Digital Twin: Starting the journey

Editor's Notes

1. Digital Twin is a means to an end, not an end in itself, so journey is an apt title for our talk. A digital twin is all about enabling new business outcomes. Over the next few slides we will explain what new business results are possible, and how these are enabled by Digital Twin technology.
2. As you’ll see throughout the presentation, you’re not simply installing a Digital Twin without first having the right foundations underneath. To build the right Digital Twin requires review of your current platforms, security, cloud providers, and business operations as most were not built to scale to the new velocity/volume of digital data. Beyond just technology, you must consider various vendor approaches to IoT and Digital Twin to ensure the offerings are created to take EVERY step of your journey, such as Blockchain, but those services scale up/down as you require.
3. Across all industries, executives are responding to a flood of raw data. Beyond traditional systems of record or systems of engagement comes a new volume, variety and velocity of diagnostic data. Machines, personnel, and buildings (i.e. assets) are sending off signals from various modalities (sight, sound, touch) at rates of a thousand times per day, per hour, or per second. This data offers engineers insights, in real-time, of assets in-use or in various stages of its lifecycle.
4. Big Data and IoT enable companies to move to new business models, such as the Rolls Royce example of selling their product as a service. By incorporating real-time data gathering from their products and applying continuous analysis of the data companies. New modes of interacting with data enables workers to become more efficient across many industries. These new user interaction modes include augmented reality, conversation or speech interaction
5. Collecting data on assets themselves is a recent shift in manufacturing. Historically, engineers built models and ran simulations to predict how assets might respond over use, temp, time, etc. Traditionally, a complex braking system in a car might be simulated, built as prototype(s), and put into production. Once in-production, it was difficult to determine the performance of the braking system in various uses, locations, times. Seeing this usage means IP can rethink what features are most valued, was experiences are missing, and what outcomes clients want more of.
6. It by incorporating digital data into stages of design, build, and operate that a firm can bring greater insights to each role and create nearly-custom products, services, and experiences for clients. By truly knowing what clients use and want from your products you can change the level of engagement, adjust supply chain complexity and risk, and change the nature of machines being ‘sold’ into outcomes or ‘as-a-service’ offers
7. Let’s break down this definition to understand its true meaning. A virtual representation means a digital model. There are actually many forms of digital model possible, and the digital twin is the integration of these various digital models to provide views to various users. Depending on the use case and the user, the digital twin may present different views. From here on out, we’re going to simplify the term “object or system” and just call it a “thing”. The thing could be a singular object, like a product, or even a person, or it could be a more complex system, such as a building, a factory, or an energy distribution network, comprised of many different parts. The lifecycle means that we have digital models of a physical product when it is just being designed, when it is being manufactured or produced, and as it gets deployed and “lives” in the real world. The methods of designing and producing the product or system will vary depending on what that product is, and so the models that make up the digital twin will be different for different kinds of products and systems. Real-time data is data collected from the sensors that are either built in or attached to a physical object, or in the environment surrounding the object. The idea of “real-time” suggests that data is collected over time and captured with high frequency. How frequent the data needs to be captured varies a lot across different applications. In some cases to be able to gain an accurate understanding of how a system is performing we may need to capture data every second, or even more frequently. In other cases, it may be good enough to capture data only every few minutes or even just a few times a day. The words “real-time data” also suggest how frequently the data needs to be captured, it doesn’t say that the data MUST be analyzed in real-time. Again, this will depend on the application and the use case. In some cases, decision making must happen nearly instantaneously (think of a safety critical situation), and in other cases such as predicting when maintenance needs to take place, a decision within a day is usually adequate. The words “or other data sources” mean that in addition to the sensor data we just talked about, there are many other data inputs that might go in to the views for a digital twin, depending on the use case. Data about the original design of the thing, about its manufacturing, about its maintenance records, and even textual data such as technician’s notes, customer feedback, and social data such as twitter comments, can all be useful pieces of the digital twin. What we want to enable with the digital twin is that the “computer” (in this case, typically a cloud service) provides the user with useful insights and even proactive recommendations to help the user work with the thing. The digital twin is a computer-based proxy with more intelligence than the actual thing has. The digital twin should be able to understand important things about its data, such as important relationships, trends over time, and potential impacts of decisions. The digital twin should be able to learn from the data, extracting patterns and relationships to provide new information without programming. Finally the digital twin should be able to use the learning and understanding to help the user reason about the state of the thing and about decisions that need to be made in that context. Dynamically recalibrating means adjusting operating parameters, issuing digital commands or executing other changes to change an expected outcome. This is something that a good digital system should do based on available information to optimize the business outcome according to the instructions and guidance given to it by the users. Improving decision making by the organization and its many users and stakeholders is the ultimate goal of the digital twin.
8. Digital Twins should be considered critical building blocks for an organization’s digital strategy. Capturing measurable process metrics and finding ways to instrument and measure these metrics for improving the business is an important best practice on the path to implementing a digital transformation strategy. The Digital Twin is instrumental in translating the performance of things in to the digital metrics that a business can use to improve decision making.
9. A Digital Twin helps serve users with key performance information. Understanding user needs and how better information about things can improve their decision making is the first step in building a digital twin. There is a vast amount of data that *could* be included in the Digital Twin. What should be included is just the data and visualizations most important to those users.
10. IBM sees Digital Twin far beyond just a virtual image of a physical object. Putting on VR/AR googles it what most Digital Twin vendors will demonstrate but they ignore the valuable work and assets that lie underneath. Visuals loose their appeal if they’re not fully describing and predicting asset performance. To offer an analogy, no movie can be created without a script, Director, Actors, and Producers; similarly a Digital Twin needs a foundation to provide the best views inside an asset.
11. Ultimately the digital twin is a data integration problem. The IBM Digital Twin approach includes organizing relevant information in to a knowledge graph. IBM Watson Services, with cognitive capabilities, can then analyze the graph to discover and present important insights hidden in the data.
12. At the heart of the Digital Twin there is a mechanism for creating a digital abstraction of the Thing. In the IBM Watson IoT Platform, we create a device abstraction that allows us to model not just the properties and data coming from the device but also its relationships to other devices and Things.
13. The IBM Watson IoT Platform includes a number of functions that allow us to bring the data from the device itself and incorporate that data in to the digital twin abstraction. Sometimes device data comes in different formats or needs to be modified in some way to fit in to the information model we want for running analytics, and the Watson IoT Platform Information Management capabilities allow the device data to be easily modified for integration in to the digital twin.
14. Digital Twins should be considered critical building blocks for an organization’s digital strategy. Capturing measurable process metrics and finding ways to instrument and measure these metrics for improving the business is an important best practice on the path to implementing a digital transformation strategy. The Digital Twin is instrumental in translating the performance of things in to the digital metrics that a business can use to improve decision making.
15. We believe in creating a platform for business that brings together not just the best infrastructure, but the best platform, the best software solutions as a service, business process as a service, and the expertise to apply it to your business when needed to help you succeed. These are just a handful of the areas where IBM cloud leads today: market share leadership, technology leadership, industry recognition, and thought leadership. Sure, IBM is a leader in public cloud IaaS, according to IDC, TBR & IHS. But that’s where it starts, not ends. We lead in hybrid cloud according to: Forrester, TBR, Ovum, and Synergy We were cloud company of the year in 2016 according to Frost and Sullivan. We lead in Object storage, IoT, Business Intelligence, API management, Mobility management, managed private cloud, managed security, dev-ops, application testing, application lifecycle management, agile project management, as well as in overall cloud thought leadership. We’re building a platform for the future of business. Your business. Your unique requirements. Your cloud.