ADVANCED MATERIALS & PROCESSES | SEPTEMBER 2026 23 A digital twin is a system that copies a real object—a product, a piece of equipment, a process— into a digital space. This digital copy is updated with data from the real world and uses the results of the analysis to improve the real object. Closing this loop—visualizing behavior in a digital space, feeding the results back into the real world for testing, and sending the verified data back to the digital side again— can move research and development forward a great deal, in either field. The closed loop is the gate through which digital twins will open a new frontier for materials science. This article introduces one such attempt at digital-side visualization, built on a particular software platform. As described in the box below, the DIGITAL TWIN AND FLOW SIMULATION FOR IMPROVED MATERIALS PROCESSING Beyond simulation, the real power of a digital twin is the closed feedback loop, which provides valuable digital-side visualization, enabling accelerated materials development. Hideyuki Kanematsu, FASM,* The University of Osaka, Suita, Japan, and BEL Inc., Sakai, Osaka, Japan Jeremy Knopp, PacForce K.K., Tokyo Akiko Ogawa, National Institute of Technology, Suzuka College, Suzuka, Japan Takayoshi Nakano, The University of Osaka, Suita, Japan *Member of ASM International authors’ research interests are centered around additively manufactured metal powders and bacteria-based biofilms. In both fields, the usual approach is trial and error, repeating the experiment and adjusting it in search of a good solution. This is slow, it costs money, and it is reaching its limits. A DIFFERENT KIND OF SOLVER: PARTICLES, NOT MESHES Most computer-aided engineering (CAE) tools for fluid and particle flow are built on a mesh: The domain is divided into a fixed grid of elements or cells, and the equations are solved on that grid. Finite element and finite volume methods have done this job well for decades. But a fixed mesh has trouble when the shape of the domain itself keeps changing—a free surface sloshing, particles separating and coming back together, a powder bed collapsing under a blade. Each time the free surface or the particle arrangement moves far enough, the mesh has to be rebuilt to keep up. This costs computing time and can also be a source of numerical error. This study utilizes two platforms from Prometech Software—Particleworks and Granuleworks. Each takes a different approach. Both are particle methods: The fluid or the granular material is represented directly as a set of particles that interact with each other, not as values fixed to a grid. Thus, there is no mesh to distort or rebuild as the powder moves. Particleworks was developed first as a general fluid solver for problems such as gearbox lubrication. Granuleworks extends the AUTHORS’ RESEARCH FOCUS Two of the article’s authors, Dr. Jeremy Knopp and Dr. Takayoshi Nakano, work on laser-based additive manufacturing of metal powders. The other two authors, Dr. Hideyuki Kanematsu, FASM, and Akiko Ogawa, work on accelerated-test devices for the laboratory evaluation of biofilms formed by bacteria on material surfaces. In 2025, they presented the biofilm case in part at the International Materials, Applications, & Technologies (IMAT) conference in Detroit, and the powder case for additive manufacturing (AM) technology at the Materials Science & Technology (MS&T) 2025 event. As the latter meeting was held in Ohio, the authors also delivered an invited lecture at Wright State University in Dayton. They hope to report on the physical-space results at an IMAT conference in the near future.
RkJQdWJsaXNoZXIy MTYyMzk3NQ==