Lighter ships that adapt to the conditions at sea can ensure more sustainable shipping. This is made possible by the rapid data processing of machine learning.
Anyone who understands the sea knows how ships have to be built for it. How high are the waves, how strong is the swell? "If such data is linked with data on the construction of ships, new ships can be optimally adapted to the conditions on the water," says Prof. Norbert Hoffmann. He is conducting research in an interdisciplinary project at Hamburg University of Technology and is using the possibilities of machine learning to link the key figures provided by the sea with those from shipbuilding. "For the first time, we can make concrete calculations. This is because it is possible to accurately process the huge amounts of data required for this," says the TU scientist.
A digital twin helps scientists at the TU Hamburg
The project "Predicting Ship Hydrodynamics to Enable Autonomous Shipping: Nonlinear Physics and Machine Learning" combines the two institutes for Ship Structural Design and Analysis and Structural Dynamics, whose head, Prof. Norbert Hoffmann, is an expert in waves. In his research, he is supplied with the necessary data from shipbuilding by shipbuilding expert Dr. Franz von Bock and Polach. As a first step, the scientists are constructing a digital twin from this data. This will help them to build a ship that can move as optimally as possible on the water. Once the dynamics that affect the hull are known, the design can be adapted accordingly. Von Bock and Polach says: "We don't even know the real loads that ships are exposed to on the water. That's why their steel structures have so far been designed to withstand all conditions in any case." None of this is particularly sustainable, as the average lifespan of ships is 25 years at most, although it could be significantly extended with a design tailored to the conditions. The researchers are focusing on medium-sized ships with a steel structure.
TU researchers measure the North Sea as a wave field
In a second step, Hoffmann has even bigger plans: He wants to measure the entire North Sea in such a way that, for the first time, it should be possible to map the conditions at sea in real time and scientists will no longer have to rely on predicted mean values. Hoffmann is optimistic and explains how he is proceeding: "The North Sea is a relatively small sea with a total of around 2,000 waves, each of which is between 100 and 200 meters long. With the help of nautical ship radar, we translate the measured data into wave movements, which in turn can be used to create the sea as an entire wave field. If we succeed, we will be able to describe the North Sea deterministically and map the wave movements in real time, says Hoffmann. That still sounds visionary, but the rapidly increasing processing speeds of machine learning show the TU scientists that they are on the right track.
Sustainable: extending the service life of ships
However, there is one big unknown in this game that can upset current calculations: climate change. It is warming the oceans and changing waves, currents and winds. This can be seen from the climate models. "You have to know the sea state," explains Prof. Hoffmann. "We get a lot of information about the interaction between the sea, wind and waves via radio buoys and the nautical ship radar. The data obtained from buoys and ships must be evaluated if we want to make accurate wave forecasts." Although the scientists include a lot of the available data in their calculations, the level of prediction is becoming somewhat more difficult due to the influence of climate change. "We include safety factors in our ship design, but we assume that significantly less steel will be used, which will make the ships lighter and consume less fuel," says von Bock. "If, together with a design that is better adapted to the waves, the service life of the ships is significantly extended, this would be a major step towards greater sustainability for all means of transport used at sea."
At the end of the interdisciplinary project, the scientists want to work together with the DLR Institute of Maritime Energy Systems, further develop the digital twin over the next few years and then construct the first ship models based on the new calculations.
More information
The I3 project "Predicting Ship Hydrodynamics to Enable Autonomous Shipping: Nonlinear Physics and Machine Learning" combines the two institutes for Ship Structural Design and Analysis and Structural Dynamics. The I3 program stands for interdisciplinarity and innovation in the engineering sciences. The aim of this program is to identify new interdisciplinary projects and promote them to such an extent that the projects can subsequently attract external funding.
