On-Board Trend Analysis for Cargo Vessel Hull Monitoring Systems
Gaidai, Oleg; Storhaug, Gaute; Wang, Fang; Yan, Ping; Naess, Arvid; Wu, Yu; Xing, Yihan; Sun, Jiayao
Peer reviewed, Journal article
Published version
Permanent lenke
https://hdl.handle.net/11250/3039961Utgivelsesdato
2022Metadata
Vis full innførselSamlinger
- Institutt for matematiske fag [2357]
- Publikasjoner fra CRIStin - NTNU [37292]
Originalversjon
ISOPE - International Offshore and Polar Engineering Conference. Proceedings. 2022, VOLUMES 1~4, 3628-3632.Sammendrag
With the increased focus on sustainability, de-carbonization and digitalization also the shipping industry is scrutinized. High steel weight implies high initial CO2 footprint but also higher operational CO2 footprint. With requirements to decreased emissions through IMO indices EEDI and CII, it is beneficial to lower the steel weight but without compromising safety. Utilizing sensor technology may enhance the safety and potentially allow for a more optimized structural design. For hull girder loading this is related to hull stress monitoring systems. Container ships are becoming increasingly important in the shipping industry. As manufactured goods are increasingly containerized, the container ship fleet has expanded.
According to DNV's hull monitoring rules, it is required to provide a forecast prediction based on recent measurement data to alarm the captain of potential extreme hull girder loading. If this forecast prediction is too low then a false impression of safety is provided, and if the prediction is frequently too high the captain may lose confidence in the system. These results are not acceptable, and it is therefore necessary that this prediction is reliable. This paper highlights modern challenges and ideas, incorporating novel statistical method to serve safe navigation. Trans-Atlantic voyages along with monitored onboard hull girder response data are discussed.
Sequence of the latest data from each individual sensor shall be displayed as a trend. Current rules states that 4-hour data sequence from each individual sensor shall form the basis for a forecast trend prediction of the expected response for at least the next hour. Encountered maximum stress may overshoot predictions by 50%, which is regarded unreliable. If too high values are predicted all the time, captain lose confidence and trust in the system trend analysis. If too low values are predicted the system provides false impression of safety. Both outcomes are unacceptable.
Special focus is paid to whipping, which is defined as a transient hull girder vibration phenomenon caused by wave impacts. The vibration decays slowly because of low damping, which results in whipping being superimposed on both the wave sagging and wave hogging cycle. The challenge is that the whipping may be of more freak nature than the more conventional wave bending considered in design.
Predicted stress level obtained by extrapolation is focused on extreme value for the sake of vessel hull safe navigation. Extrapolation by ACER (averaged conditional exceedance rate) method was done, including uncertainty bands.
Main motivation for this paper is to contribute to the development of reliability assessment methods for decision support on board ships which may facilitate an improved balance between safety and structural design. Seamanship is already a factor in ship design but through digitalization it may be enhanced allowing for some steel optimization without compromising safety.