Generative AI (GenAI) uses large language models which learn from the patterns and structure of available data. Inspired by its ‘training data’ it can create fresh text, images or tunes based on new data with similar characteristics.
To the general public, it may seem like AI is something new, but this is far from the truth. Like many IT developments, AI's roots extend back half a century or more. What is new from the application perspective are the language models of Generative AI. It is also a tale of trying to find the most useful applications for this extraordinary tool.
Industry stakeholders will know that this longer story includes plenty of content from the maritime industries, but may be less aware that it is ‘analytical’ AI which marine systems have used to improve efficiency, optimize ship performance and advance efforts towards decarbonization.
This type of AI focuses on analyzing structured data and uses statistical methods to mine data for patterns, trends and correlations on which to base the machine learning algorithms that predict output values from a given set of input variables.
Kalevi Tervo, Corporate Executive Engineer and Global Program Manager at ABB Marine & Ports says that ABB has been using its deep sector experience and domain expertise to discover the opportunities with established analytical AI for many years. According to Tervo, Analytical AI has become “a natural part of our automation, service and routing solutions”.
Making sense of AI
“Combining the new possibilities brought by AI and ABB’s domain-specific marine expertise and product knowledge has been transformative,” says Tervo. “We work with and look for opportunities to use AI when it improves safety and efficiency for crew, passengers and cargo, where it can bring unique added value to our customers. We bring to the table ABB systems performance insights, product development knowledge, and service data that nobody else has access to.”
Enterprises focusing on AI alone can respond to marine market opportunities quickly, Tervo adds, but their solutions may not be fully sensitive to the subtleties of the way a ship works as a system of systems, how those systems interact with crew, and what the actual needs of the crew are.
“We believe in using AI when it makes sense,” says Osku Kälkäjä, Head of Digital Business, ABB Marine & Ports. “We have pushed at the boundaries of analytical AI, drawing on our role as technology leader in electrical power, control and automation systems, and using our experience in integrated ship systems to capture the complex relationships that can be optimized to improve operational efficiency.”
The combination of analytic AI and ABB’s domain-specific expertise has been key in two ‘off-the-shelf’ ABB solutions which enhance specific aspects of performance of vessels in services. These two example applications in the vast ABB portfolio utilize the AI approach in very different ways.
“ABB Ability™ OptimE learns from the relationship between a vessel’s speed and its toe angle by leveraging accumulated data to optimize propulsion efficiency, reduce power and minimize emissions. In terms of AI, this is quite a straightforward use case, but we are also introducing variables to the mix, such as weather or currents.”
In another case, AI’s probabilistic inferences, ie educated guesses based on probability, have been deployed to develop a model to automate aspects of vessel control – again with the goal of boosting efficiency.
“ABB Ability™ Marine Pilot Control consolidates mission critical vessel controls at one operator position,” explains Tervo. “AI has been used to develop models for predicting future vessel behaviours based on accumulated data. Marine Pilot Control itself is not AI, but the identification of the vessel model is based on measured data and algorithms which are part of the AI toolbox. The resulting state-of-the-art control technology helps to increase safety by automating key operations, such as position-keeping, speed control, braking assistance and docking.”
Bringing products to market which exploit AI in ways that customers find directly useful has demanded what Kälkäjä describes as “a laser focus on where our core strengths lie, in power plant optimization, control systems, energy storage and management, vessel stability, weather routing and remote maintenance”. However, he adds: “We have also recognized that others bring unique capabilities. Where AI is concerned, there is no one-size-fits-all for the maritime industry.”
Community is key
A shift in industry thinking is evident in the ABB Marine & Ports’ Routing Services. Complementing ABB’s existing digital offering, the recently acquired DTN Shipping business brought a market-leading Weather Routing optimization system to ABB. This expands the number of vessels connected to ABB networks to over 5,000, while the total number of vessels served by just ABB Routing Services is approximately 13,000 annually, enabled by the application programming interfaces (APIs) offered to third parties and integrators.
At the same time, Kälkäjä acknowledges that 50 percent of routing business works “through partners” which requires an agnostic approach with respect to the user interface on route selection. “Our routing services collect the data so that the customer can optimize the route. However, there are multiple interfaces available on the market that allow customers to access ABB Routing Services.”
Already, in its broader exploration of analytic AI, ABB has identified how collaborating with customers, class societies, third parties and even competitors helps to maximize vessel efficiency, fuel economy and decarbonization.
“We have also used AI to develop the performance condition-based monitoring and maintenance models we use to work with the OVERSEA ‘fleet support center as a service’ we offer with Wallenius Marine, for example,” says Kälkäjä. “Clearly, in these scenarios, data relating to ship engine performance is key.”
As the development of data-enabled solutions and AI accelerates, collaboration will become even more essential, if full value is to be realized from ABB’s sector expertise. Kälkäjä describes this as “a different business environment than we are used to, where the new added value comes from the community”.
Analytic, generative or both?
As part of a broader group which works in open collaboration with the world’s leading software developers to maximize the power of data and algorithms, ABB Marine & Ports will be at the forefront of the maritime industry when it comes to putting GenAI to good use.
“Like many of our industry peers, solutions such as Copilot are having a huge impact on our internal processes, for example by speeding up the work we do in documentation or even software development,” says Tervo. “This is increasing productivity, for sure, but the human in the loop still plays the critical role of evaluating how data is interpreted and verifying the results. These are benefits that, in the first instance, will be felt by the customer in accelerated product development, rather than in the product themselves.”
ABB Ability™ OptimE – Toe angle optimization for propulsion is an example of a product which uses AI and is fully ready for market. The subsequent development of the technology will be carried out along traditional lines towards ‘next generation’ software with added functionality in the years ahead. Based on the step change being brought by GenAI, future developments may need to include a more outward looking and fluid collaborative approach, where ‘service-based’ solutions emerge from pooling data, expertise and effort and a shared vision of user benefits.
No company is an island
“The pace of AI development means that maritime stakeholders are beginning to appreciate the benefits of looking at ship optimization as a service, where data models are continuously updated to bring real value to customers over time. Building on this, ABB’s role as a technology leader in developing, supplying, installing and maintaining ship systems also becomes a virtuous circle to collect and interpret accurate, reliable and invaluable data.”
“Our product knowledge, integrated system expertise and accumulated performance data have allowed us to lead on maritime AI. However, part of that leadership has been to understand how collaborating with partners maximizes the benefits of AI for the customer,” Kälkäjä confirms. “No company can master everything alone and get the most out of the opportunities.”
