Transforming harbour operations with AI

Harbour operations are an essential part of day-to-day activities in port, ensuring ships make it safely to and from their destinations, but despite the importance of a system that touches every port call, co-ordination is often left to labour-intensive manual operations. ‘We go to ports where they do the planning of harbour operations with no digital systems in place; they’re either using Excel sheets, or they’re using a big whiteboard’. He recalled a recent trip to a large port with complex operations. When he stepped into the control room, he was met by a five-metre wall of whiteboards filled with pilot and vessel names co-ordinating the harbour operations. ‘Essentially, those systems are based on one person’s understanding of the vessel and port movements, not data. Basing all these decisions on manual planning creates very inefficient outcomes.’

Deeper Insights

By accurately estimating the time each operation takes, an AI platform can help turn multiple complex data streams into actionable suggestions for harbour operations planning, says Yeo.

The details considered can also include the experience of the pilot assigned to a particular operation. ‘All of these movements are controlled in terms of speed of movement in the harbour area. A senior pilot with extensive experience can make a huge difference in operational timings. I’ve seen experienced pilots who really know how to turn a vessel and get it in place. It’s impressive to watch, and it makes a big operational difference.’

‘Understanding the duration of these operations and the complexity of operations helps us optimise the whole planning process,’ said Yeo. The data from the AI model that predicts the time specific operations will take is then fed into an optimisation engine which suggests how to assign resources in the most efficient way to reduce fuel usage, cut costs, and save time. The outcome is a set of recommendations on the necessary assignments it sees for the harbour’s assets. The system can produce multiple approaches to the same load of tasks with cost breakdowns for each iteration for comparison.

‘The whole premise is that the operator will be aided by AI to make the right operational decision. At any point of time, the planners can override the decision by changing these parameters if they wish,’ said Yeo, adding that their approach as a company has always been to augment decision-making rather than make the final decision.

Cost savings

An external audit by one of the ports that adopted Innovez One’s system showed savings of around 20% in terms of tugboat travel times.

These small optimisations translate into fuel savings, time saved by pilots, and reduced maintenance costs for harbour craft. The cost savings and operational improvements quickly stack up, said Yeo, and the benefits of employing the AI system scale with port size as the number of daily movements increases. ‘You can imagine, as the number of times you send out pilots goes up, being able to effectively share pilot boats to transport a number of pilots at once makes a big difference compared to an unoptimised system.’

Fear of AI

Another component of the efficiencies AI systems present is in employee headcount. Yeo said he had seen a team of 10 people handling the pilot logistics at one port.

The staff were making phone calls and updating spreadsheets with information on which pilot was where, which pilot boat was assigned, and even car arrangements. The work of that team of 10 could be handled by a single person observing an AI system and checking the schedules it generates, according to Yeo. ‘I think the fear of AI, and of technologies replacing jobs is very real.

Oftentimes, when organisations need to implement AI, the biggest issue that they are faced with is with employees and their unions.’ Dealing with the challenge of reducing headcount can be difficult, but employees in at-risk positions can be upskilled to work with the AI system and find cost efficiencies for the benefit of the entire organisation, Yeo believes.

Learning on the job

Some point to the adaptability of human decision-makers as a strength AI would struggle to match, but Yeo says the models are constantly learning and getting smarter from the information they observe. This means estimates become sharper over time, but also that changes in operations due to procedure updates or policy changes can be picked up by AI models. ‘AI can be modelled and retrained so that the outcomes match the new policies that have been put in place. If you get a human planner to do this, they need to relearn this all manually without that live data supporting their plan.’

Casting a wider net

While the savings from optimised harbour operations are not to be sniffed at, even greater efficiencies are to be had once Just-In-Time (JIT) arrival and accurate Estimated Times of Arrival (ETAs) of ocean going vessels into port becomes are integrated into systems. Having accurate vessel ETAs and scheduling harbour operations based on those predictions will require further adoption of technology and data sharing both in port and on ship but will bring significant savings potential for many parties.

Ocean-going vessels can save time or fuel on a voyage through the exchange of information and forecasts, adjusting speed to account for the optimal arrival time. There is plenty of room for improvement in the current way of working, where ships often rush to their destination, burning extra fuel, only to wait at anchorage until a port is ready to accept them. Systems allow for communication between ship and shore, with ports sending out a Requested Time of Arrival (RTA) to a ship, which can then be confirmed by the master or adjusted depending on circumstances. Through the improved communication supported by AI predictions, ports can optimise their operations by using the same accurate forecasts to understand upcoming demand for services. ‘If a vessel says it’s going to arrive tomorrow at 0900hrs, but then the prediction says it’s going to be three or four hours late, that may impact logistics arrangements, the pilots, and the tugboat that need to be in place. The vessel may arrive at a time where it is not suitable in terms of tidal height.’

‘I foresee that in the next two to four years, ports that do not have JIT or vessel ETA to support their decision-making processes will lose out to those ports that do, which is why we are gearing up to offer our customers those functionalities.’ The adoption of AI tools holds the potential for extreme refinement of harbour operations and wider maritime operations, but a co-ordinated adoption of connected digital solutions is essential as a foundation for more advanced technologies.

See also our article on Digital Silo Busting for more of in-depth insights.

Find all our articles within the Global Ports Report 2024 and stay up-to-date with the latest news.