Digital twin is the ability to take a virtual representation of the elements and the dynamics of how an Internet-of-Things device operates and works. It's more than a blueprint. It's more than a schematic. It's not just a picture. It's a lot more than a pair of glasses. It's a virtual representation of both the elements and the dynamics of how an Internet-of-Things device operates and lives throughout its life cycle. It's an understanding of all of its dynamics, whether those are electrons that move, or whether it's the device that's moving itself. It's about understanding the elements that compose it and the dynamics of how that device is put together. Done correctly, a digital twin will influence how design, built, and operations of a device are constructed in a single life cycle.
The design phase is where engineering tooling comes together, bringing together physical elements, physical bill of materials, pulling together virtual elements. You heard about the software on the cars and all the different elements and the chips that are there. Being able to coordinate and collaborate those together into a single facility of operational oriented design that is designed to bring out the highest quality product.
In the build phase, it's about yield. It's about understanding how the devices that make the product influence the product's tolerances and stresses and designs. And it's about better manufacturing, to drive the correct tolerances and correct outcomes that you want to see for the product that you're actually making.
And then third, a digital twin facilitates the actual operation of the product as well. Products age, products go through different environments. They deal with things like weather, and they have different tolerances, and they shift, they drift. And so, your digital twin needs to drift along with those products as they age. And that feedback when done correctly, not only facilitates the operations of the product, but helps facilitate better design and better manufacturing by the lessons that are learned and the re-calibration that takes place along the way.
So quite simply, a digital twin, virtual representation of the elements and the dynamics of an Internet-of-Things device. It affects both the design, the build and the operations, of how products pulled together.
There's essential capabilities that must be present for you to be active in a digital twin. First, you have to apply analytics at every step. The amount of information that we're dealing with to apply digital twin to a small device or to a complex device, such as an automobile or an aircraft, is staggering. Analytics has to be both real time. It has to be operational. It has to be quality. And it has to be predictive oriented in its nature.
The data that comes from a digital twin needs to be open. You have to be able to access it from a variety of different sources. You have to be able to pull it together into a federated model. And you have to be able to bring it together, so that you can get that interaction, that dynamics to play. It's not just a schematic you're making. It's not just a picture you're making. You're actually making a dynamic model that you're going to shift, as you go through both the design, the build, and the operations phase of what you do with the life cycle of that product
And then last, you're going to apply industry context. You may actually use the same product differently in two different industries and have two different digital twins for that one product, based on how the industry uses that product. Whether it's a pump that's used at oil and gas, or whether it's a pump that's used in municipal or wellness and water, the outcome is based on the industry context of how that device is going to be used. And so, a digital twin not only captures the engineering aspects, but it also captures the industry context, the dynamics of how that product is used at the same time.
Let's talk a little bit about how we do this at IBM. At IBM, we start with the Watson IoT platform. It is where we bring the data in. We connect to a variety of different data sources, direct devices themselves. We connect to engineering capabilities. We bring data in through partners, such as , from the physical side. We bring logical data in from our own portfolio. And around the IoT platform, we start to build context of the relationships of the information, both over time and relationship to each other, collapsed and collect the information.
Second, we apply cognitive insights. We operate on the data to understand its variances and tolerances and how it's used. We apply techniques in machine to machine, natural language processing, video, acoustical analytics, and more to help understand the dynamics of the information that you're being presented. So when you have the information and you have cognition taking a place on it, what you can then do, which is the third step, which is the most valuable step, is you can dynamically re-calibrate your environment.
So the digital twin, when operating correctly, not only represents a picture. It's not only something that you can see in glasses and explode, but it actually works to dynamically re-calibrate your environment, affecting the design, the build, and the operations phases of everything you do around that particular device.
Well, where can you go see one of these? It's really simple. The first place that you go see one is go visit our own Watson IoT center. We've instrumented several of the floors. We collect information. And we apply our digital twin models across both how we can show you dynamic recalibration around comfort. We can show it around efficiency of how the workspace is laid out and the interactions of the different types of office capabilities that are there. And last, we can show you the economic impact or the environmental impact of how the representation takes place at the same time.