Digital twins, or highly precise digital copies of real things or systems, will allow the next phase of industrial virtualization and AI, according to NVIDIA and BMW officials.
NVIDIA’s Rev Lebaredian, vice president for Omniverse and simulation technologies, was joined by Michele Melchiorre, senior vice president for product system, technical planning, and tool shop at BMW Group, to kick off the ISC 2022 conference in Hamburg, Germany.
“There are many things you can do with this if you can build a virtual environment that equals the real world in terms of complexity, scale, and precision,” Lebaredian said.
While Lebaredian discussed the general themes and technological breakthroughs driving the creation of digital twin simulations, Melchiorre provided a comprehensive look at how BMW has used digital twins in its manufacturing facilities.
Melchiorre described real-time cooperation with digital twins and prospects for teaching AIs as a “revolution in industrial planning” as part of BMW’s aim to become more “lean, green, and digital.”
Melchiorre’s description of the BMW iFACTORY effort, which uses real-time data, simulation, and machine learning, shows how quickly digital twins have become workhorses for firms like Amazon Robotics, BMW, and others.
According to Lebaredian, these systems will connect our world representations with data coming in real-time from these realms.
“We’re looking for a system that connects the two right now,” Lebaredian said.”Any changes in the physical version can be detected and reflected in the digital domain.” “We’ll have some fantastic superpowers if we can link the dots.”
Every field of research is being transformed by supercomputing.
It’s also another illustration of how supercomputing technologies — notably its emphasis on simulation and data center-scale GPU computing — are influencing the rest of the world.
On the other hand, Converging technologies have changed high-performance computing, according to Lebaredian. GPU-accelerated systems are becoming commonplace in scientific computing and edge computing, data centers, and cloud computing.
GPU computing with AI acceleration has also become a staple of current high-performance computing. That’s how supercomputing is positioned to achieve the initial goal of computer graphics: simulation.
Computers, algorithms, and artificial intelligence (AI) have all advanced to the point that we can now simulate sophisticated worlds enough to be helpful on a large scale and even utilize these simulations as AI training grounds.
At a Crossroads in the World Simulation
A new kind of simulation is achievable with digital twins, according to Lebaredian.
Precision timing — the capacity to simulate many autonomous systems simultaneously — is required for this.
They need physically precise simulation.
They also need reliable data intake from the “actual twin” and continual synchronization.
We shall gain “superpowers” due to these digital twin simulations.
Teleportation was the first one Lebaradian looked into. “Any human anywhere on Earth may teleport into that virtual environment, just as in a multiplayer video game,” Lebaradian explained.
The next step is to travel across time.
“Time travel is possible if you record the status of the planet across time and can remember it at any moment,” Lebaradian explained.
“Not only can you now teleport to that planet,” he added, “but you can also obliterate your chronology and travel backward in time, as well as explore that destination at any moment.”
Finally, if these simulations are precise enough, we will be able to predict what will happen next.
“If you have an incredibly accurate simulator and truly predict what will happen in the future, you can practically time travel to the future if you understand the rules of physics well enough,” Lebaredian explained.
“You can compute not just one potential future, but many different futures,” he continued, describing how this may allow city planners to understand what can happen when they remodel a city, build a road, and adjust traffic systems to discover “the greatest possible future.”
These digital twins, which are tremendously compute-intensive and need precisely timed networking with extremely low latency, are being unlocked by modern supercomputing.
“We need a new supercomputer,” Lebaradian added, “one that can speed artificial intelligence and execute these large simulations in genuine real-time.”
To achieve precise timing, GPU-accelerated systems must be tuned at every system layer.
These systems will have to function in the data center and at the network’s edge to deliver data into virtual simulations with pinpoint accuracy.
Such systems will be critical for progress on both local and big dimensions, such as medicine development and climate simulation.
“We need to mimic our climate in a way that’s never been done before, with a precision that’s never been done before, and we need to trust that our simulations are truly predictive and correct,” Lebaradian added.
“Lean, Green, and Digital” is the slogan of BMW’s iFACTORY.
BMW’s Melchiorre gave an example of how this wide vision is being used in the automaker’s current efforts to become “lean, green, and digital.”
BMW has created intricate digital twins that simulate its manufacturing in real-time, with people and robots interacting in the same environment.
It’s a project that spans the production floor, the data center, and the company’s whole supply chain. Millions of moving components and pieces are involved in this digital counterpart, linked to a massive supply chain.
Melchiorre took his audience through various instances of how digital twins might be used to mimic how industrial machines, robots, and people would interact in the future.
He also highlighted how NVIDIA technology is used to mimic whole factories before they are built.
Melchiorre displayed an aerial photograph of the location in Hungary where BMW is constructing a new facility. The digital factory is 80 percent built, but the real-world facility is still mainly open.
Melchiorre noted, “This will be the first plant where we will have a complete digital twin long before manufacturing begins.”
