Catalyzing the next industrial revolution
The metaverse is coming!
At least, that’s what many headlines are saying. You might have even heard of a company changing their name to Meta. The metaverse has been around since 1992 but really has gathered steam in the last ~5 years with the larger investment by tech companies in immersive environments and the adoption of consumer virtual reality. This is summarized well by McKinsey Senior Partner Lareina Yee:
“What’s exciting is that the metaverse, like the internet, is the next platform on which we can work, live, connect, and collaborate. It’s going to be an immersive virtual environment that connects different worlds and communities.”
While most people have a decent understanding of what the metaverse can do from a consumer perspective, but the question that comes up often is how this applies to industry. How does a car manufacturer transform their operations and create a digital replica of the world, complete with relevant data, 3D visualization, and AI, while allowing that information to be accessible by people regardless of geographic location or device. This is the industrial metaverse.
When describing what the “Industrial Metaverse” consists of, I look at four key areas as being critical for success: a data platform, virtualization, environment, visualization, and data science. When these systems communicate with each other and can be used by people to collaborate, they are able to transform how businesses operate and catalyze the fourth industrial revolution.
Data Platform
In terms of building a foundation for an industrial metaverse, a rich data platform is imperative for success. Without telemetry, environmental, and operational data then there can be little insight gained from a digital twin or visualization, and certainly there can be little to no data science applicable without data to analyze. For the facility of the future, this will consist of environmental sensors that will share things like temperature and humidity, telemetry data for robots and machines that share exact positioning, voltage/current, and pressure, and information regarding the inputs and outputs (both goods and waste) of the system. By itself, the information from this can be used for a variety of things such as environmental operational tuning, power optimization, and even waste reduction. In the industrial metaverse, this information feeds the other components that encapsulate the solution. Data can be viewed in context, with 3D visuals by operators to improve operations. It can then be analyzed to identify multivariate anomalies that reduce productivity and profitability. Further, this information can also be used to help a company better understand the true environmental impact of whatever it is they produce. Not just the carbon impact from the final product, but that of their machinery, their people, and all aspects of the value chain. This information also provides critical inputs for simulation…but more on that later.
Virtualization
A digital twin (or omniverse depending on which product you prefer) of a facility is a digital representation of real-world things. This can be buildings, places, people, processes, connections, etc. It’s a bit like the Matrix but more of a copy and less of a façade.
This is the platform that connects the components of a facility and allows users to view their data in context. In built correctly, these digital representations of the real world allow users to better understand how things interact with a live graph representation and the right APIs tying everything together. Like other aspects of the industrial metaverse, this may seem trivial by itself. Yes, things are connected. So what? When that’s paired with telemetry information and IoT data it allows for the digital twin to echo the real world and consistently refactor itself to ensure that it is always current, up to date, and showing how things are operating. No more dated engineering diagrams that you hope are up to date. No more red-lined documents to review for the latest updates. This allows for those things to happen in real time. The critical component of this is a 3D visualization engine for this data.
Visualization
We have data and we have a digital representation of the facility, but what is really making these things powerful is visualization. Much of what people think of when it comes to the metaverse is virtual reality, which is absolutely a component, but that is much too narrow of a perspective for the metaverse. The data platform and digital twin are truly revolutionary when they have a 3D representation attached to them that is independent of device used to access. Mobile phone? Yep. Laptop? Of course. Virtual reality headset? Duh. Mixed reality headset? YES! These all need to be viewed as different ways to view the same 3D model and not a single solution that will solve all ills of the world. This is because users will be consuming this information in different ways, in different locations, for different purposes. They all need to have a single source of truth for what they’re looking at when they’re looking at it. This cross-platform design allows for solutions to scale across companies, geographies, and industries to truly bring about Industry 4.0. It’s simple for me to say that this is all needed and would be easy to ignore the elephant in the room — scale. Who is going to build it? High fidelity 3D models aren’t simple to create, especially to be consumed by multiple platforms, and many of the solutions to date are not equipped to render the massive number of polygons that are included in large scale CAD models. These hardware limitations will obviously be reduced in the future (Moore’s law implies this will happen in the future) and there will likely be better ways to reduce polygons in a more dynamic fashion. Heck, there could be options available today that I’m not aware of.
Data Science
With the ever-growing sources of data (Number of connected IoT devices growing 18% to 14.4 billion globally), there is also a growing need to better understand that data, in context, to drive business insights. Many people believe that more data is better. How could they not? More is always better unless it’s more traffic, more cholesterol, or more pineapple on pizza, except when it’s not. Ideally, more data isn’t better, better data is better. Relevant information, presented in context, can be extremely powerful. Data outside of this is just noise. While I advocated building a data platform previously, to collect terabytes of data, the proper use of that is what will drive the right business insights to do things such as reduce waste, improve output, increase margin, or reduce failures. The ability to turn the data collected from large arrays of sensors, telemetry, etc. can be aggregated to uncover relationships in a production line that would previously have gone unnoticed. Further, this builds a platform for AI and reinforcement learning to dynamically optimize operations in a more agile way and pulling in other factors that would have previously been unconnected. The last thing that this enables is simulation. The ability to simulate operations, scenarios, configuration changes, SKU changes, etc. in a virtual environment is truly what unlocks the potential of Industry 4.0 and allows for companies to differentiate themselves from others who cannot simulate their operations. The ability to simulate production processes, physics, and permutations of faults that can happen from start to finish and allow a company to proactively resolve improves productivity, reduces waste, and allows for businesses to extract much of the lost value in their operations. This allows for more agile operations that can optimize for cost, margin, waste, carbon impact, or a variety of other factors that are important to the business, all of which empower businesses to create more value.
While many think of immersive virtual spaces as the “metaverse”, I think a more appropriate way to view it is as a dynamic digital representation of the real world. This obviously includes virtual spaces, but also IoT information, simulation, robotics, API connections, 3D visualization, data science (not limited to AI/ML), and the integration of of these disparate systems into one single solution. Companies that use the industrial metaverse have thousands of configurations, partners, scenarios, and products to account for so the industrial metaverse won’t be “click to deploy” ready for years and will consist almost entirely of bespoke solutions in the short term. The exciting part of this isn’t creating custom solutions. The exciting part is that the components that make up the industrial metaverse, the digital twins, reinforcement learning, 3D visualization, mixed reality, etc. have already been created. They’re ready to be used.
The industrial metaverse is here.
Quelle:
Foto: Photo by Lenny Kuhne on Unsplash
https://arvrjourney.com/the-industrial-metaverse-5402de9fba4b