Thursday, October 11, 2018

Digital Twin: A Bridge Between Virtual and Physical Processes

Digital twin technology might be some of the most exciting breakthrough technologies invented in the last few years, yet surprisingly few people have ever heard of the term. However, despite being overshadowed by buzzwords like “blockchain” and “artificial intelligence”, digital twin might be even more exciting.

NASA's twin Mars Cube One (MarCO) spacecraft.
NASA has some of the first use cases for digital twin, driven by the need to monitor and repair spacecraft
millions of miles away.

So, what exactly is a digital twin? It’s a virtual model of a physical process or product. In other words, it’s a connection between the virtual and physical worlds. Digital twins are made of three components, according to G.E., those being a data model, analytics, and knowledge. The data model layer is data collection from components on the physical machine or process; this is usually done through sensors. The analytics layer consists of predicting, describing, and prescribing the behavior of that asset; this is usually done through artificial intelligence and machine learning. Finally, in the knowledge layer, subject matter expertise from scientists, historical data, and industry best practices are implemented to create the best possible use of the physical asset. There are also several types of digital twins:

Credit to G.E. for this diagram.
For example, think of a machine in a factory which has to wrap candy bars. Imagine hooking every moving part of that machine up to a sensor, collecting that data and analyzing it in real time. You could get information about potential wear-and-tear and schedule maintenance, saving any time the machine could be broken down, find a faster way to wrap candy bars based on experimentation with the digital twin, and have a greater sense of reliability that the machine will perform well the next day, since your model predicted the next issue and steps have already been taken to fix it before the issue has occurred. Can you imagine how much the risk has been reduced, maintenance decreased, and productivity increased? That is the power of a digital twin.

Digital twins are part of the growing trend in the Internet of Things (IoT), a term used to describe the connectedness of computing devices and other hardware by means of data transmission over secure networks. The data layer of digital twins is usually a cloud-based system. This means that the sensors must have a secure connection to the internet, and that is where the immense amounts of data are stored and, furthermore, analyzed. This is all part of an increased need for real-time data collection and analysis, a sector where Hadoop technology operates: providing massive data storage and operating power on infinite concurrent tasks.

G.E. is one of the forerunners in digital twin technology. Many projects, such as their wind turbines, either currently have or are acquiring digital twins. In fact, GE predicts that they will have 93-99.49% increased reliability in less than 2 years, 40% reduced maintenance in less than one year, 75% reduced time, and an incredible 11 million dollars avoidance in lost production by detecting and preventing failures before they even occur. IBM is also leading the way, with a really cool project, Maximo, which uses augmented reality to let engineers actually “see” the digital twin and interact with it in the virtual world. IBM has said that they believe artificial intelligence ideas such as natural language processing, machine learning, and virtual recognition, can take us even further in using digital twins to design even more efficient systems. Other cool IBM projects include Rational Rhapsody and Rational Engineering Lifecycle manager.

Digital twin technology really is the next big thing. That’s why Gartner has had digital twin in its top ten technologies for 2017 and 2018. Things like visualizing products as they are used by customers for better product design, building digital threads, and engineers troubleshooting processes and hardware when they aren’t in the same country, forget the same room, could become a reality with digital twins. Imagine an engineer in San Francisco working on a jet engine in New York City, a wind turbine in Idaho being made more efficient by scientists in Washington, D.C., or a dairy farmer’s windmill in Wisconsin having its productivity analyzed in Seattle. Businesses could innovate and grow at unprecedented speeds, unhindered by things that today could ruin a day of production but tomorrow will be seen and fixed before they occur, and pushed on by making every part of every process as efficient as possible.

If you would like to read further about digital twins, give these links a visit:
- IBM Watson IOT: Introduction to Digital Twin
- Microsoft Azure: Digital Twin Services

1 comment:

  1. Organizations are presently utilizing AI to increase significant bits of knowledge already distant.machine learning course