Shaping the Future of Aviation: CAELESTIS Workshop at the 14th EASN Conference: Leveraging Machine Learning, Virtual Prototyping and Manufacturing

As the aviation industry faces growing challenges in meeting future demand and sustainability goals, the 14th International EASN Conference on October 8-11 brought together over 500 experts from across the globe to Thessaloniki, Greece. With the theme of “Innovation in Aviation & Space: Towards Sustainability Today & Tomorrow,” the event was a combination of high-level discussion panels, presentations, and collaborative workshops.

During this flagship event, we had the opportunity to organize our CAELESTIS Multi-Actor Collaboration (MAC) Workshop, on the future of next-gen aviation through machine learning, virtual prototyping, and advanced manufacturing. We had four presentations from engineers, researchers, and industry stakeholders as well as an interactive Q&A session with all attendees on certain topics within the field.

With a vibrant atmosphere of innovation and collaboration, the workshop offered a sneak peek into how aviation will evolve, using technology not only to streamline processes but also to address the environmental goals set for 2050. As Ms. Marylin Bastin, Director of the European Green Sky Directorate at EUROCONTROL, powerfully put it in her keynote:

 

© RTDS Group

“On a peak day, we have between 30,000 to 75,000 flights. We expect to have up to 50,000 flights per day by 2035. So, if we don’t find any way to help us manage that complexity, it will not work.”

With those challenges in mind, CAELESTIS’s MAC workshop provided a timely and successful gathering to discuss how technological advancements can help.

Sebastian Rodriguez – expert from Arts et Métiers ParisTech, presented Hybrid Twins for the Resin Transfer Molding Process. This presentation dived into Hybrid Twins and particularly how they are applied to real-time control during the Resin Transfer Molding (RTM) process used in manufacturing composite materials for aviation. The hybrid twin approach fuses real-time sensor data with virtual models to provide unprecedented control and optimization during complex manufacturing processes.

o Real-time parameter identification: Using sensor data, the system can automatically adjust material parameters, improving the efficiency and precision of the RTM process.

o Material behavior predictions: Machine learning algorithms are employed to predict how the resin will flow inside the mold under specific conditions, providing control, minimizing errors and speeding up production.

In summary, Hybrid Twins are the future of composite manufacturing, offering real-time insights and adjustments to ensure quality and consistency in complex components like engine fan blades.

© Arts et Métiers ParisTech

AI and Hybrid Twins for Large-Scale Systems

Our partner, Mustafa Megahed from ESI Group presented a compelling case for using machine learning and AI-based hybrid twins in large-scale manufacturing systems. By reducing the time needed to set up workflows and improving the accuracy of predictions, AI is revolutionizing how industries handle increasingly complex systems.

© ESI Group

o Data reduction techniques: PCA and manifold learning are used to manage large datasets efficiently, which is vital for processing the immense amounts of data generated in aviation manufacturing.

o AI-based surrogates: Deep learning models, such as CNNs and GNNs , are driving the future of predictive accuracy in hybrid twin systems. These technologies allow manufacturers to optimize processes faster and more efficiently than ever before.

This work focuses on making better, faster, and larger systems possible through the integration of AI, which ultimately allows manufacturers to handle the complexities of modern aviation while minimizing costs and errors.

Digitalization Driving Aviation Manufacturing

In his presentation, Janusz Poplawski, representing Lortek from CAELESTIS External Advisory Board, delved into the critical role of digitalization in transforming aviation manufacturing processes. Digital tools are streamlining everything from design to production, allowing companies to predict and solve problems before they occur.

Poplawski presented several case studies, showcasing projects such as:

o Numerical and Topology Design Against Distortion: A digital approach that predicts and minimizes distortions during additive manufacturing, improving the quality and accuracy of aerospace components.

o The FLOWCAASH Project, which uses virtual prototyping to develop titanium alloy flow control actuators, crucial for improving performance in harsh environments.

o The FLASH-COMP Project, focused on composite manufacturing using AI-driven simulations and real-time monitoring tools, aiming for zero defects and reduced waste during production.

© RTDS Group

“The integration of AI, machine learning, and digital platforms is crucial in helping the aviation industry meet its 2050 environmental objectives.”

The adoption of these tools promises to enhance sustainability by optimizing manufacturing processes and reducing resource consumption.

Setting the scene for future exchanges with stakeholders

The CAELESTIS Multi-Actor Collaboration (MAC) Workshop was a success, bringing together stakeholders from diverse sectors to share their insights and expertise. This workshop also set the scene for future exchanges with stakeholders, with the next one planned at the beginning of 2025. The goal of MAC is to refine the CAELESTIS project’s outputs, ensuring they align with the latest technological advancements and industry needs. Throughout the workshop, participants engaged in insightful discussions on how machine learning, virtual prototyping, and AI can revolutionize the aviation sector, not just in terms of efficiency but also sustainability. These advancements will not only make aviation more efficient, but also help achieve ambitious sustainability goals set for 2050.

facebook
Twitter
Follow
Pinterest