As an example, adding enough sensors to create a digital twin of a car will require not only digitally replicating the shape of the vehicle, but also the tires, seats, engine, and even mirrors.
But things get far more complicated under the hood, where the inner workings of the engine will require a real-time simulation including every spark, explosion, and movement inside the cylinder block, pistons, crankshaft, valves, and plugs, capturing every distortion, glide pattern, and even the slightest bit of of friction happening in real time.
By advancing from a single asset view to a larger asset population, organisations can unlock new opportunities to enhance operations. For example, an engine, transmission, and braking system may all have discrete digital twins, but will need to interact with each other every bit as much as the real engine, transmission, and braking system do in order to achieve even deeper insight into overall system behaviour..
But Digital Twin progress doesn’t happen instantly. As we step from detail and accuracy to micro-detail and micro-accuracy, these super high-tech 3D models will enable us to visualise how our physical products are performing and changing in the moment.
If something breaks down, we can instantly tell what went wrong.
Experiments like this will not only require huge amounts of computing power, but also a massive range of computational approaches to simulate unique techniques for organising and interacting with functional responsibilities of work.
To put things in perspective, the drones we fly today have been in development for decades. It has taken that long to get to get the vehicles this good. With emerging technology, we still have to work our way through the primitive stages before we get to the good stuff.
At the same time, we are building a digital infrastructure that is layered over everything physical in the world. This is another form of digital twin advances and eventually the two will align.
A digital representation of a physical object, digital twins allow directors to create a crystal-ball-like-view into the future. They enable simulation, analysis and control to test and explore scenarios in a practice setting before initiating changes in the real world.
While digital twins have historically been associated with more complex technology environments, its impressive ability to both eliminate problems and deliver next-level operational performance is making these models a must-have technology in every unit toolkit.
Some of the first digital twin cases involved complex—and usually expensive—capital assets such as diesel engines, turbines, and heavy-duty construction equipment. Their digital representations are equally complex, comprising finite state machines with never before seen numbers of discrete states. However, digital twins offer even the simplest constructs a vast number of benefits.
The single biggest problem with digital twins is that one size does not fit all. In other words, a new digital twin is needed for every single product that is produced and the process that creates them. That’s because every product, no matter how precisely it’s made, operates differently. This is especially true if there are workers involved in manouevres.
Adaptive Digital Twin offers an adaptive smart model user interface to the physical and digital twins. The adaptive user interface is sensitive to the preferences and priorities of the user/operator. A key capability at this level is the ability to learn the preferences and priorities in of human operators in different contexts. These characteristics can be captured using network-based supervised machine learning tech.
The models employed within this digital twin are continually updated based on the data being “pulled” from the physical twin in real-time. It can also accept information in batches after system use. This digital twin can support real-time planning and decision-making during operations, maintenance, and support.
Advanced digital twins have unsupervised machine learning capability to discern objects and patterns encountered in the operational environment, and reinforcement learning of system and environment states in uncertain, partially observable environment. This digital twin at this level has a high degree of autonomy. At this level, the digital twin can analyze more granular performance, maintenance, and condition data from the real-world counterpart.
Digital twin system simulations can be run to explore failure modes, leading to progressive design improvements over time. For example, a manufacturer can link a digital twin to its service history, manufacturing process, design history, real-time network data, configuration-specific simulation models, and expected failure modes.
The ability to compare simulation outputs with actual results can provide valuable insights about the physical twin. Users can generate event-driven or agent-based simulations to explore the behaviour and interactions of the digital twin. The digital twin can incorporate 3D data and simulations, along with their characterisations using methods such as response surface models. To gauge customer experience and the impact of innovation on that experience, the digital twin can be employed to simulate a plant, product or service. For example, barcoding line replaceable units for analysis in logistics support.
Engineers can potentially use simulations linked to the digital twin to predict how the physical twin can be expected to perform in the real-world. Contrast this with having to rely on the ideal and perceived worst-case conditions typically employed in the design process.
Actual system performance data can be compared to data from the digital twin, prompting adjustment decisions that can contribute to successful mission outcomes. Also, by incorporating data from the physical twin into the digital twin, engineers can improve system models, and subsequently use the results of the analysis with the digital twin to improve the operation of the physical system in the real-world.
The value added by a digital twin to its physical twin stems from its ability to optimise both physical twin operation and maintenance schedule. Simulating digital twin behaviour enables the determination and adjustment of real-world system behaviour. Specifically, insights gained through simulation can guide changes needed in system design and manufacturing with the digital thread providing the necessary connectivity across the system life cycle.
The fidelity of the simulation will typically vary with the purpose of the simulation and the stage in the system life cycle. A potential benefit would be to rehearse missions in specific operational conditions, terrain, weather, etc. using the capabilities of the digital twin. For example, in the system design phase, a relatively slow, non-real-time simulation may be enough as long as it enables exploration and investigation of multiple different use cases under real world conditions. With access to the right data, actual operational conditions can be simulated with high confidence to yield insights into expected outcomes.
An important use of simulation is in the assessment of the expected operational life of the system i.e., low long the system can be expected to be operational, so digital twin can keep track of the wear and tear experienced by the physical twin. By employing simulation, the digital twin can estimate the remaining working life of the physical twin and proactively schedule maintenance. In other words, predictive maintenance can be used to estimate how long the physical system can be expected to operate normally and use that knowledge to proactively schedule and perform system shut down rather than wait for the breakdown of the physical twin which can be both expensive and potentially catastrophic.
For simple applications, digital twin technology offers value without having to employ machine learning. Simple applications are characterised by a limited number of variables and an easily discoverable linear relationship between inputs and outputs. However, most real-world systems that contend with multiple data streams stand to benefit from machine learning and analytics to make sense of the data.
Machine learning is applied to a data stream to uncover/discover patterns that can be subsequently exploited in a variety of ways. For example, machine learning can automate complex analytical tasks. It can evaluate data in real-time, adjust behaviour with minimal need for supervision, and increase the likelihood of desired outcomes.
Machine learning uses within a digital twin include: supervised learning of operator/user preferences and priorities in a simulation -based, controlled experimentation testbed, unsupervised learning of objects and patterns using, for example, clustering techniques in virtual and real-world environments and reinforcement learning of system and condition states in uncertain, partially observable operational environments.
Linking digital twins to networks brings the data needed to understand how the physical twin e.g., manufacturing assembly line, autonomous vehicles network behaviour and performance in the operational environment to enhance preventive maintenance and AI optimisation of the physical system and operational processes.
Acting as a bridge between the physical and virtual worlds, the network can deliver performance, maintenance and condition data from the physical twin to the digital twin. Combining insights from the real-world data with predictive modeling can enhance the ability to make informed decisions that can potentially lead to the creation of effective systems, optimised production operations, and new manoeuvre models.
Multi-source/multi-sensor information e.g., temperature, moisture content, production status of current batch can be delivered to the digital twin along with information from traditional sensors to facilitate predictive modeling, providing much needed flexibility when it comes to system mobility/location options such as selling a capability i.e., product as a service versus selling the product itself.
Importantly, the combination of digital twin and the network allows an organisation to gain insights into how a system/product is being used by customers. Such insights can enable customers to optimise maintenance schedule and resource utilisation, proactively predict potential product failures, and avoid/reduce system downtimes.
Perhaps the greatest potential benefits of the network is in the service area. For example, service that is continually informed about the operational state and condition status of the system can be effective in ensuring cost savings and high availability rates to conduct manoeuvres. For example, predictive analytics can be employed to pre-fetch and rapidly deliver a required part to a maintenance crew.
In the future, digital twin technology can be expected to become a central capability in system engineering models because it can span the full system life cycle while at the same time helping to penetrate new markets such as manufacturing, construction, etc. . Specifically, digital twins can be exploited in upfront engineering e.g., system concepts and model verification, testing e.g., model-based system validation, system maintenance e.g., condition-based maintenance, and smart manufacturing.
In the near-term, digital twin technology can be expected to be integrated into manufacturing and maintenance activities to enhance predictive maintenance and design. At the same time, digital twin technology will continue to gain ground in each sector because of real-time access to system data.
Maintenance will be a major contribution area for digital twins. For example, digital twins will help organisations transition from schedule-based to condition-based maintenance, thereby substantially reducing system maintenance costs while also enhancing system availability.
Digital twins can be exploited in aircraft engine maintenance. Currently, aircraft engines are routinely taken apart and rebuilt based on the number of hours flown, regardless of whether most of those hours are simply cruising at altitude or performing high-G maneuvers. Digital twin technology can help us better understand maintenance needs and schedule maintenance accordingly.
The need for digital twins originates from the need for increasingly advanced dashboards for users to view everything they need to know about their projects and properties in a convenient form and format. However, as many sectors continue to digitise and data becomes increasingly more complex and abundant, there has been an exponential growth in the number of tools, data formats and data services.
As a result, generating actionable information from diverse data supplied by different tools is becoming increasingly difficult. A next-generation dashboard tied to one or more digital twins can dramatically improve timely decision-making and plan execution on many projects.
For architects, engineers and planners, digital twin technology can become a source of sustainable competitive advantage by linking projects and to real-time data with user-customisable smart dashboards. Manoeuvre zoning, and traffic data can be linked to the digital twin, and accounted for in making siting and structural component placement decisions. For example, traffic information can be used to determine where to place an entrance to a building.
Digital twin technology can provide a window into system performance, for example, a digital twin can potentially help identify equipment faults and troubleshoot equipment remotely to alleviate key customer concerns. In addition to improving virtual system models, a digital twin can potentially improve physical system operations and sustainment.
A digital twin can also help with product differentiation, product quality and add-on services. Knowing how customers are using the product post-purchase, can provide useful insights including identifying and eliminating unwanted product functionality and features, as well as unwanted components, saving both time and money in the process.
A digital twin can enable visualisation of remote physical systems to troubleshoot a landing gear problem of an airplane parked at another airstrip. Multimodality sensors e.g., sight, sound, vibration, altitude can serve to deliver data from physical systems to digital twins anywhere in the world. These flexible capabilities can potentially lead to clear understanding of the state of remote systems through multi-perspective visualisation.
Digital Twin Planners are working extra hours on answering questions standing in the way of wide-scale digital twin deployment. In the meantime, digital twin technology is continuing to make impressive progress in aerospace, defense, manufacturing, site construction and other applications.
Once we’re able to produce a virtual pairing with the physical/behaviour world, we suddenly have the ability to analyze data streams and monitor systems so we can head off glitches before they occur, prevent interruptions, uncover new opportunities, and even test new strategies with quickly contrived digital models.
The combination of new manoeuvre formations assignments, information and data analysis will quickly turn most skeptical leaders into strong advocates of digital twin tech.
1. Smart Command Centres
With advanced networks entering our battlespace, having a smart command center becomes a logical extension of our need to monitor and manage operations movements. Every ship, airplane, ground vehicle, or turban in a power plant has the potential for being digitally replicated
2. Directing Effective Field Operations
Troops will soon have their own fleets of drones, with scanning capabilities, to create digital models of the battlespace. As scanners, sensors, and resolutions improve, planners will be creating increasingly functional digital twins of their roadmaps, and activity centers.
3. Platooning
The first phase of remote robotics for trucking will involve platooning where human drivers control the lead vehicle, followed by driverless vehicles. Since the driver is still in control, additional support won’t be needed until it arrives at the delivery location where either addition human operators can take the controls or remote drivers can manage vehicles for the final positioning of the truck.
4. Remote Operation
The actual operators may be working in another location, but having a person at the controls is critical for certain situations. Drivers, pilots, and captains do far more than just drive their vehicles. They provide a contact person to talk to, provide security, situational awareness, and the type of oversight and responsibility that only a human can provide. Since there is no such thing as an infallible machine, things will go wrong. When this happens, we will need a live person to manage the problem. The solution may be as simple as a system reboot, but in extreme cases, emergency rescue people will need to be involved, and having a contact person to coordinate the response is critical.
5. Search Engines for Formation Detect
Online search technology has framed much of our plans around our ability to find things. In general, if it’s not digital and online, it’s not discovered. In the future, drones and sensors will replace much of the work of today’s web crawlers when it comes to defining our searchable universe. Search technology will become far more oriented to specific attributes in the future. Over time, search engines will have the capability of finding virtually anything in either the digital or physical world.
6. Monitor and Enhance Performance
We already have several tools that can create a digital map of our troop performance, both external and internal, like 3D laser scanners. We also have a growing number of contact and embedded sensors that can track what is going on. We are developing complete digital image of troop behaviour that can be rotated around, zoomed in for close-ups, watching physiological episodes and operational resources. When we finally develop holographic displays, our ability to gain relational perspectives, as well as cause and effect relationships will only increase.
7. Predict the condition of an asset
Down the road, it will be necessary to transcend a slick dashboard view of a device’s current anatomy, and truly understand its behaviour. Digital twins provide a closer look at what conditions and events influence it to change, to regress or thrive, from one environmental state to another.
8. Increase accuracy
Understanding behaviour patterns, and leveraging advanced machine learning tools, enables meaningful digital twins to be played forward or backward in time. This modeling allows operators to better understand how a device might perform in a certain scenario, for example, to alleviate a potential mechanical failure before it happens.
9. Avoid Equipment failure
Military places high value on being able to avoid costly breaks or errors. Digital twins enable teams to explore innumerable possibilities so they can deliver, with a higher level of confidence, a recommendation around the longevity or reliability of an asset. Units are wanting to improve uptime and increase production and can leverage digital twins to do so more quickly.
10. Create composite digital twins
Digital twins make it easier for applications to interact with remote devices, whether to query them for conditions or instruct them to perform certain actions. However, digital twins containing only real-time physical state information are limited in their utility. The next step in digital twin development is to assess behaviour schematics.