Passenger ship operators can monitor vessel performance and machinery condition, while digital twins model hull structures
Digital performance, standards and internet of things (IoT) connectivity is becoming increasingly important in passenger shipping as operators learn to work remotely. The sector’s implementation of digitalisation has accelerated as it struggles with technical and market challenges during the global coronavirus pandemic.
In the short term, demand for cruise ships and ferries has plummeted, but when it return,s ships will have more automation with higher levels of remote technology. And passenger ship operators will have more real-time monitoring, diagnostics, software and analytics for remote decision support.
DNV GL chief executive for maritime Knut Ørbeck-Nilssen says the global Covid-19 crisis has forced shipping companies to implement remote working and digital technologies for business continuation and vessel optimisation. This has resulted in the sector entering a renaissance period for innovation. “Digitalisation [has been] turbocharged, accelerating cross-industry collaboration and innovation,” says Mr Knut Ørbeck-Nilssen.
“The pandemic has accelerated digital developments by half a decade,” he adds. “New technology has spurred developments in digitalisation,” and created opportunities for operators, vendors and class.
Suppliers of unmanned aerial vehicles and remote-control services have found new markets in ship inspections. “There are great opportunities with drones for inspection, taking video footage and thickness measurements,” says Mr Ørbeck-Nilssen.
Further, shipping companies and original equipment manufacturers are using artificial intelligence for predictive maintenance, intelligent scheduling, real-time analytics and improving performance.
“The mentality has changed and [maritime] is open to digital ways of working and interaction,” says Mr Ørbeck-Nilssen. Automation on ferries will also increase, in some cases to the level of systems on cruise ships.
“There is no doubt automation will continue to be developed on board,” he says. “Fundamental to this is connectivity and bandwidth to facilitate this gradual development in modern vessels.” Central to IoT and vessel automation are the sensors on board, communicating data to shore. This enables owners to “do more diagnostics using AI machine learning for predictive maintenance of equipment” says Mr Ørbeck-Nilssen.
DNV GL has benefited from digitalisation with higher demand for remote inspections using drones and video cameras. This reduces the need for surveyors to travel to ships for physical surveys.
Digital twin modelling
Digitalisation also enables DNV GL to develop digital twins of ships to manage hull inspection programmes and identify sections of passenger vessels at risk of structural damage.
Digital twins can help owners, operators and class societies predict the structural performance of ships, says DNV GL principal specialist Gaute Storhaug. “Structural damage can have severe consequences,” he explains, such as hull breakup and sea pollution. “We have inspection regimes to mitigate these risks.” This involves full inspections every five years by class surveyors, but this could be extended to seven years for certain ships.
“We want to have an inspection regime that relates to the risk of your vessel,” Mr Storhaug says. The design assumptions could differ from real operations. Two similar ships, operating in varying environments could have different risks of structural damage, such as fatigue cracks to hull structures.
For a ship operating in calm environments, the period between class surveys could be extended. But for passenger ships operating in harsher environments, this period could be shortened, with inspections focused on the most vulnerable points.
Digital twins of ships could be used to monitor structural performance. DNV GL has an indicator tool to calculate the risk of fatigue cracking and overloading on ships within a fleet.
“This uses machine learning, ship position (AIS) and global wave data, so we can reveal which ships are most at risk among a fleet,” says Mr Storhaug. A traffic light system is used – with green, orange, and red indicating the most risk – but it has limitations.
On top of this, DNV GL created a numerical twin of a ship’s hull structure based on design models which matches ship positions with wave data.
Engineers can virtually slice ship models to reveal critical details such as stiffeners for inspection. Another traffic light system indicates “where you should focus inspections” across the ship’s hull although there are uncertainties such as the wave conditions these structures face in real operations.
To account for this, DNV GL created a hybrid twin. “We mix the design model and the numerical twin with sensory information,” says Mr Storhaug. “This brings much better accuracy and complete modelling of the structure” including transferring loads across the structure.
By optimising the sensor positions, shipowners can gain more detail of stress on hull structures. “One of the benefits of the hybrid twin is you can calculate where you should have the sensors,” says Mr Storhaug. Owners can use this information to reduce operating and compliance costs.
“You can use this for maintenance planning or inspection planning,” says Mr Storhaug. “It can be useful for lifetime extension and troubleshooting. There can be many purposes for these different digital twin concepts.”
Predictive maintenance
Digitalisation technologies enable equipment manufacturers to diagnose issues before they become problems on cruise ships and ferries. Rolls-Royce Power Systems is developing predictive prognosis capabilities to diagnose engine problems before they impact operations, says its director for application engineering and automation marine and defence business Kevin Daffey.
“Ultimately, we want to move to where we can provide condition-based asset management based on how an engine is operated,” he says.
To achieve this, Rolls-Royce Power Systems is using machine learning techniques, specifically a neural network trained on “about three months of good engine data" from a testbed with sensors to detect anomalies.
Corresponding alerts are set for the anomalies and if the machine learning model has previously classified the cause of the anomalies, it provides the alert which then triggers action from a remote operating centre and provides advice back to the engine operator.
When a new anomaly pops up, experts use the data to classify it, aiding the predictive maintenance software’s progress and building a library of potential issues for recognition and diagnosis by the software.
“That means we can update the models for all the engines and all the operators," Mr Daffey says. “Our body of knowledge is continually updated and captured forever, not only for that customer, but for all customers of a particular engine class,” he said.
Rolls-Royce Power Systems is using the approach to develop a holistic diagnostic system covering engines, gear boxes, waterjets and other components for a full powertrain health management solution, according to Mr Daffey. This could be linked to the company’s shore-based service for providing advice and intervention.
Ultimately, that information can be combined with external and real-time data on conditions and vessel positioning to offer operational insights to optimise fuel consumption and other critical aspects to help operators make timely decisions.
Mr Daffey says his company sees these data-driven elements as enablers for remote operations where decisions are taken from shore, leading eventually to autonomous operations where ships (or the software that powers and monitors them) “take decisions themselves”.
Mr Daffey says its Artificial Chief Engineer development programme is at “technological readiness level four” and the company hopes it will “become a reality over the next 24 months to enable autonomous technology of the future”.
Sensor packages
These prognosis solutions require packages of sensors. A propulsion monitoring system integrates real-time fuel consumption and maritime condition monitoring with cloud storage and analysis applications, according to VAF Instruments director of research and development Erik van Ballegooijen.
By installing fuel monitoring sensors on ships, owners can replace noon reports with real-time monitoring. “It is good to go for automatic data collection, to show this data on board and to store this information somewhere in the cloud,” he says.
Data should be displayed on board for crew to make immediate performance improvements. It could then be analysed by VAF, a third party or the owner to identify trends across the fleet.
“Once in the cloud, you can do a lot of additional things, like data enrichment, creating key performance indicators and long-term trend visualisation,” says Mr van Ballegooijen.
Shipping companies and technical experts can access cloud-stored data to analyse fuel consumption across fleets of ships and add information from other sources.
“You can get external data, like weather data, or you might also have your own business intelligence tools,” he continues.
These resources can be connected to propulsion management tools or sent to charterers for analysis. “Exchanging and combining data is valuable for analysis,” says Mr van Ballegooijen. “It is also important to find the quick wins and know which data is immediately relevant.”
For propulsion performance management it is important to have a minimum dataset, including real-time measurements of fuel consumption from flow meters on the fuel supply lines. Other measurements could include speedlogs, GPS and ship draught. External condition information can include wind, wave and current data from onboard sensors or external databases.
This information can be fed into a cloud system such as VAF’s IVY propulsion management solution that uses Microsoft Azure for data cloud storage and applications.
“With this minimum dataset, you can create ship-speed and fuel consumption curves,” says Mr van Ballegooijen. Shipowners and operators can “create baselines and see improvements or why things are going wrong” to decide how to optimise operations further.
Digital twin enables waste heat recovery system development
Lappeenranta University of Technology (LUT) is developing a digital twin of a waste-heat recovery system on a cruise ship. LUT associate professor Teemu Turunen-Saaresti
outlined the project’s progress at Riviera Maritime Media’s ‘How digital twins drive vessel efficiency and voyage performance’ webinar in May 2020.
LUT’s digital twin represents the waste-heat recovery system set up in its test centre. This system could be used on cruise ships to produce energy from excess waste and LUT is testing its effectiveness.
The digital twin was generated for an organic Rankine cycle physical unit. Data to create the 3D simulation was collected and transferred to cloud storage using Beckhoff’s technology.
Mr Turunen-Saaresti said this digital twin was based on a dynamic model using Simulink.
“We made multiple pressure and temperature measurements available so we can calibrate our model,” he said. Simulink was used for its suitability for dynamic modelling of thermodynamic processes.
LUT’s digital twin required computational resources for online monitoring, predictions and 3D modelling. It also needed capacity to increase the metadata of system components such as heat exchangers and pumps.
“The different components, thermodynamics, fluid dynamics and rotor dynamics - all kinds of behaviours needed to be considered,” said Mr Turunen-Saaresti. “This is why we needed a versatile and dynamic digital twin.”
He shared the main findings from this process. “Data is crucial for this digital twin,” he continued. “We needed to get the data from real systems.” But, data from ships has its limitations. “There are limited measurement points and a variety of data acquisition systems installed,” said Mr Turunen-Saaresti.
One challenge in developing digital twins from operational data is securing access to the data. “A question is, who owns the data?” asks Mr Turunen-Saaresti. “There are many different stakeholders and companies working in the sector and the ship operators, so ownership of the data is not clear, or who has access to the data,” he added.
A solution is having close co-operation between the various stakeholders to create a dynamic and versatile digital twin that can model complicated systems on ships.
© 2023 Riviera Maritime Media Ltd.