Make Digital Twins Be Your Enterprise’s New Best Friend in an IoT World

As connected devices—anything from the computer built into a car or a fitness wearable to a smart gas meter or wind turbine—have become cheaper to manufacture, their use has grown rapidly. Some form of connected device or system is now in almost every industry and sector. This proliferation means that traditional maintenance models requiring engineers to have easy physical access to real-world devices and systems are often no longer viable. As such, the need for an alternative management model is growing, which is where “digital twin” AI technology has enormous potential for any enterprise doing business through connected devices. This report looks at the benefits companies can gain from digital twin technology, the options available on the market to take advantage of it, where digital twins are already in use, and what some industries stand to lose by ignoring this AI trend.

A digital twin is a software simulation of a real-world object or system created with AI technologies that include machine learning, deep learning, video and image analytics, and augmented/virtual reality (AR/VR). A digital twin can make remote sites accessible and observable, ensure compliance and productivity in remote locations (e.g., construction sites or deep-sea oil rigs), enable predictive maintenance of complex hardware, help design smart city programs, maintain commercial connected devices, and help track, model, and predict online consumer behavior (see Exhibit 1). This could be especially useful for manufacturing and utilities firms distributing their sites and operations ever-more widely across the globe, and consumer-facing firms catering to an ever-younger demographic of users who transact largely through mobile devices.

The IoT, or internet of things, isn’t just relevant to manufacturing firms. Consider things like phone sensors, Fitbits, smart speakers, and smart home hardware. By now, the IoT is an opportunity for businesses of all stripes to communicate more effectively with customers, predict their clients’ needs, and keep their operations running smoothly. To do this, however, maintaining a comprehensive, real-time overview of a vast network of IoT devices and systems is paramount. Enter digital twins.

Exhibit 1: Potential benefits of digital twin technology in industry

Source: Microsoft

Digital twins are becoming an imperative in an IoT-saturated world

Digital twins are virtual simulations of real-world assets. To enable generation of a digital twin, a real-world asset must have either an embedded connection to the cloud or have a sensor attached to it retroactively to enable such connectivity, converting it to a smart or IoT asset. These can be objects—like a wearable device, a smart TV, or a piece of heavy machinery—or a system—like a social media platform.Pairing and simulation technologies—precursors to digital twins—have been around for decades. NASA’s efforts to model how its equipment would function in space and then design it to eliminate preventable errors pioneered this technology. The modern term “digital twins” was coined in 2002 by Michael Grieves at the University of Michigan, but has only become an applied reality in the past few years due to the major technological advances and changing commercial circumstances outlined below.

• Emergence of sophisticated AI capabilities. Digital twins draw on multiple types of AI technology to generate realistic and accurate simulations of real-world objects and environments, including machine and deep learning, computer vision, and AR/VR. Just a few years ago, technology in these areas wasn’t sophisticated enough for commercial or industrial use, but greater availability of data and more computing power have changed this.
• More powerful computational models. More nuanced and powerful computational models have made it possible to build digital twins for non-physical systems, such as online customer behavior. Two models have been game changers: agent-based modeling, which simulates the actions and relationships of autonomous agents to determine their effects on a broader system; and system dynamics, which analyzes the nonlinear behavior of large and complex systems based on a wide array of inputs including time patterns and feedback loops.
• Proliferation of IoT devices. Connected devices have become cheaper to build and deploy compared to their early days, leading to broader commercial deployment and more physical objects visible to computers. Today’s connected devices encompass a wide variety of applications: wind turbines, deep-sea oil rigs, smart cars, smart airplane engines, Fitbits, and IoT fridges.

These factors have led to a boom in digital twin awareness and adoption and growing demand for affordable simulations. A growing number of vendors are now capitalizing on this demand. We’ll take a look at some in this report.

There is by now a wide array of digital twin offerings to suit your enterprise’s needs

The digital twin market is still nascent and in flux, with active vendors dividing along both industry and approach lines. Some vendors, particularly smaller players, take a more targeted approach and develop productized digital twins for specific verticals, while established older vendors tend to take a platform approach, enabling enterprises across a wide range of industries to compile their own simulations to reach a broader market. Here are several standout digital twin solution vendors from both small and large players:

• Microsoft’s Azure Digital Twins platform uses its Azure IoT platform to port the IoT’s platform scaling, compliance, and security benefits to customers using its digital twins solutions. Azure Digital Twins enables clients to model complex interactions between people, locations, and devices; features include a spatial intelligence graph, twin object models, and integration with broader Microsoft services like Office 365. The platform is designed to help customers transform any environment—from a school to a parking lot to a stadium—into a “smart” environment monitorable on a granular scale. This flexibility enables customers to leverage IoT devices for predictive maintenance, track energy consumption, optimize space usage, and make processes more efficient. Microsoft says its offering is differentiated by its model-first, rather than device-first, approach. In its own words: “We’ve found that customers realize huge benefits by first modeling the physical environment and then connecting (existing or new) devices to that model.” In other words, Microsoft aims to add value for customers by beginning with a strategic overview of their operational environments and plugging smart devices into it for optimal results rather than trying to change that environment piecemeal around available hardware.
• IBM offers a suite of digital twin solutions that an enterprise can use individually or combined, depending on its needs. IBM Rational Rhapsody is a modeling and design solution for product and system development and management. It enables engineers to create a secure development and test environment for connected systems and helps ensure compliance with strict industry standards. The Rational Engineering Lifecycle Manager gathers, analyzes, and visualizes engineering lifecycle data from across an organization to allow engineers to make more informed decisions. IBM Rational Lifecycle Integration Adapters, meanwhile, is an application lifecycle management tool that plugs into customers’ existing systems and hardware to help maintain device health and security. IBM’s deep AI expertise across a vast array of domains garnered through its Watson platform has enabled it to create a digital twin suite applicable to almost any sector by focusing on engineers as a horizontal, rather than on just a few industry verticals.
• US-based SWIM.AI, founded in 2015, launched its real-time edge analytics solution EDX in April 2018. The solution, which clients use in the smart cities, industrial automation, utilities, and IT infrastructure spaces, helps enterprises gain real-time data insights from edge-computing—distributed devices such as smart meters and machinery. EDX automatically generates digital twins in real-time from these data inputs. Moreover, EDX also enables the twins to analyze and learn from their own activity to generate accurate behavioral models for their devices using deep learning. SWIM.AI aims to give enterprises optimally accurate representations of their assets’ real-time and predicted performance, which they can then weave into ERP logic and workflows to enable the enterprise to make maximally data-driven business decisions (see Exhibit 2). SWIM.AI says its unique selling point is that all this is done in real-time, with no need for centralized or batched data analysis. The company raised $10 million in a series B round in July 2018 from major investors including Cambridge Innovation Plc. and ARM Holdings, as well as existing backers Silver Creek Ventures and others.

Exhibit 2: SWIM.AI’s digital twin framework

Source: SWIM.AI (HFS screenshot)

Veerum. Canadian IoT startup Veerum, founded in 2014, offers a Digital Twin platform for improving cost efficiency and project management for capital projects. The projects include large-scale undertakings such as oil and gas exploration or infrastructure construction. Unusually, Veerum largely deploys and leverages data inputs from its own IoT hardware rather than plugging into clients’ devices. Using hardware including drones, lasers, and ground robots, the platform creates a digital simulation of a site and updates it daily, enabling clients to track how their project is progressing against original design specifications (see Exhibit 3). The goal is to catch inefficiencies or errors early to avoid these problems escalating and becoming more expensive to resolve further down the line, as well as track materials use and other project factors on a more granular scale. Veerum is a participant in GE’s Zone Startups Calgary accelerator and has major clients including Cenovus Energy.

Exhibit 3: Veerum’s digital twin technology

Source: Veerum (YouTube screenshot)

Digital twins can give your enterprise a more holistic view of its operations

Historically, investing in such simulations only made fiscal sense to enterprises dealing with remote physical sites such as deep-sea oil rigs, spacecraft, and out-of-the-way construction projects. Their demand was driven by the need to be able to prevent system failures where human engineers wouldn’t be able to intervene easily, or at all. However, this is now changing. Some of the world’s most complex and mission-critical systems are no longer purely physical. For some businesses, being able to gauge consumer behavior on a social media platform could make the difference between a good and a bad financial quarter. Or, being able to predict a flaw in a smart fridge or thermostat could vastly improve customer satisfaction and brand loyalty in an economy where customer experience standards are being set very high by high-tech giants like Amazon.

Regardless of whether a digital twin relates to a physical asset or a digital one, the key benefits of the technology boil down to the following points:

• Lower spending on asset repair and maintenance. Being able to model asset behavior to predictively eliminate flaws and errors means both less expenditure on sending engineers out on-site and having to hire fewer engineering FTEs in the first place, lowering overhead.
• Predictive rather than responsive business decisions. Being able to holistically observe the behavior of a complex system and model its future behavior will enable enterprises to make strategic business decisions to anticipate future trends and events, giving them a competitive edge over less data-driven peers.
• Improved regulatory and industrial standard compliance. Remote monitoring and tracking of site conditions can help ensure workers wear appropriate safety gear, follow protocols, and use appropriate materials, potentially saving businesses fines in tightly-regulated industries such as heavy industry and chemicals.
• Better tracking of efficiency and productivity. Monitoring spatial usage and measuring actual output against projected output using a computer model, as opposed to on-the-ground oversight, can both greatly reduce expenditures and increase efficiency for enterprises.

Fundamentally, these benefits boil down to a holistic overview of their operations and systems, which becomes an invaluable advantage as the ecosystem of smart devices attached to an enterprise’s operations grows exponentially.

On the flip side of this coin, companies that neglect to view their expanding ecosystem of connected devices – whether these belong to the enterprise itself, or to its customers – as invaluable new sources of data collection and competitive insights, stand to fall behind in an increasingly data-driven market. As we’ve seen, there are plenty of varied options to help enterprises begin leveraging the information the IoT ecosystem provides, and little excuse not have them on their radars.

Bottom line: Digital twins will become essential to enterprises’ management of their assets—cross this bridge now or regret it later

The facts are clear and undeniable. The IoT is growing around enterprises and beyond their control, driven as much by their customers’ connected device purchases as it is by their decision to use these devices in their operations. Further, operations are becoming more globalized, increasing the distance between an enterprise’s nervous center and its peripheral assets. Both factors make the need for technology that can provide a holistic, real-time, and highly accurate overview of an enterprise’s entire asset ecosystem essential. This is where the leaps and bounds that have been made in the past decade in AI are making a tangible impact and making helpmeets like digital twins possible. Digital twins were once the turf of heavy industry and the exclusive right of very well-funded scientific organizations, but this is no longer the case. With offerings from players from IBM to SWIM.AI now on the market, enterprises have little excuse not to, at the very least, investigate this technology. What it comes down to is this: start looking to digital twins to control your IoT ecosystem, before it starts controlling you.