Materials teams want to find new alloys that outperform existing solutions or fill gaps in the market. They do a lot of expensive experiment, testing, and simulation to identify the best candidates. Can they get there faster at lower cost per new material? With Alchemite, you can use the data you have to design new and improved alloys, optimise processes, and more effectively target your testing and experiment.
Because this data comes from diverse sources and tests, it tends to be ‘sparse’, causing problems for conventional analysis tools. Alchemite is designed to handle sparse, noisy data, and to work for problems with multiple target parameters. It’s ideal (and proven) for design of new alloys/materials.
Case studies and demonstrators
Designing a new superalloy for aero engines
In a research project for the aerospace sector, the Intellegens technology met a complex design challenge, proposing a new nickel-base superalloy that was simultaneously optimised against eleven different physical criteria.
Combatting wear with Welding Alloys Group
Welding Alloys Group (WAG) and Intellegens applied machine learning to identify a new hardfacing material with considerable environmental and cost/benefit advantages when applied to combat wear, one of the most challenging problems faced by heavy industry.
Design of a titanium alloy with GKN Aerospace
Intellegens in collaboration with GKN Aerospace found a new titanium alloy composition for heat exchanger applications, seeking to maximise thermal conductivity without diminishing the current mechanical properties. The material design process that would normally take two years was reduced to less than three months.
Virtual experiments for steels
Chromium makes steel corrosion-resistant and hard, but is also toxic and costly. Replacing Chromium often relies on trial-and-error experimentation. Alchemite can run ‘virtual experiments’ to propose low-chromium steels with the right property profile.
Alchemite for alloys and superalloys
With the Alchemite software scientists and engineers can apply powerful deep learning methods to design, characterize, and optimise production of alloys.
- Gap-fill and validate sparse, noisy data from suppliers, testing, simulation, and production
- Auto-generate models that identify key property / process relationships
- Quantify uncertainty to support a rational business case for key decisions
- Identify which tests to do in order to characterise or qualify materials with maximum efficiency
- Propose new alloy designs, optimising against multiple targets
- Optimise both composition and processing parameters
- Respond to challenges such as how to incorporate recycled feedstock