Digital twins are a complicated topic. Essentially, a digital twin’s purpose is to replicate a complex, multi-tiered manufacturing process in a digital or simulated environment to iterate on design or process parameters. In additive manufacturing (AM), the concept of a digital twin is often no more than an abstract idea. However, the advent of a true AM digital twin will be a springboard for standardization, allowing companies to test the feasibility of design constraints for applications with ultra-fine margins of error, such as those in the aerospace and defense industry.
A&D engineers have many tools available to simulate AM builds, from topology CAD models to structural part-scale models, but the devil is in the details. Capturing the intricate physics and understanding outcomes is imperative for creating a truly representative digital twin. For example, in laser-powder bed-fusion, what happens when the laser hits the powder? There’s heating, melting, vaporization and solidification, all of which have an effect on materials used for flight-critical components.
That is somewhat straightforward, at least conceptually. But what does it mean in terms of microstructure evolution, thermal stress evolution, part distortion, cracking or surface roughness? What does that mean for part integrity and mechanical properties? How does the rate at which the metal powder melts and resolidifies affect these outcomes? Can we model physics to predict these results, and, if so, how? Answering each of these questions is high on the list of any A&D project that aims to leverage the power of the digital twin.
When working with any fluid, computational fluid dynamics (CFD) is usually the numerical method of choice. In AM, CFD comes into play where raw feedstock (powder) is melted down and processed into a completed part by modeling the transition from solid to molten to solid.
During this transition, many forces are at play. There’s laser heating, which will vary depending on heat flux distribution, speed and power; material vaporization, which depends on phase transitions specific to each alloy; recoil pressures, which depend on vaporization; Marangoni effects, which depend on surface tension and temperature; and dynamic pressure forces, which stem from gas flow and vapor plume dynamics. Clearly, it’s much more complex than fluid flow.
In the CFD software FLOW-3D AM, the physics necessary to capture these details have been implemented to make it easy for adaptation in scientific research on material behavior for AM. Researchers in A&D and other industries commonly use CFD modeling to make more accurate predictions on molten behavior leading to porosity formation or interlayer fusion, but what really makes this approach so powerful is the ability to extract outputs that can feed into other models in the digital twin framework. These include temperature gradients and cooling rates for microstructure models, or temperature and pressure data for finite element methods for modeling thermal stress, or even solidification velocities for hot cracking predictions.
To develop a digital twin for AM that accounts for all its many details and dynamics—from choosing a raw material to building a finished part—there will undoubtedly need to be a mechanism for delivering data seamlessly between various modeling tools. This is no small feat, and it is recognized in the AM industry as a necessary feature to allow for more rapid development and optimization of AM processes.
At the moment, most tools are manually connected via custom-developed application programming interfaces (APIs), but in the race toward bridging these tools and creating an end-to-end digital twin, we will likely see more universal interfaces and file formats, broadening access to this technology.
So, while CFD software continues to improve and deliver robust solutions for exceedingly challenging problems in AM, we must remember that these solutions will also bring value to a collective network of engineering and design tools and help to drive AM from a (relatively) new and untested manufacturing process to a reliable solution for some of the greatest challenges in manufacturing.
Connect With Us