vor 1 Jahr

HANSA 02-2021

  • Text
  • Hansaplus
  • Maritime
  • Hansa
  • Ships
Kooperation HANSA & MRP · DNV GL · Hafen Hamburg · Port Hub · MPP-Marktbericht · 3D-Druck in der Schifffahrt · Interview IMO-Chef Lim · Maritime Future Summit · SMM Digital 2021

Smart & Digital 1 2 Ship

Smart & Digital 1 2 Ship data analysis: Benefits and realities In order to identify operational anomalies through data analysis, shipping company Canada Steamship Lines and the National Research Council of Canada joined forces. The setting was a bulk carrier Ship owners and operators are becoming ever more interested in data collection. However, the next step of analyzing the data is still in the early stages of being realized. The National Research Council of Canada (NRC) has been conducting vessel analytics in collaboration with Canada Steamship Lines (CSL). The data described here is collected and transferred daily aboard a 225.5 m bulk carrier operating the Great Lakes and St. Lawrence Seaway in southern Canada. Operational data from between 2017 and 2020 was analyzed culminating in performance characteristics and operational metrics. The main goal of this work was to identify operational anomalies and report them in near real time to the vessel operator. To assess the operational performance of the vessel, the data was split into segments based on geofenced zones. The limits for each is defined on discrete operational areas such as lock transits, lake crossings, and constricted waterways. Figure 1 is an example of the vessel transiting through five zones within the northern section of the St. Lawrence River. Segmentation and statistics Statistics for each segment were calculated such as the distance, fuel consumed, fuel consumed over distance, and the total transit time. Multiple passes through the zone were then analyzed for consistency. To exemplify this, consider Figure 3 which summarizes passages through Lake Superior. The transit that occurred in 2017 has a higher fuel consumed over distance and lower mean speed than the other transits. Upon further investigation, the wind speed during this transit was higher than typical, as can be seen in Figure 4, highlighting the wind for that transit overlaid against the wind for 3 weeks around this period. Thus, the likely culprit of this anomaly was environmental. Performance curves A set of performance curves were generated to characterize the baseline performance of the vessel. Historical data was used to create the baseline performance curves and these curves were used as a means to identify anomalies in near real 4 4: Shaft Power vs Speed Through Water 5: Performance Curve with Percentiles 6: Curve 1 - Anomaly Detected 42 HANSA – International Maritime Journal 02 | 2021

Smart & Digital 1: Geofenced Zones 2: Lake Erie Transit Comparison 3: Wind Speed During Lake Superior Transit 3 time from the operational data. The Shaft Power vs Speed Through Water performance curve is provided in Figure 5. While the data does somewhat follow a third order curve, there is significant spread, which is typical for operational data. The irregularity of the data results from both operational and environmental factors. Consider the 5 - 5.7 MW cluster of points, where 60% of all the data actually resides. The operators of the vessel typically set a cruising »power« rather than a cruising »speed« when in unrestricted waters since machinery is optimized to run at a specified power output. The variance of weather, current, and vessel loading tend to spread the data between 10 - 14 knots. As such, a method beyond a simple polynomial fit was necessary to define the performance curves. Anomaly Detection Methodology A percentile bin method was used to determine outlying anomalies. First, the data in a performance curve was separated into bins. Each bin contained an equal number of data points and were separated based on vessel speed. The more time the ship operates in a given speed range, the higher resolution of that bin. The median, The St. Lawrence Seaway is a busy waterway 60% and 70% of each bin was calculated in terms of the shaft power range within that bin. These data limits were added to the baseline performance curve for each bin as in Figure 6. © U.S. Department of Transport The data from the previous day was added to the curves to compare the near real time data to the historical performance. An anomaly was flagged if 5 minutes of consecutive data lay outside the 70% tolerance lines. Consider Figure 7 which illustrates a single anomaly (red markers) that occurred. This event resulted from a high acceleration where the power was brought up quickly and the vessel took several minutes to reach speed. Complexity The analysis of operational vessel data is challenging and complex. There are many factors that influence the data including external, operational, and vessel conditions. Insight from this type of analysis can lead to increased operational efficiency and decreases in fuel consumption. This can result in savings for vessel operators and result in an overall greener fleet. Authors: Trevor Harris Allison Kennedy, National Research Council of Canada 5 6 © National Research Council of Canada HANSA – International Maritime Journal 02 | 2021 43

HANSA Magazine

HANSA Magazine

Hansa News Headlines