Tracking Vessel Performance Patterns
Nautilus Labs makes use of high-frequency data (HFD) to assist shoreside teams to make optimal decisions for vessel marketing and maintenance actions by offering more responsibility for the upkeep of ship operating conditions.
In order to do this, Nautilus Platform analyses the efficiency of each aspect of a vessel’s systems with our proprietary performance models that surface actionable insights to track performance degradation across a fleet in real-time.
The key performance indicator for each ship is ultimately its speed and consumption profile. At Nautilus Labs, we use sensor data to model the relationship between speed and consumption of a vessel, stripping out the effects of weather and sea state, in order to confidently predict how much fuel the vessel will consume at a given speed under any environmental condition.
This relationship is grounded on thousands and perhaps millions of data points, providing powerful insight that helps our clients minimise fuel cost on upcoming voyages and more precisely price their vessels in the market. However, while speed and consumption is critically important as a lagging indicator of vessel health and a leading indicator of future market performance, there is a large benefit to being able to understand the health of other components of the ship’s system such as the propeller, hull, and main engine, for targeted monitoring and timely maintenance.
Today, most shoreside teams leverage time-based planned maintenance to conduct certain activities at standard intervals and react to unplanned vessel degradation after it has already had an impact on voyage economics.
Nautilus Platform provides a real-time dashboard of vessel performance across all core systems, isolating the cause of degradation and providing the necessary insights to drive maintenance decision-making. Nautilus surfaces these insights through the continuous monitoring of the relationships between main engine fuel consumption, shaft power, RPM, and speed through water, ensuring maintenance actions are properly prioritised in real-time.
To do this, we must understand the relationship between these key vessel performance metrics. Given the established relationships depicted in the image above, we have built machine learning models for understanding the effects of draft, trim, wind, and sea state on a ship’s performance. With a model for how the vessel responds to these various conditions, we can also understand how the vessel might have behaved without the effects of those same conditions. This process yields us a high-signal, normalised view of performance over time, and it’s a great way for us to identify problems in real-time.
Let’s dive into a real-world example, in which we witnessed a propulsion curve for a vessel undergo significant performance degradation immediately after an idle period in March-April 2018. This information, surfaced to the user, demonstrates the vessel’s added resistance for propulsion. It displays the change over time in the relationship between shaft power and speed through water.
In order to answer the question of what element of the system is experiencing degraded performance, we can look at the two underlying relationships that compose propulsion: hull and propeller. After similarly displaying the changes over time for these two relationships, as seen in the image below, we can conclude that the hull is responsible for this increase in resistance, as shown by the similar spike in resistance during the same time period.
Due to Nautilus’ Historical Analysis tool and our contextual map, we were able to observe that prior to the experienced degradation, the vessel was sitting idle in warm waters for a number of weeks, raising the possibility of hull fouling. Armed with this information, a shoreside team can schedule an inspection or cleaning in order to resolve the issue, and that’s exactly what happened in this example.
Furthermore, beyond just the identification and rapid response to a potentially impactful hull fouling event, the user can then track the efficacy of a hull cleaning by measuring how much hull efficiency was improved after it was completed. In this case, we can also see how well a hull cleaning was performed in September, which resulted in improved efficiency, but interestingly did not return the vessel to its pre-fouling levels. This added resistance presents itself in the overall speed and consumption curve and has implications for the environmental and economic efficiency of the vessel.
In conclusion, we have seen the incredible impact of high-frequency data matched with machine learning and a comprehensive vessel performance model through a real-world example. These are just some of the powerful workflows and opportunities that are available through Nautilus Platform and we are excited about what the future holds for the industry and our partners.