Digital Twins - Game changer or just shiny toys?

The use of twins to duplicate and mimic an operational platform has rapidly increased since the first recognised use case purportedly on the Apollo 13 program in 1970. Immediately after the realisation that Apollo 13 experienced a critical failure from the sudden explosion of its oxygen tanks, NASA was able to replicate the scenario and plan recovery using the twin it had built back on Earth. The use on that twin ultimately saved the entire crew.

However building a physical twin is obviously not always possible especially when the platform is the size of an aircraft carrier or a submarine, enter the Digital Twin. Advancements in technology have rapidly expanded the digital value chain across manufacturing organisations. PLM, CAD, IoT and extended reality have collaboratively brought together the ability to do what was previously impossible, the ability to be immersed within a platform or environment without physically being there. Enter virtual operations.

Live IoT sensor data transposed on a desktop digital twin view of an engine room

This capability is a true needle mover however most organisations are yet to document or even begin to assimilate the value proposition

This extended value chain and digital twin technology has the ability to unlock immense value across numerous use cases in turn amplifying business outcomes. A digital twin can now in effect become like the cockpit of an aeroplane. The entire manufacturing environment can be controlled in a central location. And not only can the twin view information being sent from within the environment, it can also send information back to that environment in order to turn things on and off, control machinery or initiate evacuation procedures. This capability is a true needle mover however most organisations are yet to document or even begin to assimilate the value proposition. So in this article I want to pull apart some of those value outcomes in the hope that it provides clarity to the most commonly asked question "Digital twins sound great but do they actually provide business value?".

Once again, I am going to focus this article on shipbuilding due to the sheer volume of global activity at the moment. To keep it relatively brief, I will focus on three key areas of a platform's life-cycle that we can apply digital twins to:

1. Twinning the manufacturing or operational environment (e.g. a shipyard)
2. Twinning the to-be platform (prefabrication design)
3. Twinning the operational platform (post fabrication)

1. Twinning the manufacturing or operational environment (e.g. a shipyard)

Although the key focus over the past decade has been around modeling digital twins for specific platforms such as ships, there is enormous value in doing the same for the environments that manufacture and support those platforms. In fact the value has been considerably underestimated as these environments are generally steady-state, last decades to centuries and cater for the build and support of numerous platforms. It therefore makes sense to suggest that the ROI is easier to define and quantify and therefore the business case is more straight forward and self supporting.

When I thought about writing this article I wanted to challenge my notion that there are so many valuable business outcomes that could be realised by digitising a shipyard that it would be hard to know what to include in this article. I achieved this by a) engaging the services of my daughter as I love to get her involved (I paid her with XBOX credits) and b) getting her to write down as many ideas that came out of my mouth in 2 minutes. The result is the list below (excuse her spelling, she is only 10). Total of 17, so an idea every 7 seconds, not bad. Imagine how many your organisation could cobble together in an hour with a clean whiteboard and a table of caffeinated team members!

From that list I decided to hone in on five prominent use cases which we (Accenture) have subsequently incorporated into development sprints to bring them to life. Hopefully I don't need to go into too much detail for each in order to draw out the business value outcomes as I believe they speak for themselves. I have however added outcome icons (blue circles with icon) to each image to clarify which include: SAFETY, EFFICIENY, PRODUCTIVITY, COMPLIANCE, QUALITY and NWOW (New ways of working) As you look at each I want you to envisage the value outcomes with the context that an average naval program lasts 35 years and that it is possible to implement a proof of value (POV) for each use case for under $100k AUD. So for under half a million Aussie dollars, you could implement the capability to detect toxic gas anomalies, track vehicles, view temperature anomalies and locate parts within a shipyard. But before anyone goes and asks the CFO for some dollars, bear in mind there needs to be adequate IT infrastructure to provide data access and you will eventually want to spend money on the right security architecture before rolling it out into production. What I am trying to highlight here is that for the value returned, the cost to implement is trivial.

TOXIC GAS DETECTION

It is not hard to understand the value outcome of this use case as anything that helps prevent injuries or heaven forbid fatalities, should be a zero brain cell decision. By deploying gas sensors in strategic locations across a shipyard a digital twin can show a very intuitive and easy to pinpoint view of where issues might be unfolding. As you can imagine with a capability like this you don't want to wait for a user to open the twin and assess if there is an issue rather the twin would sense and send a notification (sms / email / wearable) to notify user/s of the anomaly. There are also options to extend this use case so the twin automatically initiates the vibration of wearable smart devices (e.g. smart bracelet / watch) on anyone within a certain radius of a gas anomaly. The wearable could also display a summary of the issue and direct individuals to the appropriate muster point.

VEHICLE TRACKING

Vehicle tracking might be a less obvious use case however it has the potential to save both lives and money. Based on evolving events in a shipyard such as fire or chemical spillage, vehicle management can ground vehicles, allow them to only drive away from the incident and even autonomously move vehicles to a safe location. Tracking vehicles can also provide an insight into routes taken between various points in order to optimise task sequencing to save time and money.

MATERIAL TRACKING

For a reason unbeknownst to me, material tracking seems to be one of the least matured processes in a shipyard. At the mature end of the spectrum I have seen an organisation track material with 2d barcodes and an Access Database. At the bottom end I have seen everything from spreadsheets to physically writing locations on paper to "Ask Wayne where it is". And although humans have a pretty remarkable brain, committing more than about fifty part locations to bio storage (memory) isn't going to happen. It is also not currently possible to "back up" a brain and therefore if that brain is stuck inside its host who is lying in bed with the flu, nobody is going to know where their parts are stored.

The main reason all of these methods are ineffective is because they rely on an attribute labelled "Last known location". If someone has moved a part and didn't use the barcode scanner or didn't write it down or didn't tell Wayne, the information is now useless. This scenario leads to tens of millions of dollars of material waste each year in Australia alone.

So how does the twin help? As most of us now know, IoT goes hand in hand with digital twins. Without real time, dynamic data feeding the twin, it is nothing more than a 3D mashup. So with the help of RFID sensors and smart barcodes, material can be tracked automatically as it moves around the manufacturing environment. With this information being fed into the twin, a user can quickly find and actively confirm a part is in the recorded location. This is also a very effective way to track and locate classified, sensitive or export controlled materials. For example, if an item is tagged as "UV susceptible" the AI within the digital twin will understand if the item is being stored outside and if so, raise a notification.

TEMPERATURE MANAGEMENT

Managing an organisation's workforce when climatic conditions become dangerous is another easy to solve use case for the twin. If temperatures escalate to unsafe levels such as in the image below, the twin can be used to undertake various actions based on business rules such as:

• Alerts can be issued to phones and wearables
• Evacuation lights can be activated
• Backup AC can be automatically switched on
• Machinery can automatically be turned off

EMERGENCY MANAGEMENT

Who could possibly argue that a use case involving the potential saving of life isn't the most relevant of all to enhance with technology? A "Person-down" alarm is nothing new however integrating it into a twin in order to enable proximity context and direct integration into response management systems is. The scope of person-down has also broadened with the evolution of wearable technologies. The issue with traditional person-down alarms is that it wasn't until the person had suffered an event that it activated the alarm. With modern wearable technology, lead indicators such as heart arrhythmias can activate alerts, life-saving minutes before an event.

Rules to manage different scenarios can be created and edited to define what constitutes an event and more importantly, what to do with it. There are some obvious events that should trigger an immediate alarm without any requirement for validation such as the press of a panic button. For other scenarios such as changes in vital signs like heart-rate, checks can be made to determine whether an alarm needs to be raised or not (i.e. They are not just going for a lunchtime jog).

When an alarm is triggered, for example when an individual presses a panic button, resources can automatically be notified and dispatched. For less critical scenarios such as when a heart-rate or temperature fluctuation has been detected, the twin can provide options to the operator to either dispatch resources or ignore the detection.

CONTRACT TRACING AND EVENT SIMULATION

Undoubtedly this is very topical. What would the introduction of one or two individuals with a highly infectious virus mean for a shipyard with thousands of workers? Using a twin, this can be modelled (simulated) and more importantly back-traced once an infectious individual is diagnosed. The example below demonstrates the use of AI to simulate how quickly the infection might spread across an operational area like a shipyard. Using wearables to collect tracing data, you can literally press rewind on the digital twin to see how many people an actual infected individual came into contact with over a certain period. With that information you can then go and isolate those individuals from the environment.

2. Twinning the to-be platform (prefabrication design)

Using a twin for the design of a platform could arguably be the most valuable application of this technology. Why? Because it has the potential of savings thousands of hours and millions of dollars of rework. How? By removing the majority if not all of the design flaws before even a single line of weld is laid. As the icon ribbon above would indicate, all desired business outcomes can be addressed using a twin in this phase including safety, the most important of all. Disasters such as that which happened on HMAS Westralia could have been avoided by effective modelling and simulation in the design phase using twin technology.

There are a number of ways simulation through a twin can remove potential defects. Artificial intelligence (AI) and high performance computing (HPC) can be used to simulate processes and events that previously could only be done after the platform was built. Similarly, design flaws that would normally be found after the platform has been built can now be found in design by virtually walking through the design with virtual reality. Issues like 'line of site' and walkway access through a ship can be assessed by touring the vessel virtually. Some of the key areas that can be assessed with simulation include:

• Fluid and flow dynamics (hydrodynamics and resistance coefficients)
• Damage simulation (i.e. explosion)
• Thermal dynamics and fire risk
• Radio frequency and electromagnetic penetration
• Piping and HVAC optimisation
• Design efficiency (ease of access, route and task optimisation, etc.)

The use of virtual reality as an entry point into the twin has proven to be incredibly value in assessing whether a design will be practical from an operational perspective. As the video below shows, a user can walk around areas of the design using VR, detect any issues such as access obstructions and modify the design dynamically.

Fire simulation is another great example of how a digital twin can model safety issues prefabrication. Getting this right first time is absolutely essential. Through the use of a Digital Twin and the cognitive power of Artificial Intelligence, the subsequent spread of fire from an ignition can be predicted. This simulation is not just based on the physical layout of the ship but also on a number of other factors such as compartment materials, flammable substances, ventilation, installed fire suppression systems, etc. Physical and behavioural attributes are tagged to the twin so the AI layer can determine factors such as how long it will take to burn through different materials. At the bottom of the image below you can see that the application has determined there are eight walls that are vulnerable to fire and should be further insulated. It has also determined that there are four zones that are not protected by fire suppression (sprinklers) which should also be addressed.

3. Twinning the operational platform (post fabrication)

The use of digital twins in operational environments is arguably the most exciting, forward thinking and needle moving application of this technology. It also represents the most complex and risky application so an investment needs to be made in securing and architecting the solution with care. With robust architecture and a secure data delivery mechanism, there are some powerful use cases that can be bought alive with twins.

Virtual Maintenance

Maintenance is a key activity on platforms of any size, from personal automobiles all the way up to large warships and skyscrapers. A twin can be lit up with information collected across thousands of sensors integrated into these platform. This can create a three dimensional dashboard of information and identify a vast number of events, interactions and issues happening in the physical twin in real time. So in theory I could pop on a Virtual Reality headset and walk around the digital twin of say a ship and view the exact same events, interactions and issues virtually as they are happening in the physical world. That is insanely powerful as it can provide visual and dimensional context to a mechanical issue. If we wanted to simplify that result, we could just provide collected sensor data in the form of a dashboard on a PC. An individual could then drill down through sensor information from a web application. However in order to expedite root cause analysis we could overlay that sensor data onto the 3D visualisation through the use of VR provide a view that has the added benefit of visual context giving rise to the birth of the "Virtual Maintenance Engineer". Given visual context, an experienced engineer who knows the physical platform intimately can pinpoint issues far faster than with just data and they don't even need to be on the ship.

So in theory I could pop on a Virtual Reality headset and walk around the digital twin of say a ship and view the exact same events, interactions and issues virtually as they are happening in the physical world.

As you can see in the image above, a virtual maintenance engineer can use a VR headset to navigate the Digital Twin, in this case a 3D scan of a compartment, and view real-time data being collected from the Physical Twin. This can provide the engineer with visual context to understand what the issue might be, in this case a coolant flow issue raising the temperature of a diesel generator. Although there would most likely be an engineer attending to the issue on the Physical Twin, a virtual engineer who may have subject matter expertise could be in another location viewing the same interaction within the Digital Twin. Not only can the view provide key sensor data from systems and equipment within the compartment but the virtual engineer can also pull in sensor data from other parts of the ship as required.

In fact I could bring up the 3D view of an engine room on my mobile phone for a vessel that's out at sea, whilst I am standing in my front yard watering my garden

The same can be done using a twin viewed through augmented reality. In fact I could bring up the 3D view of an engine room on my mobile phone for a vessel that's out at sea, whilst I am standing in my front yard watering my garden. Like with VR, that view can show me exactly what's going on in that physical engine room out at sea by overlaying the environmental data collected by the IIoT ecosystem on the vessel. That's pretty powerful and if I wanted to take that interactivity to the next level, I could program that digital twin such that I could actually turn systems off on the ship in response to my conclusion that something is about to fail if I don't. The concept can be see below in the video which as it turns out was taken in my front yard whilst I was watering the garden...

Predictive Maintenance

I am sure it's not just me that has a suspicion some products, especially electronics, seem to be manufactured to fail shortly after their warranty expires. I had the pleasure of a TV failing 2 days before its warranty expired recently and thought to myself "Yes, they didn't quite get the timing on this one right!". Luckily, predictive maintenance has since taken the guess work out of when systems and parts will fail.

Predictive maintenance is a powerful capability that is already attracting a lot of attention. It uses two rapidly evolving capabilities, Artificial Intelligence (AI) and Machine Learning (ML). In short, it learns what is happening to different parts of a vehicle or platform and can predict if there is an imminent failure. For example it will learn from previously collected data that when a sensor on a wheel detects higher than average vibration, the wheel bearing is at risk of failure. To maximise the value of predictive maintenance, it should be integrated into the Digital Twin. With the Digital Twin at the core, any number of views can be configured to display valuable information required for different tasks. As an example, an automotive engineer can quickly determine which parts of a customer's vehicle might be close to failure when they bring the vehicle in for a scheduled service. With that information the engineer can discuss the option of replacing those parts during the scheduled service instead of running the risk of parts failing between services. Ultimately the objective of predictive maintenance is to maximise the up-time of the vehicle.

Uptime of a Defence asset is particularly important, in fact it is the number one KPI when it comes to in-service support. It could literally mean the difference between the success, and not, of an engagement with a foreign threat or a humanitarian aid effort. Therefore, the value a digital twin can bring to predictive maintenance and in turn, platform availability is significant.

Definitely not just shiny toys

Good business acumen is predicated on an understanding that every business investment needs to be tied to a positive business outcome. So although I have spent my career focused on innovation and technology I never recommend investing in either unless their is a value outcome. Having said that, intangible outcomes can often be as valuable as tangible when it comes to technology. For example, I have seen talent attrition reduce by over 10% simply by providing employees access to technology that may or may not make their role any more productive but makes them feel supported. The reduction in attrition and therefore recruitment costs offset the cost of that technology and actually provided a nett saving.

Ultimately if you are not disrupting you are being disrupted.

Digital twins provide both tangible and intangible outcomes, some of which can be quantifiable needle movers to an organisation. They can take cost out, add revenue in and address the safety, quality and productivity of an organisation. They have the ability to drive organisational nirvana; Faster, Better, Cheaper. They are shiny however they are also immensely valuable and will exponentially grow in numbers over the coming decade. They drive competitive advantage and therefore should be a part of your investment strategy to drive disruption and avoid becoming the distrupted.