Many people think of a digital twin as a 3D rendering of a physical object. And while that might be part of it, we’ve had computer-aided design (CAD) models for decades. Why the hype now?
It’s because the digital twin concept involves far more than that.
Digital twin technology helps businesses visualize assets and optimize operations by synchronizing the virtual world with the real world. Internet of things (IoT) sensors instantly transmit assorted data from an object to its digital twin. As the conditions of the object change, so too do those of its digital twin.
A digital twin is not simply a 3D rendering; it is a dynamic digital representation of a real-world object in real time.
Digital twins can be made for anything, from a single part to an entire power plant. They can also be created to visualize a process, such as a complex business project. Over time, data accumulates and a profile of the object or process emerges, enabling clear understanding of the past, acute awareness of the present, and improved decision making about the future.
In some cases, a digital twin can control its real-world object using actuators to convert human inputs into physical actions (e.g., an operations manager remotely suspends a machine after its digital twin indicates excessive heat).
While digital twin technology is relatively new and used primarily by large enterprises, it will eventually change the way information is shared and how asset lifecycles are managed. Small and midsize businesses must learn about this emerging technology now to prepare for the disruptions and opportunities it will bring in the future.
The seemingly sudden rise of the digital twin has been driven by several factors: ubiquitous cloud-based computing, effectively unlimited data storage, advanced analytics capabilities, and the sharply declining cost of internet of things (IoT) sensors.
Think of a digital twin as the central online data repository for one specific real-world thing. The digital twin isn’t the model-nor is it the sensors, data, or analytics. The digital twin is the rich and versatile resource that results from their convergence.
The model is a virtual mirror of a real-world object. It can also represent a system (e.g., supply chain) to make it easier to visualize and thus more tangible. Advanced modeling software can produce strikingly realistic digital versions of real-world objects that allow omnidirectional views and various methods of manipulation. However, the model does not necessarily need to be an exact duplicate, depending on its purpose.
To make a digital twin, you must connect it to a real-world object. This is done by attaching IoT sensors and other instruments that extract information about conditions, status, and context. In recent years, IoT sensors have quickly decreased in cost while greatly increasing in variety and utility. IoT sensors make assets smart by instantly and wirelessly transmitting measurements-such as motion or chemical composition.
Real-time data syncs the real-world object with its twin. For example, the digital twin for a specific section of an underground pipeline would include data regarding conditions (e.g. pressure, vibration), operational status (e.g., online, offline), and context (e.g., angle, location). Digital twins also commonly integrate other relevant business data such as product details, technical specifications, and warranty status.
Software runs the model while monitoring and organizing the vast amounts of incoming data. Analytics programs crunch data to glean useful information and make predictions. Digital twin software is also used to run off-line simulations and test the results of different variables in what-if scenarios.
One of the primary benefits of digital twin technology is its data sharing capability. Digital twins collect massive amounts of data that is often relevant to a diverse array of stakeholders. In the following example, we consider how various stakeholders benefit from the data generated by a single automobile’s digital twin.
The owner accesses an online portal that displays information about the car's status, specifications, and warranty.
The insurance company gauges the owner's driving habits to inform a usage-based rate policy.
Mechanics access information to identify problems, view service history, and make repairs.
Regulators assess data to verify compliance with safety and emissions standards.
The manufacturer and suppliers analyze incoming data to enhance design, hone engineering, and improve safety. Composites of this car and every other car like it are used to determine average wear of parts, run simulations, and predict failures before they occur.
Though many people associate digital twin technology strictly with manufacturing, it is incredibly versatile and can enhance business in many ways.
Digital twins make a huge impact on field service. Real-time feedback can alert users not only that a problem has occurred, but can relay the exact nature of the problem. For example, in the field service industry, it’s common to arrive at a service location with only a vague notion of the problem at hand. Statistics show that nearly a quarter of all field-service requests are not resolved on the first attempt, requiring a follow-up to complete the job.
Digital twin technology can alert the technician to the specific problem, allow visual inspection prior to arrival, and provide access to historical data that might have led to the issue. This ensures that the technician arrives with the tools, knowledge, and context needed to fix the problem on the first try. This is particularly useful for sites that are not readily available for inspection due to complexity, distance, or danger.
Digital twins illuminate the entire supply chain, from procurement to delivery. As a supply chain is a process and lacks a distinct physical form, its digital twin is metaphorical in nature but still reflects the current status of each stage of the operation.
For example, a regional paper company might use a digital twin to monitor production, gauge customer demand, and track inventory. Meanwhile, the digital twin automatically orders raw materials from suppliers, monitors warehouse humidity levels, and predicts the potential for shipping delays due to inclement weather.
CAD modeling can show you a realistic digital blueprint of your office. A digital twin of that office, however, indicates how many people are in each room, which resources are being used, and whether a light bulb needs replacing. Sensor-laden buildings are now being developed with digital twins in mind from the ground up.
These systems collect and disperse information in innovative ways that are fundamentally changing building management. Air quality sensors alert facilities to change filters, signs in the parking garage direct drivers to open spots, and various seating arrangements are simulated to determine which configuration is most efficient.
Beyond complicated design and high cost, digital twin technology faces several hurdles before it can be adopted widely.
As industry standards are not yet well defined, achieving interoperability between manufacturers is a significant challenge for digital twin development. A discrete digital twin represents a single part made by one manufacturer and is relatively simple to create. However, complications arise when creating a composite digital twin that integrates numerous parts from different manufacturers.
Another challenge is coordinating and securing the vast amount of data required to run a digital twin. As discussed previously, digital twins typically have multiple stakeholders who should each have access only to specific data. This presents risks to proprietary information, data privacy, and regulatory compliance. Great care must be taken to ensure that the right information ends up in the right hands.
Getting involved with digital twin technology at this early stage comes with risk. The technology is evolving rapidly an, as the industry matures, some digital twin software vendors will end up more successful than others. Signing up with the wrong vendor now could result in a project that is obsolete before it’s finished.
Furthermore, proprietary software and a lack of standards means that data portability is difficult. Thus, early adopters might become overly dependent on a particular vendor or find that changing vendors is impractical, a problem known as vendor lock-in.
It’s clear that digital twin technology has only just begun to show its potential. Gartner predicts that by 2021, half of large industrial companies will use digital twins, resulting in those organizations gaining a 10 percent improvement in effectiveness (report available to clients).
Digital twins will increasingly incorporate machine learning, AI, and augmented reality to automate processes, invoke self-repair mechanisms, and enhance training simulations. Farther in the future we will create digital twins of unimaginably complicated systems, from our bodies to entire cities.
For now, the needs of most small and midsize businesses can be met with less involved business intelligence and analytics software. However, as best practices are developed and use cases are proven, the technology will become more affordable, spread rapidly, and be adapted to the needs of SMBs.