IEEE Robotics & Automation Magazine - September 2023 - 18

ship and shipboard crane were received
in real time via a 4G connection, as
detailed in Figure 3. Figure 12 depicts
the corresponding four channels of crane
data for the whole operation, based on
which the crane's status can be evaluated
and used in the twin system. The three
digital twins in Figure 11 are all realtime
representations of the same system,
consisting of the ship and crane subsystems.
The blue system on the left mirrors
the real crane, but the movement of its
hull is determined by harsher weather
controlled in the simulation. The ship
and crane in the middle are the exact
presentation of the real system in realtime
operation. The hull of the green
ship follows the movement of the real ship, but its crane is
controlled by an operator in the remote center to test an alternative
path trajectory in the lifting operation. In this way, it
allows the direct testing of " what-if " scenarios based on a
real-time current state of remote systems. All the functions
have been integrated into the NTNU forskingslab in Ålesund
based on cooperation and support with the Offshore Simulation
Center. A real-time, online digital twin will enhance
operation safety and security in the following aspects:
20
15
10
0 1,000 2,000 3,000 4,000
Time (s)
100
50
-50
0 1,000 2,000 3,000 4,000
Time (s)
200
100
-100
-200
0 1,000 2,000 3,000 4,000
Time (s)
400
300
200
100
ngl
gl
ng
5,000 6,000 7,000
Slewing AngleS wing Ang
nglgle
Outer Boom Angle
OOuter Boom Angle
Main Boom Angle
Ma nBoo
Ma n Boo
Maiiain Boom Angleglle
Angle
"
ALL THE FUNCTIONS HAVE
BEEN INTEGRHATED INTO
THE NTNU FORSKINGSLAB
IN ÅLESUND BASED
ON COOPERATION AND
SUPPORT WITH THE
OFFSHORE SIMULATION
CENTER.
„
■ It will explore the limits of procedural
safety by simulating the operation in
more extreme weather conditions, with
identical current crane loads, deck
arrangement, and ship ballasting.
■ It will find a proper way or safer, alternative
ways of performing the operation
under certain harsh weather conditions.
DISCUSSION
To build a digital twin ship, we propose
an architecture and concept using an
open model and simulator platform. As a
prototype, the Gunnerus digital twin
combines expertise from complementary
disciplines including sensor technology,
artificial intelligence, computer science,
and the maritime industry.
Many challenges arise when trying to fully represent a physiExtension
Length of Outer Boom
5,000 6,000 7,000
5,000 6,000 7,000
cal counterpart by the twin system in a stochastic environment.
One of the challenges is high-fidelity models. As learned from
the Gunnerus twin ship, the amount and quality of data are fundamental
to establishing models for the twin system, e.g., either
for identification of system parameters, or for development of
ML models. Therefore, it is necessary to establish mechanisms
to update models regularly, especially for these ML models.
Thus, the computational cost of ML
models, such as the training time for the
model in the " PHM System " section,
can be relatively low, and the model is
able to adapt to operation/environmental
changes. Another challenge lies in
cosimulation for digital twin implementation;
for example, scalability shall be
considered when it comes to industrial
use of the digital twin. Insufficient synchronization
and tight coupling issues
may arise if the system becomes too
complex and requires high computational
effort. A domain-specific challenge
in FMU implementation also
needs to be discussed.
The Gunnerus twin ship is our
best attempt at digital twin systems
for marine operation. Even though
there is still room for improvement in
terms of accuracy and efficiency, we
believe that the implementation will
enlighten both the maritime industry
and academia in developing twin
ships in the near future.
0 1,000 2,000 3,000 4,000
Time (s)
5,000 6,000 7,000
FIGURE 12. Four channels of sensor data for crane-status monitoring.
18 IEEE ROBOTICS & AUTOMATION MAGAZINE SEPTEMBER 2023
CONCLUSION
This article illustrated a digital twin of
the RV Gunnerus. All the related key
technologies, including cosimulation,
Outer
Main
Slewing Angle (°)
Boom Angle (°)
Boom Angle (°)
Extension
Length (m)
IEEE Robotics & Automation Magazine - September 2023

Table of Contents for the Digital Edition of IEEE Robotics & Automation Magazine - September 2023
