| dc.description.abstract | The safety and efficiency of marine operations at sea rely on accurate information about the sea state, which includes the dominant wave height, wave direction, and wave period. Unfortunately, many areas at sea lack this crucial data due to a scarcity of measuring instruments or inadequate measurement resolution. However, ships have the potential to address this issue since they are omnipresent at sea and situated near the waves, making them optimal platforms for both measuring and reporting wave conditions.
Shipboard sea state estimation uses sensor measurements of the sea surface from a vessel to determine important wave characteristics through model-based or signalbased approaches. Signal-based approaches have several advantages over model-based methods as they estimate waves directly from sensor measurements without the need for any complex ship model. However, these approaches often rely on expensive instruments and expert assistance for installation and maintenance.
This doctoral thesis investigates a relatively new and unexplored signal-based approach for shipboard wave estimation that is cost-effective and easy to implement. The approach uses the phase-time-path-differences (PTPDs) between an array of inertial measurement units (IMUs) to infer the directionality and frequency characteristics of waves. Only a few works have considered using a PTPD approach for wave estimation based on shipboard IMUs. However, these studies are restricted to model-ship wave tank testing in regular waves, and its application appears to overlook the differences between sensor delays on a rigid body and those directly obtained from sensors situated on the ocean. Moreover, it is presently unclear how many IMUs are needed, how far they should be separated, and how they should be geometrically arranged to determine the prevailing sea state.
The present study proves that the main wave direction and wave number can be uniquely determined from a minimum of two independent phase differences. Although measurements of the latter can be obtained from a minimum of three noncollinear IMUs, this work demonstrates that a single IMU is sufficient by utilizing a rigid-body measurement transformation to generate the other measurements needed. Moreover, a comprehensive theoretical assessment of the validity of the PTPD approach is conducted, determining the conditions under which it may be safely applied to model rigid body sensor delays. These conditions are validated experimentally through extensive testing with a model ship in a wave tank.
As a ship moves forward in following seas, it is a well-known problem that each encountered wave frequency can correspond to three distinct absolute wave frequencies, making it challenging to accurately determine the correct wave frequency during movement. However, through the observability results presented in this thesis, we prove that the absolute wave frequency can be uniquely determined while the vessel is moving using the PTPD approach. This interesting result is validated experimentally in a wave tank with a model ship exposed to various regular and irregular waves.
An inherent drawback of using measured ship motions to determine wave characteristics is that they are susceptible to distortions caused by the effect of vessel low-pass filtering when the waves are sufficiently short. To address this challenge, a novel analytical expression of the frequency bandwidth of undistorted waves is derived based on the main vessel dimensions. This frequency bandwidth aids in identifying the wave components that are safe to consider and those to avoid. This frequency bandwidth is incorporated into our proposed methodology for implementing the PTPD approach, which comprises a fast Fourier transform and an unscented Kalman filter. Moreover, with our proposed methodology, we are able to yield estimates of the wave direction and wave number/period close to real-time, with updates given every three minutes after an initial six-minute startup period.
The validation of the proposed approach is carried out through model-scale and fullscale field experiments. The latter involves a research vessel with a commercial wave radar operating alongside various wave buoys under diverse sea state conditions. The results of these experiments show strong agreement with wave reference systems, confirming the competitiveness of our theory and method against existing wave measurement technology. Notably, our proposed method offers advantages in costeffectiveness, simplicity, and environmental resilience, thereby establishing it as a promising alternative or complementary aid within the field. | en_US |
