There is no way that you could say that the field of naval architecture is set in its ways. On the contrary, it is a subject that is in a state of constant development. New technologies and methods are discovered and brought into play, helping naval architects and engineers with their goal of designing a vessel that effectively meets its design requirements.
Research that is currently being carried out at C-Job Naval Architects/Engineers illustrates this point, says Thijs Muller, one of the company’s Naval Architects. “We are investigating how genetic algorithms can be used to automate and optimise the concept design stage. We are on the way to achieving even more accurate estimations of construction weight, building costs and performance in addition to generating concept designs within shorter time spans.”
As the name suggests, genetic algorithms borrow certain principles from the field of biological genetics. This demonstrates that the concept of ‘survival of the fittest’ also exists in the world of naval architecture and engineering. To show how genetic algorithm are utilised, Thijs uses an example of 50 generations of ship designs, with each generation containing 20 individuals, creating a design space of a total of 1,000 designs that include all feasible and unfeasible, as well as winning designs.
Darwinian ship design
“After the first generation there are 20 designs; each of which is checked for feasibility. The unfeasible designs ‘die’ and are not allowed to pass their genes – their design characteristics – onto the next generation. The feasible designs are the winners; they are ranked according to their performance and are allowed to pass their genes on. We create progressive generations by combining the survivors from the previous generation with an amount of random design variation. The resultant ‘gene pool’ of ship designs gets better with each generation.”
C-Job is carrying out this research in cooperation with software design house NAPA. “This is at the core of our integrated design process, enabling us to use optimisation routines in combination with an integrated design approach.” The research has a clear focus of attention: the optimisation phase of the concept design stage of ship design. This is when different parameters are tested in order to find the optimum for a given objective. “You can change the parameters to fit the project requirements,” explains Thijs. “They can be fixed parameters such as a payload or deadweight. Or they can be variable; the length of the vessel, for example. Of course, the more variables you want to optimise, the more complex the study becomes.”
Reducing Total Cost of Ownership
Because different vessels perform different functions with different mission equipment, the type of parameters can also be adapted. Cargo capacity and fuel consumption are important for a container vessel, and bollard pull and manoeuvrability are key aspects of tug design. “We have worked with a Trailing Suction Hopper Dredger design for our initial research. Using the two parameters of weight and resistance, our main objective has been to optimise the return on investment. We want to accomplish this by maximizing performance by decreasing resistance, whilst minimizing the steel weight and consequent building costs. There will always be a trade-off between these two parameters; you cannot optimise one without sacrificing the other, so the best designs lie on the so-called pareto frontier, named after Vilfredo Pareto.”
C-Job’s research into accelerated concept design is more than just producing designs in a shorter period of time. It also has the potential to reduce a vessel’s capital expenses, which are largely defined by weight, and operational expenses, which are determined by its operational efficiency with parameters such as fuel consumption, cargo capacity etc. “Of course, the ultimate goal is to reduce the Total Cost of Ownership for the ship-owner,” says Thijs. “But accelerated concept design is also about removing uncertainty from the whole ship design and construction process. It is about giving the ship-owner confidence in its ship and reducing the risks.”
From the perspective of the naval architect, accelerated concept design can point the designer into a certain design direction where improvements can be found. “It can serve as an inspiration to consider more ‘out-of-the-box’ designs – ones that may perform better than the original idea.”
The graph above – regarding the initial research carried out by C-Job with a Trailing Suction Hopper Dredger design – shows some encouraging results. “This proof of concept clearly shows a pareto front that contains optimum designs. Genetic algorithms have successfully identified a feasible design space,” notes Thijs. “Our next step is that we want to move this concept out of the academic world and more into real world applications. This will involve increasing depth of detail and improving the algorithmic process. In the future, we estimate that we will be able to create four feasible concept designs in two weeks.”
Learn more during the C-Job webinar