From The Editor

Are we now on the cusp of creating our own “Metalverse?”

It's been nearly a year since Facebook announced it was rebranding to Meta, changing the name of its parent company, stock market ticker, and logo, signifying it would now focus its future on the upcoming “metaverse.” In the time since, what this means hasn’t gotten any clearer and neither has the countless other references and uses of the term. The actual provenance of the word “metaverse” comes from a 1992 science fiction novel “Snow Crash” by Neal Stephenson set in the early 21st Century, years after a global economic collapse that was sparked in part by hyperinflation. In the book, he introduces the reader to the metaverse, a phrase he coined as a successor to the Internet and constitutes his early 1990s vision of how a virtual reality–based Internet might evolve in the near future. As with the internet evolving in the industrial sector, the recently held AFS 2022 Foundry Industry 4.0 Conference broached many of the topics that are key facets of metaverse and how our industry might integrate them into a more digital-centric manufacturing environment.

From digital data collection, distributed information sharing, adaptive machine controls, and artificial intelligence (AI)-based decision-making, are we now on the cusp of creating our own “Metalverse”? What is needed to implement this potential evolution or revolution in manufacturing? At the recent American Metalcasting Consortium (AMC) annual technology review, John Moreland from the Center for Innovation through Visualization and Simulation (CIVS) at Purdue University Northwest presented their research into creating a VR (Virtual Reality) simulator of the diecasting process. This work, funded by Defense Logistics Agency under the guidance of NADCA, has been ongoing for the past four years and is now at a point that via VR Googles, like Oculus Quest 2, and handheld controllers, one can enter an immersive experience on the floor of a diecasting shop. The simulator allows users to select various learning and practice activities going through diecasting machine orientation, start-up procedures and troubleshooting, melting furnace charging, tapping, and cleaning, as well as interactive visualizations that make parts of the diecasting machine transparent and allow the user to see liquid metal behavior inside the shot sleeve and casting cavity during a shot. While this was an impressive display of what is possible, how close is it to an actual metalverse?

During that AFS 4.0 conference, Tony Del Sesto from the National Manufacturing Institute MxD in Chicago discussed digital twins. This is a virtual representation of an object or real-world physical system (a physical twin) that serves as the indistinguishable digital counterpart of it for practical purposes, such as system simulation, integration, testing, monitoring, and maintenance. It spans its lifecycle, is updated from real-time data, and uses simulation, machine learning, and reasoning to help decision-making. While listening to the presentation by the researchers at Purdue one feels they are seeing what is called a digital twin. However as viewed in non-augmented reality we see that although helpful for presenting a virtual representation of the equipment or process to train and visualize, it is not an actual digital twin. Features like process settings and parameters (shot speed, hydraulic pressure, etc.) can’t be changed in real-time to observe and measure their influence on outcomes. What about a process simulation—is it a digital twin creating our “metalverse”? While also a key component of a digital twin, it is only one part—not the digital twin.

So, what does it take if we want to create one? It will take more than just collecting a lot of data, conducting simulations, and building models. It has become apparent that first, we need to move towards collecting inputs from more performance-based tests that are quantitative, conducted either in real-time and shared online or augmented via lab testing. This means moving beyond the traditional in-process tests, like the ones we do for our sand systems, and conduct those that allow us to integrate the results into process simulation modeling. The outcomes and predictions must be shared across the value chain. With the potential for hundreds of thousands of data inputs, we must discern the key signals that are predictive of the information and noise. This will allow for better adaptive control and real-time adjustments via knowledge-based algorithms, so the corrective actions are taken in the manufacturing environment eliminating the potential wastage/ scrap/non-conformance. Jiten Shah has been leading an AFS AMC project showing how ICME can be utilized to create these types of tools. The CAD representations for the various aspects of the manufacturing processes must be created. Then these components need to be integrated and linked.

So, is it even possible to create a digital twin of a metal casting operation or a metalverse? I have seen one impressive recent example of creating a digital twin for a thermodynamic system by Gruppoo Cimbali using the Altair EDEM to model and deeply study the physics behind making a cup of coffee. This and other manufacturing examples show how particle physics can be used. While it may take some time and effort to create one for an entire metal casting process, aspects certainly can be developed. This will require investment, collaboration among the supplier, user and academic community, training a workforce with new skill sets, and funding for research and development. The result will advance our technology to not only meet customer expectations but allow us to stay competitive against alternative future manufacturing approaches.