Challenge: Additive Manufacturing (AM) is a new processing technology used in a wide range of industries to produce and repair bespoke and high-value parts including, for example, aerospace engine components, turbine blades, and oil drilling tools. To date, very few materials have been fully experimentally verified going through these processes, severely restricting the application of the technology to wider fields of use. A particular challenge is that the ability to print materials is poorly understood – direct laser deposition (one AM method) has only been applied to just ten sets of processing variables. This provides a mere ten data points which are not enough for traditional machine learning techniques to predict the properties of a wider family of processing variables.
Solution: We optimized the direct laser deposition process for AM using historical welding data and the available sparse direct laser deposition dataset. Alchemite was able to leverage abundant weldability data to automatically identify and exploit property-property relationships enabling the capture of new insights into the changing processing variables.
Outcome: By using historical welding data combined with the sparse data from direct laser deposition, we were able to optimize this AM process and broaden its application to new processing variables, saving 15 years of research. Quicker than conventional production techniques, AM has the potential to save manufacturers vast amounts of time and money. This approach is not limited to AM and can be applied to the introduction to any manufacturing process.
Swipe to navigate articles