Transforma logo

An abundance of Digital Twins

NOV 27, 2019 | Jim Morrish
region: ALL vertical: ALL Internet of ThingsHuman Machine InterfaceArtificial IntelligenceData SharingProduct Lifecycle Management

There is a lot of discussion of Digital Twins in the Digital Transformation (DX) space. The concept might at first seem like a simple one, but the underlying detail reveals a lot of complexity. Much of this complexity is caused by the fact that there’s nothing that you can actually ‘do’ with a Digital Twin by itself, kind of like the keystone in a bridge they are only really valuable when used in context. DX technologies that rely heavily on digital twin concepts include: AI, Data Sharing, Product Lifecycle Management, 3D Printing, Robotics and Human Machine Interface. To some lesser extent, Digital Twin can also support IoT, Distributed Ledger and RPA solutions.

The types of Digital Twin

First, let’s look at the different kinds of Digital Twin, which we at Transforma Insights view as lying on a spectrum from ‘digital shadow’ through to ‘digital reality’. The main types of Digital Twin are:

Embedded, where extensive information is available on-board a device, for example a printer with its’ own web page.

Representative, which is a set of data (remote to the device) that represents the current state of a device, or asset.

Shell, where software acts as a shell around a device, relaying information about the device and associated sensor information, and accepting instructions.

Historical, which includes historical information about an asset, potentially including the batch numbers of component parts, and exposure to any events that impact expected lifespan.

Model, which we define as a software programme that can emulate the performance of an asset under certain specified conditions, and which can function independently of the asset. Multiple models for different assets can be combined to create a modelled environment.

Test, where the digital precedes the physical: product (or system) development and production modelling take place in a virtual environment.

Physical (and we’re introducing a new term here: “Physical Twin”) is the logical conclusion for digital twins, where physical assets are effectively a representation of their twin in the cloud. Incremental materialisation is the system-by-system, or component-by-component, physical creation of elements of a digital environment which are then effectively spliced back in to the digital original.

Relationships between Digital Twins

That’s a lot of different potential types of Digital Twin, but the complexity doesn’t stop there since any given device (or asset) can potentially be associated with multiple different Digital Twins to support different use cases.

Take, for example, a steel mill with a (simplified) production line including sintering, a blast furnace and associated gas boiler and a rolling machine. Firstly, the manufacturers of each of these production line assets may have their own digital twin of the device, monitoring performance in detail, supporting pre-emptive maintenance and compiling a detailed history for the device in question. But these individual and quite possibly very sophisticated Digital Twins won’t necessarily ‘talk’ to each other, so it will often be necessary to construct a different Digital Twin for each device to support the end-to-end steel manufacturing process. And, again, different Digital Twins of devices may be used to support building information management systems in any manufacturing facility. True, the information in these extra Digital Twins may be sourced from the original manufacturer’s Digital Twins, but the information will reside in different systems and be used for different purposes.

This relationship is illustrated in the graphic below.

Digital Twin.jpg

Where to start with Digital Twins

Given all of this complexity, where to start? Well there are a few key steps for adoption:

• Focus on the Use Case: Digital twins should be developed with a view to a specific use case, or range of use cases. The range of information that can potentially be included in a digital twin is near infinite – in many cases, the costs will outweigh the gain.

• Start small and grow: The more information that a Digital Twin provides, the better, but it’s a relationship of diminishing returns. Basic baseline information in a digital twin can be incredibly useful and an incremental approach helps the concept to get established within an enterprise.

• Plan for the future: Once established, is it likely that Digital Twin concepts will expand in scope and sophistication. Managers should therefore plan how to include a potentially extended value chain, and share information between enterprises. Good documentation is crucial, and open software advantageous.

NOV 25, 2019| Matt Hatton Previous Post
Why does the world need Transforma Insights?
Next Post DEC 02, 2019| Jim Morrish
Digital Transformation is a spectrum
All Blog Posts