Every generation of aircraft is more complex than the last, and managing this complexity within the design process is a fundamental challenge for airframers, suppliers and others within aviation. Systems modelling is a key tool in the aviation design toolbox, and is a candidate for realising some of the early benefits in artificial intelligence within the industry. We sat down with Pierre-Emmanuel Dumouchel, chief executive and founder of Dessia, fresh from being the first French startup to complete Boeing’s accelerator, Aerospace Xelerated.
To set the context, Dumouchel tells us, “In order to make engineering more efficient by 2030, it is understood that the solution will involve a strong automation of the design processes. Currently, all engineering tasks remain iterative and manual, but it is essential that engineering becomes focused on automation within the next 10 years.”
This is likely to include model-based systems engineering frameworks that engineers can use to understand the numerous interconnections between systems and their requirements. But as these get more complicated, the burden grows of manually producing design data, putting in place specifications, assessing scenarios, and calculating impact cascades: to systems, subsystems and components. Current models also often lack an incorporation of the service engineering and manufacturing processes that are inherently linked to them.
Critical to the successful implementation of AI is identifying the problems to be solved
Dumouchel explains that “the aviation industry needs to transform, fast, and digital technologies will be at the centre of this change. Of course, when we think about digitalisation, we foresee the product, the supply chain, the operations — what we don’t see is that 80 percent of what happens in the life of a plane is decided in the first years of its inception, in the upstream design phases, when technical choices are made that will constrain how we can manufacture, operate, maintain our aircraft.”
As a result, it’s no surprise that the growing digitalisation of design and manufacturing engineering is an early focus for integrating artificial intelligence into the aviation industry. Drivers here include the increasing complexity of the systems as well as the ongoing engineer shortage.
“Specifically to the aviation industry, the experts who went through an entire development process are on the brink of retirement, which adds up to the complexity. In such a context, how do you provide the brainpower required to enable the change in engineering?” Dumouchel asks. “Part of the answer lies in the ability to use software programs, and now artificial intelligence, to augment the engineer’s ability to perform. Those programs can almost instantaneously perform hundreds of design cycles. With these new capabilities, engineers will tomorrow be able to reduce the overall design time by several months, design right the first time, find more performing designs, automate change requests evaluations, and more.”
Dessia’s proposed solution is a new approach to systems modelling in this context, which it calls generative engineering.
“Using object-oriented modelling coupled with artificial intelligence and 3D generative algorithms, it allows us to exhaustively explore all possible combinations, to manage uncertainty and variability and to explore quickly all the functional architectural and 3D integration solutions meeting the requirements,” Dumouchel says. “The future of aviation engineering would then be for engineers to supervise digital robots — agents or companions are common terms in the industry — that will take care of the repetitive screening of millions of design solutions and recommend the most appropriate ones.”
The benefits of this approach include speed, with improvements on the time taken for requirement convergence, conceptual approach and architecture design. It also enables new approaches to solutions, especially those that would be a combination of complicated and time-consuming for humans (or, rather, teams of humans) to design. It can explore many more possible configurations than other approaches, iterate more often — and more rapidly — and inform structuring decisions upstream and downstream.
Accelerators are playing a key role in speeding AI and other innovations to market within the industry
There are clear advantages to this model, so it’s no wonder that it caught the attention of the Aerospace Xelerated software startup accelerator, backed by The Boeing Company, Tawazun Council, its industry partners GKN Aerospace, the UK Defence and Security Accelerator, and Etihad Airways.
“Since this year’s cohort was focused on software editors for Industry 4.0, Dessia was shortlisted out of about two hundred startups identified worldwide, and was finally selected along with ten other companies to be accelerated,” Dumouchel explains. “The program is a short, intensive, three-month mentorship framework that helped us engage with Boeing stakeholders, as well as with the surrounding aerospace ecosystem, and benefit from the program’s team support in doing so. Part of the benefits are the end goal of securing a pilot project with Boeing as well as adding the company to Dessia’s investors, through a SAFE [simple agreement for future equity] that is granted to each of the 11 startups selected.”
Initiatives like these help to soften the sharpest of the bleeding edges of technology, maturing early innovations into applicable engineering and industrial tools.
Within aviation, these kinds of technologies will lead to nothing short of a revolution as engineering tasks that are currently manual, iterative and/or time-consuming are automated.
“When we look at the software industry, we can see that this revolution is already underway through what is called continuous integration. Essentially, when a specification or a piece of code is modified, a set of verification and integration testing processes are automatically performed,” Dumouchel says. “By analogy, we observe an evolution of engineering towards these approaches, which can be referred to as continuous engineering. This paradigm shift will enable engineers to quickly adapt their products based on needs, as well as to find innovative solutions by analyzing a large number of generative hypotheses.”
“In summary,” he concludes, “our vision for engineering in 2030 is that the engineer of the future will be greatly augmented and assisted by automated AI-based processes, enabling them to tackle major issues such as reducing carbon emissions in products designed by engineers.”
Author: John Walton
Published: 29 August 2023
;
