Using AI and machine learning to optimise vessel health

Using AI and machine learning to optimise vessel health
MarineInsight is an AI-based platform that provides a complete picture of a vessel’s performance and the health of its systems

Seattle-based ioCurrents is pioneering a new artificially intelligent (AI) platform that combines digitalisation and machine learning in real-time to provide actionable insights on ship performance optimisation.  Digital Ship caught up with Cosmo King, co-founder and CIO to find out more.

Thanks to advancing technology, greater and more accurate data on vessel performance is available to shipping companies looking to better understand performance of their ships. However, translating raw data into useful, actionable insights is still a challenge. ioCurrents has developed an AI-based platform, MarineInsight, that uses data from a vessel’s sensors to provide a complete and accurate picture of a vessel’s performance and the overall health of its systems. The solution recommends when, where and why each asset onboard a vessel needs attention, providing key information needed to predict failures, optimise fuel consumption, and drive more efficient maintenance.

Making sense of data

Spending much of his childhood on an old steamboat-turned-restaurant in Chesapeake Bay, Virginia, Cosmo King grew up with a close connection to the maritime environment. His personal interest in computers and technology led him to eventually join tech company SpaceCurve where he met Bhaskar Bhattacharyya, who would become his co-founder of ioCurrents. Eight years ago, King and Bhattacharyya were building high-scale databases for processing geospatial IoT and sensor data and realised discussions on IoT were not transferring to the maritime industry. They saw that data was being collected but analysis was lacking and believed IoT could be better applied to ships to optimise fuel consumption and drive better maintenance strategies.

“No one wants to be barraged by millions of records of meaningless and contextless data and have to try to figure out how to make sense of it,” King told Digital Ship. “Analysis is needed, and artificial intelligence and machine learning can provide the ability to automatically analyse data in a way that gives you answers that are more meaningful,” he said. “This will ultimately help to solve problems in maritime that we thought were never possible.”

Bhattacharyya and King decided to use their experience in data science and machine learning to launch software company ioCurrents, which helps shipping companies to make better sense of their data through an intelligent and scalable platform called MarineInsight. MarineInsight takes data already collected by sensors onboard ships and turns this into a Digital Twin to characterise and understand a vessel’s operations. Data can be collated from sensors and the cloud, and analysed to deliver real-time information on the performance efficiency of any vessel and its assets.

MarineInsight analyses a variety of ship components including engines, generators, Z-drives, transmissions, fuel and water tanks, winches, PLCs, or custom sensors for parameters like temperature or vibration. The tool can be used to determine how efficiently an engine is running, how much fuel is being consumed and the emissions it is producing. In the event of a ship having difficulty running a certain type of fuel, a common scenario where fuel has recently been switched from high-sulphur to low-sulphur to meet regulation, MarineInsight generates a smart alert before any propulsion issues arise. Using AI and machine learning, MarineInsight predicts when equipment is likely to fail to enable condition-based maintenance, improving the lifespan of equipment and reducing vessel downtime, while minimising maintenance costs.

MarineInsight focuses on analysing data from the core systems of the vessel to determine its overall health. “We know that data from engines and from generators and from GPS navigation systems is very useful because if any of these fail it could be catastrophic,” King explained.  Once this data is analysed and given to the operators, anyone within that organisation as well as permitted stakeholders on ship or on shore can access the data in real-time. The overall goal is to bridge the gap between ship and shore by providing real-time vessel analytics to personnel wherever they are located. Within the MarineInsight portal is a key feature called the InsightHub, where users can access real-time automated reports, showing information that can be provided to the IMO or other regulatory bodies to prove compliance in various areas. The reports are automatically generated, eliminating the need for crew to manually complete them to reduce administrative burden and enhance report accuracy. The reports and charts can be customised to zoom in on any area of a vessel’s operation. For those just starting out on their digitalisation journey and striving to keep costs low, individual reports can be purchased as and when required for a few hundred dollars.

In addition to providing information on the health of a vessel, MarineInsight provides real-time insight on fuel performance and indicates how current fuel consumption compares to a vessel’s previous voyages. Using data collected from a vessel over a one-month period, MarineInsight establishes a baseline and uses self-learning algorithms to identify whether the current voyage is less or more fuel efficient than similar previous voyages.

Voyage optimisation

ioCurrents is planning to launch its voyage efficiency optimisation tool with AI during 2021. “The tool will enable you to get recommendations on the optimal speed for the vessel based on what AI has learned about the equipment’s most efficient mode of running,” King explained. “We’re not telling you where to go but we’re saying speed up or slow down throughout your voyage, at times that the captain might not have expected to make a change. Over the course of the voyage, this is going to give you the most efficient total run. We’ve seen substantial improvements in performance while still arriving on time using this method and I think this will help companies to save money as they take on more expensive types of fuel to meet environmental initiatives.”

Installation and integration  

The MarineInsight platform consists of two key components. The DataHub, a mini-computer that collects and analyses data locally and is stored onboard the vessel. The second component is the remote analytics cloud platform, which continually receives data that is uploaded via a cellular or satellite connection.

MarineInsight can be integrated with data from other solutions that are stored on the cloud. For more rich and real-time solutions, the system can also be integrated with other onboard systems. “We have APIs on our software that can communicate with other software that already exists onboard the vessel. Some of that data never needs to be uploaded to the cloud, it can be acted on immediately onboard the vessel, even if connectivity is poor,” King explained.

MarineInsight is quick to install and begins collecting and analysing data immediately. For the first couple of months, the platform uses this data to build a model to understand the vessel’s performance. After a few months of use, MarineInsight is able to apply higher levels of machine learning and AI to carry out predictive analysis and efficiency optimisation.

While the pandemic has affected physical movement and installation of solutions onboard vessels, ioCurrents saw this an opportunity to develop a self-installation kit. The small equipment package can be shipped anywhere in the world and installed by crew members without a high degree of technical knowledge.

Customers and partners

MarineInsight was first introduced to fishing vessel customers based in Washington State but operating out of one of the largest American fishing ports, Dutch harbour in Alaska. For ioCurrents this was the perfect opportunity to trial the technology onboard small vessels operating in treacherous waters. According to King, many fisherman he spoke with were keen to resolve safety issues they were facing as well as showing motivation to tackle environmental challenges. Following successful roll out to the fishing sector, ioCurrents moved on to the tug and barge industry. “We were keen to scale up the technology and we saw from these operators that there was a lot of motivation to improve navigational safety. If the main engines fail on a tugboat and it’s moving between here and Alaska, pulling a two thousand tonne barge behind them, it’s going to be really bad if they lose the ability to navigate in waters with swift flowing currents. Not only is this a risk to their life, but it could also be an environmental disaster,” King said.

Today MarineInsight is used onboard commercial and passenger vessels including tankers and bulk carriers.

ioCurrents is currently working with top maritime security technology groups to provide third party verification.

ioCurrents is also partnering with various companies to expand its solution. A partnership with KVH is enabling ioCurrents to offer KVH Watch connectivity as part of its maritime services. KVH Watch’s dedicated connectivity enables ioCurrents to use real-time data to provide customers with actionable insights and engage in real-time video troubleshooting sessions during a voyage.

In May this year, Danelec Marine announced it was connecting its shipboard data to the ioCurrents platform.  This partnership will enable owners and operators to access even greater data on vessel performance without the need to install any additional hardware from ioCurrents.

Most recently, Kongsberg Digital announced that it was adding software solutions from MarineInsight to its Kognifai Marketplace. This will enable its Vessel Insight users to access automated reports on vessel health and fuel consumption based on automated data analysis as well as ioCurrent’s more advanced AI driven machine learning models to predict failures and dynamically optimise voyages. This is all available without the installation of additional hardware.

This article was originally published in Digital Ship’s Jun/Jul magazine. Click here to download the issue and read the article on page 28.