On the development of an onboard sea state estimator based on numerical vessel motion data

Publikationstyp
Conference Paper
Publikationsdatum
2022-06
Sprache
English
Author
Reiner, Carl 
Detlefsen, Ole 
Abdel-Maksoud, Moustafa orcid-logo
Institut
Fluiddynamik und Schiffstheorie M-8 
TORE-URI
http://hdl.handle.net/11420/13019
Citation
32nd International Ocean and Polar Engineering Conference (ISOPE 2022)
Contribution to Conference
32nd International Ocean and Polar Engineering Conference, ISOPE 2022 
Scopus ID
2-s2.0-85142158044
Publisher
International Society of Offshore and Polar Engineers
Peer Reviewed
true
Within the Joint Research Project RetroLadung, weight-optimized cell guides with integrated motion sensors and data-driven decision support systems (DSS) are developed. A basic functionality of the DSS is the estimation of sea states based on measurement data from acceleration sensors. A convolutional neural network (CNN) is created for this task and trained with simulation data. This paper presents the early development stage of the network during which unimodal, short-crested seas described with JONSWAP spectra are considered. Thereby, the forward speed is excluded. Model tank tests are conducted by the Institute for Fluid Dynamics and Ship Theory (FDS) and results are used to validate the predictions of the sea state estimation (SSE).
Schlagworte
Sea state estimation
DDC Class
600: Technik
620: Ingenieurwissenschaften