Photo: Joachim Rode

Optimal aircraft wing designed on Europe’s biggest computer

Thursday 16 Nov 17
by Lisbeth Lassen

About the article in Nature

The article ‘Giga-voxel computational morphogenesis for structural design’ was written by Niels Aage, Erik Andreassen, Boyan S. Lazarov, and Ole Sigmund—and published in Nature on 4 October.

The publication marks the completion of the NextTop project, which VILLUM FONDEN supported with a grant of DKK 12 million (EUR 1.6 million) in 2011. When PRACE gave the research group access to their supercomputer in 2014, they were able to test the method they had developed in the NextTop project.

With the help of Europe’s most powerful computer, DTU researchers developed the optimum structure for an aircraft wing in under five days.

Four DTU researchers have calculated how to achieve the best and most resistant aircraft wing structure using the least amount of material. In essence, the method known as topology optimization identifies the strongest structures in relation to a specific load while employing as little material as possible. The researchers’ findings are described in an article in the renowned scientific journal—Nature.

Associate professor Niels Aage from DTU Mechanical Engineering compares the inner support structures in the aircraft wing model with structures found in nature—e.g. bones or the interior of a bird’s beak. These structures developed with the same aim—namely to reduce weight while providing adequate resistance to the stresses to which they are subjected. In this way, their shape and function are similar to the designed aircraft wing.

“A bird’s beak has evolved over a very long time,” says Niels Aage. “Evolution isn’t necessarily smart—or rather—nature can accommodate flawed mutations and thus move in multiple directions at a time. However, every time we take a step forward in the development of our design, we move in the right direction—so you might say it’s a clever form of design evolution.”

In the same way as a bird requires fewer resources and less energy with a light and strong beak, the research group ended up with a lighter and stronger design capable of reducing the aircraft wing’s weight by 2-5 per cent—or 200-500 kg per wing. This translates into a fuel savings of 40 to 200 tonnes annually.

Large and small LEGO bricks
Niels Aage has long worked on developing models that calculate optimal solutions for very large structures—a task which the research team shared and which led to the development of a code that proved capable of solving problems associated with very large structures. With access to a supercomputer via the Partnership for Advanced Computing in Europe (PRACE), the researchers were able to apply their code in practice, as the supercomputer can handle large models in extremely high resolution.

"It corresponds to going from building something with DUPLO bricks to using ordinary LEGO bricks."
Niels Aage, associate professor at DTU Mechanical Engineering

Photo: Joachim Rode

“It corresponds to going from building something with DUPLO bricks to using ordinary LEGO bricks,” explains Niels Aage.

The choice fell on an aircraft wing because Niels Aage has been working in this area and because the rest of the group is also interested in aircraft.

“We concluded that if you take the wing of a Boeing 777 and disassemble in into over 1 billion pieces, you can actually arrive at the design tolerance—which is fun to look at, as it ceases to be an academic example.”

Dr Lee Margetts, Vice Chairman of PRACE, has plenty of praise for the article’s authors:

“The article stands out because it closes the so-called ‘high performance computing gap’ in the manufacturing industry. We’ve lacked structural engineering on HPC systems because the existing software wasn’t able to cope with scaling large elements. The research conducted here has the potential to renew many of the industrial sectors that underpin the EU’s economy.”

Many different materials
Many large structures such as bridges, wind turbines, or offshore structures comprise several different materials—something the model can accommodate.

“In this case, our calculations were for an aluminium aircraft wing. It is a wonderfully isotropic material, which means that it does not vary in magnitude according to the direction of measurement. However, we could have conducted our experiment with other—and more materials. We might decide to look at laminates and sandwich structures, as the model can also handle these,” explains Niels Aage.

In other words, the model has a very broad field of application on large or multi-material structures.

“Naturally we’ve described our approach point by point in the article so that everyone with the relevant background can repeat the method.”

Photo: Joachim Rode

News and filters

Get updated on news that match your filter.