Key points
Since the 1990s, digital twin technology has been a compelling concept. It has taken its time to come to fruition: the idea of a virtual mirror simulating real-world interactions in a digital realm was for years based firmly in the future (or if employed at all, confined to multi-billion-dollar operations such as NASA missions).
But thanks to increased processing power capable of handling the vast amounts of data involved, digital twins are finally about to have their moment. By the end of the decade, these powerful simulations enabling seamless interactions between the physical and virtual worlds are likely to become an everyday feature of manufacturing. The question is, what will this change mean for the process industry – and how can companies best harness its potential?
Data insight and its benefits in manufacturing
Knowledge is power – and so the more data that can be collected, analysed and acted upon, the more we can achieve. That’s something we already know, and it’s a maxim the process industry puts to use in everything from timesheets to ambient temperature monitoring.
In essence, digital twin technology is just another form of data insight – albeit one that is highly accurate, responsive and interactive in real time. Along with other Industry 4.0 innovations such as cloud computing, artificial intelligence and the Internet of Things, it provides vital information allowing industry to hone and streamline its processes, driving efficiencies to maximise productivity.
Why is data so important? In the manufacturing industry, data insight:
- reduces overheads
- improves product quality
- anticipates problems in advance
- informs processes to get the best yield and maximise profit
And while all forms of data insight are valuable, virtual reality simulations such as digital twins have the potential to provide a significant return on investment. Once the technology is ready to go, the virtual replica is cheap to run: it allows manufacturers to experiment with alternative approaches without the risk and costs associated with setting up real-world processes.
The role of digital twin technology in manufacturing
A digital twin can do more than simply visualise potential. When implemented in a production environment, it introduces a unique, real-time synergy between the physical and virtual worlds. Sensors fitted to measure both machine performance and external factors such as humidity and ambient temperature provide a continuous source of real-world data, which may be augmented with input from other sources such as CAD models.
This data can then be fed back to the physical world to enable changes that could either be automated entirely, or mediated by a human operator. These changes are likely to be made immersively in augmented reality, using technology such as holograms and smart glasses to interact with the system.
The real-time simulations created by digital twin technology can be used on production lines for a number of practical purposes:
- Predictive maintenance, identifying any likely problems or malfunctions before they arise, enabling decision makers to choose the optimal time for repairs so as not to interfere with production
- Virtual prototypes, which allow manufacturers to optimise product design while cutting the cost of manufacturing and reducing production time
- Monitoring and accountability, keeping track of every stage of the production process including all parts and raw materials
- Modelling different scenarios to visualise what scenarios might cause a bottleneck in production or the ramifications of a product shortage. This provides the information needed to guard against risk and make effective decisions on the fly.
Examples of digital twins already in use across different sectors
Ford (1) employs digital twin technology to impressive effect in its vehicle manufacturing process, with seven separate digital twins used in the production of each different vehicle model. High-resolution digital twins are responsible for different tasks including design, build, manufacturing and customer experience. Notably, the twin responsible for monitoring production facilities enables enhanced efficiency across the production process, particularly concerning energy conservation.
Unilever PLC (2) uses digital twins to ensure flexibility and resilience, in case of supply chain fluctuations. With eight digital twins operating across North and South America, Europe and Asia, the technology is used to simulate hypothetical scenarios and enable manufacturers to react swiftly to unexpected developments, ensuring the optimal output.
Atos and Siemens (3) are collaborating on a digital twin solution for the pharmaceutical industry, helping the sector to reinvent the manufacturing process to optimise quality and reliability. Working with GlaxoSmithKline, the team aims to digitalise the vaccine manufacturing process to allow for faster development and production.
Piloting digital twin technology
While digital twin technology is just beginning to be adopted by big-name companies, it is not in general use just yet. However, as more companies follow on from the early adopters to capitalise on the advantages of this technology, its adoption is likely to trickle down over the next few years. It’s predicted that digital twin technology will be incorporated into more IoT solutions, with up to 89% of IoT platforms including digital twin capability by 2025, and 36% of executives planning to incorporate the technology by 2028.
By the end of the decade, it’s likely that the landscape will look very different, with many of the oft-predicted benefits of digital twin technology finally becoming (more than just virtual) reality.
With this in mind, what is the best way to move towards adopting this new technology? As with any change, it’s wise to pilot a digital twin by starting off with a fairly simple, discrete implementation. To get true value from the experiment, the process you are testing must be valuable enough for there to be a potential benefit, and there may be questions that have so far remained unanswered.
Perhaps there are issues with production that aren’t fully understood, where digital twin technology could provide more insight. The digital twin can be introduced in iterative cycles, making adjustments each time, to get the project off the ground quickly.
At Adatis, we already help our customers to understand their data, create models and adopt machine learning. As digital twins evolve to become part of the landscape it’s likely they will become another tool at our disposal – arming our clients with big data analysis on a scale never seen before.
Sources:
José Mendes is a Principal Data Analytics Consultant for Adatis Consulting with over 11 years’ experience in delivering Microsoft Azure/ SQL Data Analytics solutions. Presented for the first time in 2017 at SQLBits and since then has been blogging and speaking about the Bot Framework, Modern Data Platforms, and other cool Azure services.
