Digital Twin – a concept worth exploring
The idea of creating a complete digital copy of a physical asset, system or device is a fascinating one. Having a virtual twin that is acting and aging exactly like its physical counterpart promises new insights into the condition and performance of a product, and may provide valuable knowledge for developing its next generation.
Although introduced in an engineering setting almost 20 years ago the Digital Twin concept is still highly relevant and worth exploring. However, as is often the case with tech buzzwords, the more popular and widely used they become, the harder it gets to define them precisely and to distinguish between the smoke and mirrors of marketing and the real potential of the original concept.
From the 1970s
Actually, the conceptual idea of using a “twin” dates back to the 1970s and NASAs Apollo program. In 2003 the Digital Twin concept was taken up in a manufacturing setting, when NASA consultant and manufacturing scientist Michael Grieves introduced it in a lecture on product lifecycle management, defining its the three components: The physical product, its virtual shadow or counterpart, and the data flow back and forth between the two – data going from the physical product to the model, and information and processes going from the model to the physical product.
Model fed with data
Hans Christian Lønstad, CTO of Data Respons Solutions, elaborates:
– Basically, a Digital Twin is a model of an item and how it operates. The model is fed with real sensor data from its physical twin, so that it can evolve in the same pace as the physical object. For instance, let’s take an airplane engine. It will be affected differently over time dependent on its usage. If you’re flying across a desert and you get in a sandstorm, there will be a huge wear on the engine. Everything is recorded in the airplane, and when it’s grounded the engine manufacturer can pull back all the data and feed it into the twin model. Feeding the digital twin with real-world data and observing it will give the manufacturer a pretty good idea of the condition of the specific item out in the field.
As Lønstad points out, you need to distinguish between the simulation model of the engine, which is a general model, and the specific instances of the general model, which are digital twins of individual engines fed with usage data from that specific engine. In that way you get a collection of data that can predict the condition of the physical item it’s representing with high precision.
According to Lønstad, new business models can be developed on top of this concept, for instance propulsion as a service, with the customer only paying for engine hours instead of buying the engine itself.
– If you can get really detailed knowledge about the condition of your equipment you can do maintenance and plan service more efficiently. Manufacturers of expensive equipment can build good business cases on that knowledge. You can compare it to a car insurance system, where the insurance company is monitoring your driving. That makes it easier for them to choose the right risk profile for you and come up with the correct price for the insurance.
The Cargotec gateway
Data Respons Solutions is assisting Cargotec, one of the world’s leading providers of cargo and load handling solutions, in creating the groundwork for business models based on digital twins. A team of Data Respons engineers has designed a versatile and globally applicable gateway for transmitting usage data from various cargo and load handling equipment marketed under the three Cargotec brands Kalmar, Hiab and MacGregor.
The gateway is an important milestone in Cargotec’s digitalisation journey. According to Cargotec, digitalisation enables new business models and service offerings, and the company’s goal is that in the near future 40 per cent of its net sales should come from software and service. With the technical infrastructure in place Cargotec is now working with utilizing the data collected, doing analytics and adding value to it.
On top of the hype cycle
The Digital Twin concept has gained a lot of interest in the past 5 years, both in industry and academia. It even managed to make it into the Gartner Top 10 Strategic Technology Trends of 2019, and in 2018 it was on top of the peak of inflated expectations on the well-known Gartner hype cycle.
What usually happens is that such a hyped concept then cascades down into the valley of disillusionment. However, if enough people put serious work into it, it will eventually become a productive technology and claw its way up to the plateau of productivity.
In fact, some of the people putting serious work into the Digital Twins concept are working at Data Respons and its subsidiaries. What they’re doing may not fit exactly into the strict definition of the concept put forward by Michael Grieves in 2003. But still it’s all about digital representations of physical objects, processes or systems, and the interaction between objects and their digital shadow or counterpart.
For instance when you take a look at MicroDoc, one of the German subsidiaries of Data Respons, working mainly in the automotive sector. When working in the vehicle telematics area MicroDoc uses Digital Twins to display properties of the telematics units in for instance trucks. On the backend side the individual micro services can work with the twin without having to care about the actual hardware. The twin in the truck will be updated as soon as it connects to the backend.
– We call it a Device Twin, says Nicolas Helou, software architect and developer at MicroDoc.
– Our device management in the backend consists of a database and a number of services around it. The device management knows all these telematics devices in the trucks and knows their current state and configuration.
– We have organized this data in the Device Twins. The Device Twin is also a digital representation, but in this case, mostly of the configuration and the operating states of the devices in the truck. If you want to change the configuration of the telematics device in the truck, you talk to the Device Twin. You can change the configuration in the Device Twin and then it will synchronize this data with the device in the truck. If the device is not turned on at the moment, the Device Twin will just keep the information, and transmit it when the device is active.
– Another technology we’re working with, resembling the Digital Twin concept, is something we call Health Monitoring: On a regular basis each device sends information to the backend about CPU usage, memory usage and other properties, which are all related to the health of the device. We can see if it works as it should, and all the health data is collected in the backend, where we have a graphic representation of all the devices in the system and their current state, both of every single device and of the system as a whole.
No doubt the concept of the Digital Twin will continue to attract interest, and it will be driven by the developments in related technologies like Big Data, IoT, Industry 4.0 and large sensor networks. We’re looking into the future of data-driven manufacturing, and Digital Twins will definitely be a part of that future.