For years, Norwegian platform supply vessels (PSVs) have transported food, equipment and materials to oil and gas platforms on the Norwegian continental shelf.
The fewer waves a hull generates as it moves through the water, the less fuel the vessel needs. This results in lower emissions and reduces operating costs for the ship owner.
Researchers at SINTEF Ocean in Trondheim are combining a curated dataset derived from historical model tests with new methods in data analysis and machine learning.
The resulting model uses a limited set of main dimensions to estimate the calm-water resistance of PSVs.
This makes it possible to estimate the hull resistance of PSVs more quickly and with greater precision. Henning Borgen, CTO at Sirius Design & Integration and partner in the ZeroLog project, is pleased with the development of the tool.
“This type of tool is closely tied to our core business and enables us to perform faster and more accurate calculations. For us, the development of such useful tools for industrial use is the main motivation for participating in research projects,” says Borgen.
A model that keeps improving
One advantage of data-driven models is that they can easily be updated when additional suitable data are available for this purpose.
“The method will continue to improve as suitable data become available for use in model development. So far, it has only been used to estimate resistance for PSVs, but the plan is to expand it with other tools as well,” says researcher Endre Sandvik.
Together with his research colleagues Thor Albrektsen and Øyvind Rabliås at SINTEF Ocean, Sandvik has worked through the ZeroLog project with the goal of developing the next generation of PSV concepts.
“The new method makes it possible for us and designers in the industry to evaluate many different combinations of main dimensions while assessing how they affect resistance and energy efficiency,” says Albrektsen.
The trio believes there is significant potential in combining test results with artificial intelligence, and that new methods for use in the maritime industry will continue to emerge. This methodology can also be reused to develop new approaches for other ship types.
An ocean of opportunity
Over many years, SINTEF Ocean has tested most vessel types in its towing tanks. For the new PSV resistance model, relevant results have been structured into a protected development dataset.
“The current method focuses on resistance in calm water, but we also want to look into similar approaches towards manoeuvring and wave-added resistance, which are difficult to assess at an early design-stage today,” says Rabliås.
The data made available to the maritime industry when they conduct basin tests are crucial, both for design work and for verifying vessel performance before construction. Now, these data gain new value as they collectively form the foundation for new methods in ship design.
A number of models already exist for calculating ship resistance based on measurements and data. However, these established methods are largely designed for traditional merchant vessels. The Norwegian maritime industry works extensively with specialised vessels, such as PSVs, and for these vessels the traditional methods perform far less accurately.
“That is why we believe the newly developed method will be a valuable contribution for achieving more accurate estimates of calm water resistance for offshore vessels in the early stages of a project,” says Rabliås.
Although tools based on computational fluid dynamics calculations (CFD)—that is, computer models showing how water flows around the hull—can calculate calm water resistance accurately, these methods are still time-consuming.
There is therefore a need for simpler methods that can be used early in the design process or integrated into optimisation routines. The new models, which make it possible to explore energy-efficient ship designs, will be made available to the maritime industry through the ShipX Ship Speed and Powering plugin, combining the new approach with an input-and-output structure familiar from existing empirical methods.
