Artificial intelligence (AI) company Intellegens, which is a spin-off from the University of Cambridge, created a unique toolset that can train deep neural networks from noisy or sparse data. The machine learning algorithm, called Alchemite, was created at the university’s Cavendish Laboratory, and is now making it faster, easier, and less expensive to design new materials for 3D printing projects. The Alchemite engine is the company’s first commercial product, and was recently used by a research collaboration to design a new nickel-based alloy for direct laser deposition.
Researchers at the university’s Stone Group, along with several commercial partners, saved about $10 million and 15 years in research and development by using the Alchemite engine to analyze information about existing materials and find a new combustor alloy that could be used to 3D print jet engine components that satisfy the aerospace industry’s exacting performance targets.
“Worldwide there are millions of materials available commercially that are characterised by hundreds of different properties. Using traditional techniques to explore the information we know about these materials, to come up with new substances, substrates and systems, is a painstaking process that can take months if not years,” Gareth Conduit, the Chief Technology Officer at Intellegens, explained. “Learning the underlying correlations in existing materials data, to estimate missing properties, the Alchemite engine can quickly, efficiently and accurately propose new materials with target properties – speeding up the development process. The potential for this technology in the field of direct laser deposition and across the wider materials sector is huge – particularly in fields such as 3D printing, where new materials are needed to work with completely different production processes.”
Alchemite engine
Alchemite is based on deep learning algorithms which are able to see correlations between all available parameters in corrupt, fragmented, noisy, and unstructured datasets. The engine then unravels these data problems and creates accurate models that are able to find errors, optimize target properties, and predict missing values. Alchemite has been used in many applications, including drug discovery, patient analytics, predictive maintenance, and advanced materials.
Thin films of oxides deposited with atomic layer precision using pulsed laser deposition. [Image: Adam A. Læssøe]
“Worldwide there are millions of materials available commercially that are characterised by hundreds of different properties. Using traditional techniques to explore the information we know about these materials, to come up with new substances, substrates and systems, is a painstaking process that can take months if not years. Learning the underlying correlations in existing materials data, to estimate missing properties, the Alchemite
engine can quickly, efficiently and accurately propose new materials with target properties – speeding up the development process,” said Gareth, who is also a Royal Society University Research Fellow at the University of Cambridge. “The potential for this technology in the field of direct laser deposition and across the wider materials sector is huge – particularly in fields such as 3D printing, where new materials are needed to work with completely different production processes.”
Direct laser deposition – a form of DED – is used in many industries to repair and manufacture bespoke and high-value parts, such as turbine blades, oil drilling tools, and aerospace engine components, like the Stone Group is working on. As with most 3D printing methods, direct laser deposition can help component manufacturers save a lot of time and money, but next generation materials that can accommodate high stress gradients and temperature are needed to help bring the process to its full potential.
When it comes to developing new materials with more traditional methods of research, a lot of expensive and time-consuming trial and error can occur, and the process becomes even more difficult when it comes to designing new alloys for direct laser deposition. As of right now, this AM method has only been applied to about ten nickel-alloy compositions, which really limits how much data is available to use for future research. But Intellegens’ Alchemite engine helped the team get around this, and complete the material selection process more quickly.
(a) Secondary electron micrograph image for AlloyDLD. (b) Representative geometry of a sample combustor manufactured by direct laser deposition. [Image: Intelligens]
Because Alchemite can learn from data that’s only 0.05% complete, the researchers were able to confirm potential new alloy properties and predict with higher accuracy how they would function in the real world. Once they used the engine to find the best alloy, the team completed a series of experiments to confirm its physical properties, such as fatigue life, density, phase stability, creep resistance, oxidation, and resistance to thermal stresses. The results of these experiments showed that the new nickel-based alloy was much better suited for direct laser deposition 3D printing, and making jet engine components, than other commercially available alloys.
Discuss this story and other 3D printing topics at 3DPrintBoard.com or share your thoughts in the Facebook comments below.
198 Replies to “Alchemite Machine Learning Engine Used to Design New Alloy for Direct Laser Deposition 3D Printing”
Comments are closed.