Industrial Challenges in Geometry Optimization for Additive Manufacturing
Interview with Johannes Pauli
Research Engineer – Vehicle Concepts and Technologies at BMW Group
Johannes Pauli, a specialist for vehicle structures within BMW’s research division, focuses on implementing generative design methods across concept and series development phases.
Part of the team shaping future vehicle concepts, Johannes is also responsible for the adoption of new manufacturing technologies for structural components and algorithmic development methods, including the Wire Arc Additive Manufacturing (WAAM) process, for use in BMW Group production vehicles.
As he prepares for his keynote presentation at the CDFAM Computational Design Symposium in Berlin, we delve into his role at BMW and the innovative contributions he will discuss during his address.
Could you please begin by describing your current position at BMW and the scope of your role there?
In my current role as structural engineer within vehicle research, I am on a mission to design structural components using cutting-edge technologies that can shape vehicle concepts of the future.
At BMW, we are constantly looking for advanced manufacturing solutions, such as Additive Manufacturing and WAAM, thriving to develop better products.
However, using these technologies brings new challenges in the development process of structural components. To address those challenges, we are constantly developing and applying innovative methods to create optimal designs in the least possible time.
Working at the research end of the design process which is open to experimentation and exploring new tools and techniques, what criteria do you employ to assess and determine the readiness of new tools and techniques for integration into the broader engineering workflow?
The barrier to entry is relatively low, if a significant unique selling point is given. We have benchmark cases that go down to component level. However, integrating new tools and techniques into our series development workflow can be quite challenging, especially when using full vehicle simulation with very specific modeling guidelines.
BMW has developed its proprietary shape optimization code to bridge the gaps left by commercially available design and engineering software. Could you discuss the functionalities and advantages this custom software offers that are not found in off-the-shelf solutions?
We are constantly developing new functionalities.
During my presentation at CDFAM I will give some insights about two AM specific response functions, that we implemented. Even though I cannot share too many details about new functionalities in our pipeline, the biggest advantage is that we can prioritize feature requests ourselves and have an unbeatable time to (our internal) market.
What additional features would you like to see implemented in design software, either within BMW or commercial applications, and what other design challenges do you believe need addressing?
It would be greatly beneficial, if the design software could encompass the entire development process from early phases and blank paper to the final assessment of manufacturability and cost just before start of production.
While we have seen some improvements over the past years, manufacturability for series processes (not only AM) definitely needs more attention.
With the advent of AI engineering tools on the horizon, many organizations are recognizing the gap in the requisite data needed to leverage these tools effectively once they’re accessible. How is the work you’re currently undertaking facilitating the preparation and accumulation of data essential for training machine learning algorithms in the future?
At BMW, we have large databases containing all kinds of data from the development process, so data quantity is not an issue. In my opinion the question is: Do we have the right data at the right quality to predict useful things for future products, that might look a lot different than what we know today.
There will be a number of presentations from experts in the field of AI in engineering from academia and industry. What do you hope to learn from them and other presenters at CDFAM Berlin?
I am looking forward to see how far other experts and companies pushed technology in this field. Certainly, using AI to generate geometry from scratch is not an easy topic.
Multiple examples from academia demonstrate the advantages of these methods and I would be happy to see a more comprehensive use in industrial applications, that meet requirements regarding structural performance, weight, cost and manufacturability.
Finally, What key insights should attendees expect to gain from your presentation at CDFAM regarding the Industrial Challenges in Geometry Optimization for Additive Manufacturing?
My presentation will provide some insights into our metal AM parts in series production and their development process utilizing numerical optimization methods. I will deep dive into the challenges we faced, which might not be trivial to computational engineers.
I will close with an outlook on future technologies like WAAM and the unique challenges associated with this technology. I am really looking forward to discuss these topics with leading experts from so many different industries and learn about their perspectives at CDFAM!
Register to attend CDFAM to learn more about the work Johannes, his team and others at the forefront computational design and advanced manufacturing at all scales. The two-day event features over 30 expert presentations along with networking opportunities and demos from leading software developers from around the world.
