ABSTRACT

Statistical wave forecasting methods are sometimes applied because of their simplicity and convenience. Most of them, however, include some drawbacks from statistical and numerical viewpoints. In this paper, these drawbacks are discussed and a statistical forecasting method utilizing the Kalman filter technique combined with the Principal Component Analysis (Kalman-PCA model) is applied for the prediction of long period waves having the period of tens of seconds or longer, which was proposed to mitigate the drawbacks of the conventional statistical wave forecasting methods. Applicability and reliability of the method is examined for observed wave data at Shibushi port in Japan based on wave data and weather data for 5-years.

INTRODUCTION

Accidents such as cutting of moored ropes of cargo ships frequently occur due to their oscillations in harbors, which influence the efficiency of cargo loading. The cause is a long period harbor oscillation with the period of tens of seconds or longer. The countermeasures are being expected in many harbors. Several existing countermeasures are, however, not necessarily effective and do not meet the demand of the people who are in charge of cargo loading. On the contrary, in the present situation, it is considered to be rather difficult to establish an efficient countermeasure immediately though it may not be impossible. Under this situation, the forecasting technique for sea condition is alternatively required. In order to execute safe and economical control for cargo loading, it is indispensable to develop an accurate and reliable forecasting method for a long period wave conditions at least several days ahead. However, since the mechanism of the generation, development and attenuation of such long period waves has not been elucidated, we will introduce a statistical method for the proper judgment of the possibility of executing cargo loading on the bases of weather data.

This content is only available via PDF.
You can access this article if you purchase or spend a download.