Untersuchung zum Potenzial seismischer Großausführungsmessungen im Schiffsbetrieb
Iven Topp, M. Sc., Universität Rostock
Ensuring the safe and efficient operation of seagoing vessels requires continuous monitoring of critical technical subsystems. Established approaches include on-board sensors, periodic dry-dock inspections, or optical inspections conducted in port by trained divers. Methods such as hydroacoustic measurements offer the advantage of non-intrusive monitoring without equipment on board. Shipping channels are particularly suited to such measurements, as isolated vessels can be recorded individually. However, at the low transit speeds typical for restricted waterways, the propeller-induced components of underwater radiated noise (URN), especially cavitation, can fall below detectable thresholds. That creates a monitoring gap that hydroacoustic methods cannot reliably close.
This paper presents an approach to ship operation monitoring that addresses this gap by analysing seismic signals recorded by shore-mounted geophones. These sensors are established in the context of environmental seismology and are here applied to ship operation monitoring. To explore the applicability of this approach, a measurement campaign covering two months of continuous recordings was conducted at the Kiel Canal. The dataset comprises more than 2100 individual ship passages from over 1000 distinct vessels, time-stamped and clustered into ship types via their AIS data. Recordings were made simultaneously on two independent geophone sensors at different shore positions and on all three spatial components, providing a six-channel feature basis per event.
After instrument response correction, data post-processing extracts a six-second waveform window from the continuous measurements for every vessel. Spectra are filtered to suppress background noise. Feature extraction is performed with the open-source R package eseis, especially using its statistical descriptors per event. Pairwise analysis of the 19 spectral descriptors reveals a pronounced dependence on vessel size: ships with larger deadweight tonnage and overall length occupy systematically distinct regions across the majority of the 171 pairwise combinations. This is reproduced independently on both sensors, suggesting a physical relationship rather than a site-specific or sensor-specific artefact. Among the 99 vessels with five or more recorded passages, spectral descriptors show noticeably lower variance within individual vessels than across the full fleet. This indicates that a vessel-specific baseline could in principle be established, which is a prerequisite for future condition monitoring applications.
These results demonstrate, as a proof of concept, that shore-based seismic recordings carry vessel-specific signatures scaling with physical vessel properties and remaining broadly consistent across repeated passages of the same vessel. Remaining limitations include a potential site-specific influence on the observed signatures. The reliance on AIS-based vessel identification does not capture the effect of small ships without AIS, contributing seismic energy to nominally single-vessel recordings. The absence of concurrent weather information does not allow a quantification of weather-induced noise contribution. Future work will extend the approach through campaigns at additional waterway sites and through recordings of individual vessel transits at multiple sensor positions along the channel.
Exploring the Applicability of Full Scale Seismic Measurements on Ship Operation Monitoring
