Interspectral
Interspectral and EOS collaborate
Last month, Interspectral showcased its AM Explorer platform at Formnext in Germany. The software specialists announced the reaffirmation of its partnership with EOS and also exhibited the latest features on its data analytics and visualisation software platform.
Following on from the event, Manufacturing Quality had the opportunity to speak to Isabelle Hachette, CEO of Interspectral. We discussed topics including how the AM Explorer platform is helping to provide quality assurance within additive manufacturing (AM), the challenges faced by the industry, and the potential that artificial intelligence (AI) has to innovate manufacturing processes.
Interspectral
Isabelle Hachette, CEO of Interspectral.
MQ: Why did you develop an AI-based approach to anomaly detection for AM Explorer?
IH: Our software platform solution, AM Explorer, is designed to empower customers in the additive manufacturing market by enabling them to detect, visualise, analyse, and communicate anomalies in their prints. The qualification requirements in industries such as aerospace, defence, and automotive are exceptionally stringent. To meet these standards, a robust and precise quality assurance approach is critical.
AI allows us to go beyond traditional methods by automating anomaly detection, improving accuracy, and providing actionable insights. It bridges the gap between raw data and meaningful analysis, enabling users to maintain the highest levels of quality and reliability in their production processes. It’s all about reproducibility and traceability.
MQ: How does AM Explorer address the unique challenges of additive manufacturing?
IH: Additive manufacturing, particularly techniques like Laser Powder Bed Fusion (LPBF), introduces unique defect challenges that are not typically encountered in traditional manufacturing. Common issues include lack-of-fusion porosity resulting from insufficient laser energy, keyhole porosity caused by excessive laser energy creating unstable melt pools, balling where molten material fails to spread evenly, and gas porosity arising from trapped gas during the process. AM Explorer connects to the 3D printer and addresses these challenges through advanced image analysis and AI-driven capabilities, offering users deep insights into the root causes of defects. This enables proactive quality improvements, ensuring that every print adheres to high-performance benchmarks and stringent industry standards.
MQ: Why did you choose to collaborate with AMEXCI when developing the AI features?
IH: AMEXCI is a long-term partner of Interspectral and an established service provider in metal additive manufacturing. Their R&D expertise and experience in serial production make them an ideal collaborator for developing and testing AM Explorer’s advanced features.
Their team’s focus on training, workshops, and defect detection aligns perfectly with our mission to push the boundaries of additive manufacturing intelligence. Using AMEXCI’s facilities allowed us to refine our features in a hands-on, real-world environment.
MQ: The modular solution that came about as a result of the project utilises artificial intelligence. How big of a part does AI play in the defect detection process?
IH: AI is at the core of our defect detection process. Our data analytics module is an AI-driven solution designed to automate and accelerate the quality assurance workflow.
The system automatically detects the defects and helps the end user to categorise them, reducing the reliance on manual inspection and improving accuracy. Additionally, the modularity of our platform allows customers to integrate their own AI modules into AM Explorer, making it highly adaptable to various workflows and needs.
MQ: Building on the previous question, do you believe AI holds the key to the future of defect detection in manufacturing?
IH: AI has transformative potential in additive manufacturing, from real-time defect detection and advanced data analysis to predictive modelling and efficient data management.
With machine learning, manufacturers can not only identify defects but also uncover patterns that lead to them, enabling proactive improvements in processes. AI will be central to achieving smarter, faster, and more reliable quality assurance systems Shape.
MQ: Did metrology play any role in the R&D phase of the project?
IH: We didn’t use classical metrology techniques. Instead, we focused on image analysis, visual inspection methods, and machine learning models to achieve our goals. This approach aligns with the digital nature of additive manufacturing and our commitment to leveraging cutting-edge technologies.
MQ: In the press release of the new solution, you mention that the solution is “already running live at customer sites.” Any particular areas where the solution is excelling or areas for improvement?
IH: Our AI module is live at MMB Volum-e, where it has demonstrated strong capabilities in identifying and categorising defects effectively.
When the AI feature launched, we identified the need for more robust documentation capabilities. This led us to introduce a new feature at Formnext 2024 that allows users to document analyses and share files seamlessly, further streamlining the workflow for customers.
The early feedback has been very encouraging, and we’re committed to continuous improvement based on real-world usage and customer input.
From this deployment, we realised the need for improved documentation and sharing capabilities. Based on this feedback, we’ve introduced a new feature for streamlined analysis documentation and file sharing, which was showcased at Formnext 2024.
Early feedback has been very positive, and we are committed to refining our solution through ongoing collaboration with our customers.