Abstract
The data in noon reports on a daily basis from vessel operations reflect the variation of operational
behaviors. In our previous research, a practical assessment method has been proposed to diagnose the quality of
individual sample through validation rules. The findings have been implemented and as a consequence, the
quality of data has been improved to some extent based on triggered warnings and alarms. However, the unstable
periods of data, defined as the spans with large variations that are different from expected values, and their root
causes are still hard to be identified by the method, which limits further improvement of vessel operations. In the
present research, the so-called SlideCOV method, is proposed to provide diagnosis on the unstable periods of
noon reports in vessel operations accounting for multiple samples. This method utilizes historical COV
(coefficient of variation) values of the performance indicators (e.g., Specific fuel oil consumption (SFOC)) and a
sliding window with a fixed width and step for the identification of current unstable periods. Sensitivity study
has been conducted for the selection of widths and steps of sliding windows. The unstable periods can be
identified when the COV values exceed a given threshold. The effects of short voyages and long voyages have
been taken into account in this method. Case studies have been done based on the datasets provided by shipping
companies. Typical root causes of the unstable periods of vessels are identified based on data analysis, which
will provide guidance on the improvement of vessel performance in practice.
behaviors. In our previous research, a practical assessment method has been proposed to diagnose the quality of
individual sample through validation rules. The findings have been implemented and as a consequence, the
quality of data has been improved to some extent based on triggered warnings and alarms. However, the unstable
periods of data, defined as the spans with large variations that are different from expected values, and their root
causes are still hard to be identified by the method, which limits further improvement of vessel operations. In the
present research, the so-called SlideCOV method, is proposed to provide diagnosis on the unstable periods of
noon reports in vessel operations accounting for multiple samples. This method utilizes historical COV
(coefficient of variation) values of the performance indicators (e.g., Specific fuel oil consumption (SFOC)) and a
sliding window with a fixed width and step for the identification of current unstable periods. Sensitivity study
has been conducted for the selection of widths and steps of sliding windows. The unstable periods can be
identified when the COV values exceed a given threshold. The effects of short voyages and long voyages have
been taken into account in this method. Case studies have been done based on the datasets provided by shipping
companies. Typical root causes of the unstable periods of vessels are identified based on data analysis, which
will provide guidance on the improvement of vessel performance in practice.
| Originalsprog | Engelsk |
|---|---|
| Titel | 2022 International Conference on Naval Architecture and Ocean & Marine Engineering (NAOME 2022) |
| Status | Accepteret/In press - 15. aug. 2022 |