean, a double peaked spectrum is observed when wind wave and swell is mixed.

In this study, a combined wave spectrum model for wind wave and swell is proposed and verified by using measurement data obtained at Fukushima offshore floating wind turbine demonstration site. The combined spectrum is modeled as linear combination of wind wave and swell component based on JONSWAP models.

Measured significant wave height and wave period are used for the input parameters of the model and predicted spectrum is compared with directly measured one.

According to the result of a case study, it is found that proposed spectrum can be predicted double peaked spectrum and show better agreement with measurement than JONSWAP model. Models are also validated for average spectrum in each wind speed bin. Proposed spectrum model shows better result than conventional JONSWAP spectrum, especially for the width and peak energy.

Method

In this study, combined spectrum is modeled as linear combination of wind wave and swell component, in both of which JONSWAP spectrum are used. Measured significant wave height and wave period are used as input parameter of a spectrum model and measured spectrum at the same site is used for the validation of the model. Note that significant wave height and wave period are needed to separate into these wind wave and swell component for the proposed model. For these separation, formulas shown mean relations of wind speeds and significant wave height/period are used, which were proposed by Tanemoto and Ishihara (2016) and composed by wind wave and swell components.

Results

Proposed model is validated through a case study which double peaked spectrum was measured at the site. Conventional JONSWAP model is also validated.

JONSWAP model cannot predict double peak spectrum because it has only one pairs of parameter for wave height and period. However, by using proposed model double peak spectrum can be predicted and shows good agreement with measurement.

Then, proposed and JONSWAP models are validated for average spectra, which are defined as average of spectra generated by using measurement in each 1m/s bin, in which one year's worth of measurement is used. By using proposed model, low frequency part of the spectrum can be predicted due to separations of wind wave and swell components, and it obtained better agreement with measurement than JONSWAP.

Conclusions

In this study, a combined wave spectrum model of wind wave and swell is proposed. By using proposed spectrum model, double peaked spectrum can be produced. Proposed spectra are validated through a case study and one year's worth of bin average spectrum obtained from measurements, and shows better agreement than those obtained by JONSWAP model.

Objectives

In the range of wind speed from 7 to 12m/s, which wind wave and swell is comparably mixed, proposed combined spectrum model shows good agreement. Proposed model is expressed as simple formula. Therefore it is easy to use and may contribute to improve the accuracy of external condition for the fatigue load analysis.