Multiscale modelling of additive manufacturing
CSIRO Manufacturing has partnered with Singapore’s Nanyang Technological University and RMIT University to develop cutting-edge simulation models of additive manufacturing. The robust predictive modelling, designed specifically for the Selective Laser Melting (SLM) process in the first instance, will help dial-up pre-determined microstructures in additively manufactured parts, leading to more efficient and cost effective production.
The customisation opportunity offered by additive manufacturing means that each bespoke part built using the technology is different from the one printed before it. But this value proposition can present a challenge when determining the optimal processing window for each part. Knowing the best set of parameters from the outset can eliminate ‘trial-and-error’ in production, and allow high value parts to be built right the first time. The subsequent reduction in scrap can lead to substantial cost savings, with additional savings gained through assuring part quality and removing the need for destructive testing.
It is easy to see how a reliable predictive model that supports the engineering of pre-ordained microstructures in additively manufactured metallic parts can provide businesses with a competitive edge.
Tuneable, defect-free microstructures in metallic additive manufacturing – no longer a dream
A material’s properties are determined by microstructures, and reliable predictive modelling can help optimise the ‘tuning’ of those microstructures, allowing for specific property distribution in critical sections of a part during processing.
Multiscale modelling, which considers phenomena that occur at various time and length scales, offers a strong opportunity for developing this sort of predictive capability. Such modelling has been traditionally difficult to achieve given the different software packages used for the various scales are typically incompatible. To overcome this barrier, CSIRO has pioneered the use of the same programming language for proprietary coding for the different scales.
This project will combine high-definition continuum fluid dynamics at the part scale, which models powder melting and melt pool tracking, with phase-field modelling at the grain scale, which predicts solidification rate and finished-part microstructure. ‘Live’, bidirectional data exchange between these two sub-models will permit the most accurate predictions of both the behaviour of the SLM process and the microstructure of the finished part. Both will be obtained from one multiscale simulation platform.