IDC estimates that European manufacturers had spent nearly €4.2 billion on digital twins by the end of 2019, up more than 15 per cent year over year. Utilities, transportation firms and mining operations were estimated to have added another €1.2 billion to the mix, 20 per cent above 2018. Most of this spending was for component-based twins (such as those for a valve or a pump) and asset-based twins (such as the motor or vehicle that contains the pump). Both twin types have been in use for decades, with data derived from real-world use and digital-twin tests analysed and fed back into the design and manufacturing processes to improve quality and usability, helping cement customer loyalty.

But a significant chunk of that spending will also be directed towards systems built around sets of material assets that perform a specific function. For instance, a digital twin may model a vehicle assembly line or a minerals-extraction process. And some enterprises will take things a step further, connecting their twins to operational and business data, enabling product owners, finance managers and even chief executive officers to check the potential impact on safety, costs and the efficiency of assets and processes.

Taking twins further

It is in these novel areas where digital twins can contribute significant additional business value. For instance, auto (and other) manufacturers may use the sensors and processors in their finished vehicles to collect information on quality, functionality and usability. The data can then be analysed and fed back into twins that are used for designing and engineering, prototyping and production processes, creating a so-called “closed loop”. This in turn allows further experimentation and refinement, creation of post-sale service and maintenance models, scheduling of predictive maintenance and forecasts on service-related costs, revenue and repurchase revenue from long-term customers. (See figure.)

Closed loop digital twin


Source: DXC Technology, 2019 (modified by IDC)

Adopting a systems-thinking approach has led to the scaling up of the twin concept. Ports such as Rotterdam, for example, are aiming to use their internet of things (IoT) networks to create digital twins of their operations. They already have sensors and control modules heavily peppered throughout their rails, thoroughfares, traffic control systems, vehicles, docks and container cranes. The data has been collected and pushed into dashboards, and now just needs to be analysed and organised into a virtual representation. Similarly, a European Commission accelerator programme for precision agriculture is encouraging farms to embrace the digital twin concept. IoT arrays used to monitor everything from moisture levels and machine use to animal movements and production processes have been used to pilot twins at dairy farms in the Netherlands, feed silos in Spain and bee colonies in Greece.

Already standard tools

Creating a digital twin is no small feat. It can require a plethora of aggregated data streams representing how, when, where and under what conditions a device or system is used. Infrastructure and application environments must be architected for managing and analysing the data. Modelling and graphics tools must connect the data to 3D presentations of things and places. Engineering and manufacturing firms (and utilities to a lesser degree) have made the effort because the costs of building and testing virtually are generally tiny compared to production of finished products.

These firms will continue to expand their use of digital twins. While complex digital twin deployments are still in their early days, spending is expected to soar by around 14 per cent annually for the foreseeable future. Perhaps more important, technological development is accelerating, and even conservative futurists are anticipating twins of organs and perhaps entire human bodies within the next couple of decades, a telling indicator of their potential. It’s evident that any organisation with an IoT array should at least consider looking into digital twins.