De hull inspectie van de toekomst

Toekomstige inspecties zullen drone, Virtual Reality en Digital Twin technologieën combineren, lees meer. [English]

Today, assessment of the structural condition of a vessel mostly takes place during survey, i.e. during physical tank entry. New inspection techniques such as drones and self-localizing cameras could enable cost efficient full visual mapping of tank condition on a 3D ship model. A COMPIT 2017 paper by Christian Cabos et al. (DNV GL), www.compit.info, outlines a vision of how such techniques could allow structural condition assessment to be performed remotely in the (not so far) future, reducing risk, cost and time in such surveys. The outlined concept combines three key technologies that have evolved from research to maritime industry application over the past decade: the “Digital Twin” concept, Virtual and Augmented Reality, and drone technology including machine vision. Drone technology, Virtual Reality, and 3D Product Data Modelling are not new to maritime world, but it is the intelligent combination of these technologies that now unlocks the potential of smarter surveys.

Digital Twins form the backbone of planning and assessing
The “Digital Twin” is a computer model of the ship, in this case the hull structure. The concept has progressed from 3D models in CAD systems over product data models (adding required thickness following Class Rules, steel quality, as-built information, etc.) to the “Digital Twin” (adding corrosion history based on thickness measurements and inspections, “as-is” condition with associated strength and vibration performance, captured photos of surveys, etc.). The backbone software at DNV GL for this purpose is ShipManager Hull. Photos of the structure captured by a camera with tracking system (taken by a surveyor or a drone) can be mapped on the virtual structure in the “Digital Twin” resulting in a realistic representation of the structural condition at a given date. 

Drones will be our eyes on site
Drones are increasingly used for a multitude of applications, e.g. aerial photography. However, standard GPS technology for drone orientation is not applicable inside massive steel environments such as ships. Instead, drone orientation may be enabled through automatic indoor positioning technology, e.g. using 3D maps based on the Digital Twin. Remote connectivity will allow directing the drone to additional close-up capturing. E.g., suspect areas detected in the pre-scan imagery can be followed up through remotely advising the drone to re-visit specific locations and obtain more detailed imagery. On top of regular cameras, drones may supply infrared, ultraviolet or ultrasound measurements. For higher time efficiency, hull inspections would be preferably done in a pre-scan unattended by the class surveyor, following a prepared inspection plan. The resulting images could automatically be mapped onto a 3D model. As long as autonomous drone flight is not possible, pre-scanning can be performed through a remotely guided drone carrying an indoor positioning system or even a service technician carrying a camera system. This procedure is still much more efficient than the current survey practice.

Hull condition will be surveyed in Virtual Reality
Virtual Reality (VR) refers to the technologies to navigate through 3D computer model where stereoscopic vision and the associated depth perception adds to the realism. After the pre-scan, inspection data captured on board (images and measurement readings) can be attached to the Digital Twin 3D model. The surveyor can assess the hull condition in Virtual Reality from his office and even invite other stakeholders, such as colleagues or owner representatives, who are physically somewhere else, into the same virtual space for e.g. decision support and sharing insights. Or, in complicated cases, an expert from helpdesk can be brought into the same virtual space for decision support.

In such a virtual environment, the hull condition can be assessed and documented very quickly. Moving around (e.g. examining bulkhead plates from both sides by simply walking through the virtual bulkhead and turning around), accessing thickness measurement results, comparing to previous records, etc. becomes fast and easy. Automated algorithms for image recognition can direct the surveyor to locations with suspected cracks, deformations, heavy corrosion wastage, coating break-down, etc. Findings, repair measures, memo to surveyor, etc. can directly be recorded on the structure in the “Digital Twin” (such as the ShipManager Hull). With integrated reporting, no additional work is necessary for the surveyor after leaving the virtual ship – documentation is done when survey is done. Results can be presented and explained to the owner representative in a virtual meeting on-board the virtual ship in an efficient and most intuitive way. 

Not the same, but equivalent?
To what extent can Virtual Reality replace physical presence in the real world? A class surveyor on board during survey uses more than his eyes. Touch and feel (Is a surface smooth or wet? Is this shiny liquid oil or water?), smell (e.g. for smoke or cargo vapours in a closed space), and hearing (e.g. hammering on a surface to test the acoustical response) are used in surveys, but not an option in the outlined virtual surveys. For equivalent safety, future surveys would then need to compensate for the “lost senses”. One lever would be more frequent and more extensive examination of the structure in the easier and cheaper virtual surveys based on drone scanning. Another lever would be employing additional senses for the drone, such as hyperspectral imaging (= infrared and ultraviolet). Thereby corrosion or coating breakdown could be detected with higher likelihood than with the visible spectrum perceived by the human surveyor on site now. 

From vision to reality 
The technologies described above are in different development stages between prototype and field testing.
Today, hull inspection techniques using camera-equipped drones are starting to be accepted, e.g. for inspections of large cargo holds when visual confirmation of good condition is sufficient. The effort for survey preparation on board is thereby dramatically reduced. But so far, drone inspections require two persons for the survey: the drone pilot with clear line of sight to the drone and the surveyor interpreting the screen image. The surveyor also needs the line of sight to the drone to understand its current position. Drone inspection techniques for narrow spaces – where no line of sight to the drone is possible – and contactless thickness measurement are subject to research. Once in place, the technology would allow voyages or even short unplanned downtimes of the ship to be effectively used for the pre-scanning of the hull structure.
Several systems for automated indoor positioning of cameras and sensors are currently under development. Already images captured in closed spaces can be properly associated to 3D models. With some further development, the same technology will allow drones to orient themselves inside a ship with sufficient precision. 

DNV GL has developed a working prototype demonstrating how hull condition can be assessed and documented in a virtual space, based on its VR Survey Simulator SuSi. The prototype allows VR users to experience a water ballast tank of a VLCC. Users can examine emulated thickness measurement data and drone footage, search for locations with corrosion, cracks or buckling and document possible findings. Emulated data from previous inspections of the ship and sister vessels can be compared with the actual tank condition.