Seafarers are concerned that in the coming decade, they may lose their jobs in a wave of autonomy, as self-driving, autonomous surface vessels (ASVs) take to the seas, making them redundant. Investors are hoping that a new wave of such ships will make them rich, looking to become the "Tesla of the oceans" selling or deploying self-navigating, autonomous vessel technology and emissions-free power systems to transform the marine industry, and to make themselves squillions of dollars in the process, of course.
Governments have been handing out subsidies to assist in the development of autonomously navigated and zero-emission marine technologies, including US$15 million from the Norwegian government's green investment fund Enova to Cayman Islands-owned survey venture Ocean Infinity, and multimodal transport and logistics group Samskip to build two small container vessels for service between Oslo and Rotterdam.
If you only watch one news video this week, then I would recommend you watch the BBC's excellent coverage of the world's largest and first autonomous cargo vessel Yara Birkeland here. The 120TEU open-top container ship is now sailing on the fjords of Norway for its owner, chemical company Yara. The 80-metre-long vessel is used for transporting fertiliser from Yara's plant at Herøya and its main export port in Brevik, near Oslo.
There's a lot of hype about the autonomous vessels sector, which we have attempted to debunk (here and here). The cautious and, to my mind, eminently sensible, progress of Yara and Kongsberg shows us what we can expect from the technology over the coming decade.
Yara Birkeland exists to take trucks off the road, reduce carbon emissions through complete use of battery power, and showcase Norwegian maritime technology. Yara claims that the vessel will reduce NOx and CO2 emissions by reducing diesel-powered truck transport, saving around 40,000 lorry journeys per year. It is important to note that Yara is one-third owned by the Norwegian government, which is its largest shareholder. You may know of Yara under its former name, Norsk Hydro.
The key technology – including the sensors and integration required for remote and autonomous ship operations, the electric drive, and the battery and propulsion control systems – is provided by Kongsberg. Kongsberg is just over 50 per cent owned by the Norwegian government. When the vessel was christened in April, Crown Prince Haakon attended the ceremony.
Yara Birkeland was built at Vard Brevik in Norway, and cost NOK250 million (now US$25 million). That's US$200,000 per TEU – not at all cheap.
But this is part of a bigger national plan to decarbonise by Norway. Other companies and governments pursuing autonomous vessels and zero-carbon vessels lack the same national drive and integrated planning as Norway.
First some background: the vessel is owned by an industrial player, Yara, not a standard shipowner. Yara Birkeland will likely work for Yara for its entire operational life, as long as Yara is producing fertiliser at Herøya in Norway, as it has been doing for nearly one hundred years. Yara's fertiliser factory at Herøya is one of the country's largest sources of CO2 emissions outside the oil and gas industry, and is responsible for releasing 800,000 tonnes of CO2 annually.
A crucial element in fertiliser production is hydrogen, which is converted into ammonia. Today, the hydrogen is produced using liquefied natural gas (LNG). By producing hydrogen based on renewable energy, Yara will be able to create emission-free ammonia.
Yara identified that it, and by extension, Norway had a unique opportunity to take a leading position in the green transition, but that the window of opportunity is limited. "Emission-free ammonia is the key to reducing emissions from world food production and long-distance shipping," said Svein Tore Holsether, CEO of Yara.
Yara's corporate board has made the decision to invest in a 24MW demonstration plant where the technology will be demonstrated, backed by the Norwegian government. Yara has been granted NOK283 million (US$29 million) from Enova for this development to produce green ammonia for fertiliser and ships.
This plant will be one of the largest projects producing green ammonia in the world, Yara says. Renewable energy will replace fossil fuels in the production process and thereby reduce CO2 emissions by approximately 41,000 tonnes per year. The plant will produce enough hydrogen to produce 20,500 tonnes of ammonia per year, which converts into between 60,000 and 80,000 tonnes of green fossil-fuel-free mineral fertiliser. The project aims to supply the first green ammonia products as early as mid-2023, both as fossil-fuel-free fertiliser as well as fuel for ships.
"We move from good intentions to actions, the investment decision has been made and the project begins now", said Magnus Ankarstrand, director of Yara Clean Ammonia, in the press release here.
When the newbuilding contract was announced, Yara said that Yara Birkeland would be sailing without crew by 2022. In 2021 it commenced operations with a full crew on board. Today, as the BBC video shows, it is still operating with a human crew aboard, as Yara learns from the operation and gradually prepares for a staged move to control from the Remote Operations Centre (ROC).
Certain owners have talked grandiosely about revolutionising the entire marine industry and removing all crew from vessels in the coming years. Yara Birkeland is at the cutting edge of technology and suggests that slow, steady, and safe progress is the way forward. The Norwegian tortoise will likely beat the Silicon Valley or Silicon Wharf hare.
The ship will always operate within twelve nautical miles from the coast, the designers say, and usually less than two nautical miles in its standard operations. Yara Birkeland operates exclusively between Herøya and Brevik. It doesn't sail on other routes, it doesn't conduct dynamic positioning operations alongside oil and gas facilities offshore, it doesn't serve multiple ports in different countries. It is always within coastal waters within an hour's drive from the proposed ROC in Horten, which will be operated by a joint venture between Kongsberg and Wilhelmsen.
There are some challenges, yes, and along the route there are speed restrictions under the Brevik bridge, challenging currents, and some passages that are classified as congested waters. The entire route is monitored by the Vessel Traffic Service in Brevik. However, this is emphatically not like trying to manoeuvre a platform supply vessel (PSV) alongside in Aberdeen harbour on a busy day, or transiting the Straits of Singapore at night or sailing through a wind farm with tightly packed turbines all around. Yara fully controls the berth at Herøya.
Kongsberg has designed a nifty solution to automatically moor the vessel using cranes and specially designed wire hawsers and bollards, and to load and discharge it with automatic gantry cranes. Removing stevedores from repetitive rope tying duties at the wharf makes sense, since the vessel visits only the same two berths and the same mooring bitts over and over. The repetitive monotony of the operation makes it suitable for automation.
One key learning, and a key warning for others attempting to enter the autonomous navigation segment, is that the distance between Herøya and Brevik is approximately seven nautical miles.
Yes, you read that right. Seven nautical miles.
This is a very controlled environment, and Yara Birkeland is not operating in the open ocean, but in fjords only, and over very short distances. It can operate completely on battery power because it is never more than seven nautical miles from its charging station, and it uses its batteries as permanent ballast, turning the problem of their weight into an advantage.
Over the next two years, Kongsberg and Yara say that they will gradually implement and test the ship's autonomous functional capabilities. The captain will get ready to relocate to the ROC over this period. The partners will also begin the approval process for un-crewed vessel operations in conjunction with the Norwegian Maritime Authority and third-party assessment from DNV, as classification society.
Yes, you read that correctly too. Even though the vessel is in service today with crew, its operations without human crew are still waiting for approval from the regulatory authorities.
One of the reasons Yara and Kongsberg have been so patient is the risk that an accident will undo all the progress they have put into the technology.
Consider the recent coverage in electrek magazine of the test of Israeli company Orca AI's Artificial Intelligence software on the 749GT vessel Suzaku for Japanese shipowner NYK (here). The vessel navigated autonomously on a 40-hour voyage of just under 500 nautical miles in the congested waters of Tokyo Bay under the control of Orca's AI system to demonstrate its capabilities, but with a human crew onboard.
The ship "avoided hundreds of collisions" under the control of the software, and the report states that Suzaku travelled "without human intervention for 99 per cent of the trip."
I don't need to emphasise that a 99 per cent success rate is both very impressive and functionally useless. A ship that only has a near-miss or a collision on three or four days of the year when operating autonomously is a ship that will remain crewed with humans.
Orca AI's website (here) at the moment is ambiguous on the company's final direction. It clearly has a product with excellent potential to help human navigators make better decisions through the use of AI, and so it emphasises that it can increase situational awareness for navigators, reduce sensory workload, and "empower crew with automation."
Orca AI's software can independently recognise small fishing boats and small markers that are not captured by radar and are not equipped with AIS. The system measures the distance to these targets and notifies the watchkeeper on duty of danger of collision.
At the same time, however, Orca AI says it is "providing the technology to bring autonomous cargo ships to reality." I am not sure we are there yet or will be within the next decade except in very special cases like the Yara Birkeland example.
Please don't think that I am a Luddite. Too many marine accidents are caused by human error – from the bulker Wakashio running aground off Mauritius whilst the crew were trying to get a mobile phone signal in 2020 (here), to the unfortunate destruction of the wheelhouse of the Dutch Inland cargo ship Southford when the ship struck the bridge in Scheldt-Rhine Canal near Rilland in Zeeland in the Netherlands, on the evening of August 17 (captured spectacularly on video on Twitter here).
Any technology that can make life safer for seafarers should be embraced. The bridge video (here) of the collision between the Jersey-registered crewboat Njord Forseti and a windfarm tower in the Southern North Sea on April 23, 2020, reminds us that some watchkeepers sometimes fail to pay attention, either through fatigue, distraction, or other factors. The bridge video footage from the crewboat is alarming to say the least. How was there so little situational awareness from the Spanish captain? The Master of Njord Forseti claimed that he was "adjusting of settings on the VHF radio, which is mounted immediately to starboard of his seat."
As the Jersey flag state investigation, which you can read here, stated of his excuse: "this has not been positively determined and distractions caused by other means cannot be ruled out."
Perish the thought that he might have been browsing his smartphone or texting!
What is clear is that whilst transiting between windfarms at approximately 20 knots, the crewboat hit a turbine tower, and the impact resulted in serious damage to the vessel. Interviews by the flag state investigators with the both vessel's master and deckhand revealed that neither could remember the events immediately prior to the incident, as both were knocked unconscious by the force of the impact. They were both evacuated by air to hospital.
Orca AI's software, or similar systems, could undoubtedly have assisted by warning the watchkeeper of the impending danger. In this way, I see technology continuing to augment human navigation over the coming decade, just as radar and binoculars have done. The danger of distraction in a warm and comfortable ROC is arguably greater than on the bridge of a moving ship, where motion and noise provide immediate alerts.
On certain selected voyages like the Yara Birkeland's monotonous seven-mile circuit between Herøya and Brevik and back, a phased introduction of autonomous systems and remote control from an ROC is the way forward. For very small, light autonomous vessels performing bathymetric surveys, for instance, remote control is a complete no-brainer.
But for more complex voyage patterns, for bigger vessels, and for higher risk operations, like being in DP alongside a platform or rig, I expect the next decade to see humans still at the helm, hopefully assisted by newer and better technology, and not distracted by their phones.
If you are reading this on the bridge, get back on watch now!
Background Reading
Baird Maritime's full Focus on Unmanned Craft feature with reviews of several of the smaller units entering service is here.
More Yara Birkland information on YouTube here.
Our December 2021 review of Yara Birkeland is here.
More about Yara's efforts to decarbonise are here.
NYK's press release on its co-operation with Orca AI to test the lookout support system is here.
Surprise! OSM and Ocean Infinity are advertising for deck officers for the large 85m survey vessels that the latter will take delivery of in 2023. See here to become a member of the "lean crew" onboard. Autonomy is hard work!