Racing for industrialization with Maud Chidiac
Interview with Maud Chidiac, Program Manager AI at AMEXCI AB, Karlskoga, Sweden
Who are you?
I am Maud, passionate about technology development and about the strong and positive use cases AI can bring to multiple fields. I have been working before in the field of autonomous vehicles in Berlin, connected transportation in Paris and now AI for AM in Sweden at AMEXCI. I love to explore new topics and come up with creative solutions to ensure project delivery.
What is Amexci?
AMEXCI is a Research and Development joint venture owned by a consortium of eleven Swedish companies in a broad range of industries including energy, forest, defence, transporation.
It was initiated in 2017 by Marcus Wallenberg with the purpose of exploring collaboration as a mean to speed up the adoption of the Additive Manufacturing technology, offering Research in the field of AM (from printing parameter development to material properties), Development (re-design and printing iterations in both metal and polymer with an internal park of machines) and Innovation (Guiding the customers through their AM journey with education programs).
What is your role at AMEXCI?
At AMEXCI I work as a Program Manager for Artificial Intelligence. My role was first to identify the potential applications of AI for AMEXCI’s businesses and for AM in general. For two years now I have led the development of a proof-of-concept for aluminium L-PBF defect detection using monitoring data, through different phases and with partners.
What research topics are you/your group currently working with?
I am currently working with analyzing defects from AM process data that one can extract from commercially available monitoring systems.
AI is a very hyped technology what will it mean for the AM industry?
AI techniques are getting more mature and accessible with more people trained to handle them. This represents a strong potential for the AM industry that is always more digital with more data that can be extracted.
Bridging the gap between AI and AM can bring tremendous value with early-stage defect and crash detection during the print, as well as ensuring that parts are compliant with quality requirements for L-PBF. AI can enable an intelligent analysis of the layer-by-layer pictures being recorded on real-time.
What would you say are the most important barriers for industrialization to overcome for the metal AM industry as whole?
The biggest challenge for companies moving from prototyping into serial production with AM is the qualification of printed components.
Today the traditional Non-Destructive and Destructive testing method suffer limitations (costs, lead time, ability to see defects). Machine manufacturers are developing monitoring systems recording valuable data but they lack intelligence for companies to use them. We believe AI applied to AM can support AM quality control.
You often talk about the digital thread in AM. What is it and why is it important for the industry?
AM being a digital thread means that process data can be extracted from all steps of the AM workflow. All data from print parameters to monitoring (sensors and cameras of all kinds) and testing data can be used as input data for intelligent algorithms that can generate useful information.
We at Interspectral work with making AM process data accessible to user in a easy way. What is your advice to us, what should we focus on? Which is the best way we can help the industry.
We at AMEXCI are really impressed by the capabilities of AM Explorer and we believe that we could at some point integrate the outcomes from our AI analysis into it for visualization of the identified defects. Maybe one day the data integration on AM Explorer could be done on real time during the build?
Learn more about the work at Amexci here:
https://amexci.com/open-labelling-process-annotating-metal-lpbf-data/