1. Chapter 2: Planning Your Digital Twin
In the previous chapter, we learned why we need Digital Twins and how to use
them to drive specific business outcomes. We explored the history of the Digital
Twin and the various industries in which it presents opportunities.
Now, let's examine how to plan for an industrial Digital Twin in an enterprise
setting. We will identify the key criteria that can be used to determine whether the
industrial Digital Twin is applicable to the business scenarios. Additionally, we will
explore how to develop the business case for investments in the Digital Twin.
Following this, we will explore the prerequisites for the Digital Twin in an
enterprise, including the functional and non-functional requirements. This will
allow us to identify the underlying digital technologies required for the Digital
Twin. Note that we will not lose sight of the organizational factors that are in play
for the success of the industrial Digital Twin initiatives.
In this chapter, we will learn about the following topics:
Key criteria
Expected business outcomes
Prerequisites for the Digital Twin
Organizational factors
Technological needs
Key criteria
Here, we will identify the key criteria to help an enterprise decide when the
introduction of an industrial Digital Twin makes sense. The Digital Twin can be for
a physical asset system or a process such as a manufacturing process in a plant.
Depending on the target of the Digital Twin, objective criteria have to be
established to ensure that the Digital Twin will add business value. Here, business
values and outcomes are used in a broader sense, and they could include the
following:
Improved life of the asset
Process efficiency gains
Operational optimization or lower operating costs
2. New digital revenues
Competitive advantage
Improved end customer satisfaction
Improved safety
A social good such as the reduction of the carbon footprint
As a result, once the key criteria have been established, it is easier to evaluate the
direct and indirect investments and opportunity costs versus the broader business
value generated. Figure 2.1 shows this visually:
Figure 2.1 – Evaluation of the business benefits of the Industrial Digital Twin
In Chapter 1, Introduction to Digital Twin, we discussed the business value and
transformational value as a result of the deployment of Digital Twins in the
industry. The transformation value that is added can, often, be in form of new
digital revenues. The monetization of Digital Twins can be based on a single Digital
Twin or an ecosystem of twins, such as the concept of the Twin2Twin (T2T)
ecosystem. (Please refer to https://www.hcltech.com/blogs/twin2twin-
new- paradigm-enterprise-digital-transformation). The concept of T2T is
similar to Digital Twin Aggregate (DTA). However, T2T can also include disparate
twins, such as twins of buildings and spaces along with twins of heating,
ventilation, and air conditioning (HVAC) and security systems to
efficiently operate a commercial building. These T2T twins are not hierarchical in
nature.
A single Digital Twin can be monetized by the original equipment
manufacturer (OEMs), that is, by charging the customer for the use of the Digital
Twin of the physical product. Such a Digital Twin can be sold as an add-on to the
physical product and have a specific purpose, such as the predictive maintenance
of the asset. The article on T2T, as referenced earlier, states that Digital Twins from
different stakeholders in a value chain can help "to create a vibrant TWIN2TWIN
economy." This will provide monetization opportunities to the different providers
in the industry value chain. Today, marketplaces are common in the cloud
3. computing world. One such example is Oracle's Cloud marketplace, which allows
various stakeholders to create a vibrant ecosystem of cloud computing services
that they can monetize jointly. (Please refer
to https://cloudmarketplace.oracle.com/marketplace/oci.)
We believe similar ecosystems of the Digital Twin will arise in which the different
participants of the ecosystem will contribute to the interdependent parts of the
Digital Twin of the system or Digital Twin aggregate.
The commercial sector is often driven by profitability, so the Digital Twins should
drive business value in general. However, in the public sector, the driver might be a
social good, offering sustainability or an improved citizen experience. Singapore is
experimenting with its own Digital Twin of a city. The government agency
National Research Foundation has created Virtual Singapore, which is useful for
simulating traffic or emergencies. (Please refer
to https://www.youtube.com/watch?v=QnLyy0owGL0&feature=youtu.be.)
Therefore, it is clear that the key criteria for a Digital Twin would vary based on
the stakeholders.
In the next section, we will take a deeper look at the expected business outcomes
for a specific industry in the commercial sector along with public sector scenarios.
Expected business outcomes
In the Industry use of Digital Twins section of Chapter 1, Introduction to Digital
Twin, we discussed the applicable industry segments for the industrial Digital
Twin. Now, let's take a look at the more specific business outcomes that are
expected, or are possible, in some of these scenarios.
The manufacturing industry
When dealing with physical products, the manufacturer is responsible for the
following:
Product design and development
Manufacturing/assembly
Supply chain and distribution
4. Product warranty and reputation
Optional service contracts
Let's take a further look at discrete and process manufacturing within this context.
Discrete manufacturing
Let's take the example of aircraft manufacturing. A commercial aircraft is a fairly
complex product and consists of assemblies such as the body or the fuselage, wings,
two or four engines, landing gears, and the stabilizer. To build the Digital Twin of
the aircraft as a composite asset, the prerequisite would be the Digital Twins of the
major parts of the aircraft. These parts can be manufactured with different OEMs;
for instance, the aircraft could be manufactured by Boeing but the engine could
come from General Electric (GE). The Boeing-designed landing gear might be
manufactured by United Technologies, which is now called Raytheon Technologies
Corporation. Figure 2.2 depicts an aircraft and its major parts:
Figure 2.2 – The major parts of an aircraft
When the composite asset, such as an aircraft, is an assembly of supplier-provided
assemblies, the composite Digital Twin would depend on the collaboration of the
entire supply chain. In turn, each of the major OEMs, such as GE and United
5. Technologies (now Raytheon), depends on several supplies for the smaller
components.
Since we are looking at a commercial aircraft, which is used by airlines to fly
passengers and cargo, we can list the high-level business outcomes as follows:
Aircraft manufacturer: They provide reliable aircraft to the airlines.
Aircraft services provider: The manufacturer might also be the
provider of the maintenance services, and responsible for ensuring
reliability, uptime, and safety in operations.
Aircraft owner or operator: The airline that operates the aircraft is
responsible for the safe and on-time operations of flights offered to its
passengers.
The airline's passenger: The end customer expects timely and safe
flights at air travel costs that are on par or better than other airlines.
The preceding list consists of easy-to-understand business outcomes in the context
of a discrete manufacturing industry where the product is the aircraft. This is a
good example of a business-to-business-to-consumer (B2B2C) business model
where a business such as Boeing or Airbus is selling to another business such as
American Airlines or British Airways and, in turn, offering services to an
individual end consumer who is an airline passenger.
Now that we understand the simplified value chain of the commercial airline
industry, we can evaluate how an industrial Digital Twin of an aircraft would fit
here. If the Digital Twin of the aircraft helps the manufacturer to build better
aircraft or provide better service offerings to the airlines, then it delivers in terms
of the business outcomes. Likewise, if the Digital Twin helps to reduce downtime –
especially any unscheduled downtime for the aircraft – and improve the safety and
efficiency of flights, then it delivers the business value.
While the airline passenger is not directly making a decision about the adoption or
use of the Digital Twin of the aircraft, their customer satisfaction will provide
direct inputs to the business outcomes of the airlines. Likewise, if the Digital Twin
of the aircraft, and the jet engine as part of it, helps to improve fuel efficiency or
carbon emissions, then it improves regulatory compliance and promotes social
good. In our over-simplified model of the aircraft value chain, we have not looked
at several other non-aeronautical stakeholders such as the airports and the ground
services providers. However, every stakeholder benefits from efficiency in the
6. basic aeronautical value chain. If the use of a Digital Twin of an aircraft leads to
less unscheduled disruption of flights, airport operations run smoothly.
In order to provide the required business outcomes that were discussed in the
preceding section, let's summarize the key business outcomes that are expected
from the industrial Digital Twin of a commercial aircraft:
It will help improve the current and future models of the aircraft.
It will reduce the unscheduled downtime of the aircraft – this is often
measured as Aircraft on Ground (AOG). (Please refer
to https://www.proponent.com/causes-costs-behind-grounded-aircraft/.)
It will improve fuel efficiency and the carbon footprint.
It will improve safety and reliability and reduce the variability within
operations.
In the next section, we will take a look at an example from process manufacturing.
Process manufacturing
In the Discrete manufacturing section, we looked at the example of commercial
aircraft. The aviation industry is heavily dependent on fuel, which is one of its
largest operating costs. Hence, a natural segue would be the oil and gas industry
where process manufacturing is critical.
Process manufacturing is often used in the oil and gas, chemical, semiconductor,
plastics, metal, pharmaceutical, and biotechnology industries along with consumer
packaged goods, including the food and beverage industry. Process manufacturing
uses liquid and other forms of ingredients that are often mixed according to
established recipes. Propane gas is an outcome of process manufacturing, although
the final packaged product sold might be measured in discrete number of
cylinders.
The petroleum industry consists of three major segments. They include the
following:
1. Upstream industry: This is involved in the exploration, drilling, and
production of crude oil or natural gas via oil wells.
2. Midstream industry: This is involved in the storage and transportation
of petroleum products.
7. 3. Downstream industry: This is involved in the refining and distribution
of petroleum products so that they reach end consumers via gas stations.
Figure 2.3 shows these three segments of the industry:
Figure 2.3 – The three segments of the oil and gas industry
Now, let's take a look at the applicability of the Digital Twins in the different parts
of the oil and gas industry. Oil wells use a critical piece of equipment, called a
blowout preventer (BOP). BOP is used to monitor, seal and control the oil
and gas wells, to prevent a blowout. However, a BOP is manufactured via discrete
manufacturing. In the downstream industry, the refinery is a part of process
manufacturing. A petrochemical refinery, as shown in Figure 2.4, is an important
part of the production of fuel for the transportation industry, including aircraft:
Figure 2.4 – A petroleum refinery
8. Now, let's examine how a Digital Twin can add value to a petroleum refinery. In
this context, a commonly used term is a digital refinery, which refers to the
digitization of refinery operations to provide an end-to-end view of the operations.
A few initiatives will be listed here, such as the Digital Twin of the crude
distillation unit (CDU). In 2018, TANECO and ChemTech partnered to create a
Digital Twin of the CDU, where the business objective was to optimize the oil
fractionation process. This Digital Twin used the thermodynamic model of the
production process. (Please refer
to https://www.hydrocarbonprocessing.com/news/2018/06/taneco-and-chemtech-
create-digital-twin-of-refinery.) AspenTech has also focused on the Digital Twin of
the CDU to reduce the operational risk within a plant.
Often, the petrochemical industry is challenged to control its environmental
impact due to emissions. Bharat Petroleum Corporation Ltd, in India, is working
with AspenTech on a Digital Twin for the Refinery-Wide Emission Model, to help
control the impact on the environment and stay within the regulatory guidelines
refer
when operating a refinery. (Please
to https://www.worldofchemicals.com/media/digital-twin-for-refinery-
wide- emission-and-efficiency-monitoring/4697.html.)
In the preceding section, we looked at two major industries, namely, the aviation
industry and the oil and gas industry. We studied examples of discrete
manufacturing and process manufacturing and how both provide opportunities
for the use of a Digital Twin to help drive business outcomes.
Next, we will take a look at the role of industrial Digital Twins in smart
manufacturing.
Smart manufacturing
Smart manufacturing, or a smart factory, is a broad term that is used for improving
the manufacturing industry by applying digital technologies such as Digital Twins,
the Internet of Things (IoT), and additive manufacturing (also called 3D printing).
The providers of factory automation equipment such as industrial automation
providers, namely, Siemens, Rockwell, and GE, have focused on providing
connected equipment, to facilitate smart manufacturing. However, here, we will
focus on robotics providers such as Kuka and ABB (formerly Asea Brown Boveri).
A Digital Twin provides opportunities for the providers of such manufacturing
9. robots to deliver digital services to its customers. Figure 2.5 shows this concept at
a high level:
Figure 2.5 – The Digital Twin of a manufacturing robot
When a robot is used in the assembly line process of discrete manufacturing, such
as a car in an automobile factory, a specific stage of the process can be digitized to
optimize that stage. In this case, a Digital Twin of the robot, provided by its
manufacturer (such as Kuka in Figure 2.6), can be used to map out the assembly
process digitally, leading to simulation models to optimize the performance,
including throughput and quality. This scenario allows Kuka to sell digital services
powered by the Digital Twin of the robot to the factory operator. In turn, the factory
operator can operationalize it for smart manufacturing and improve its
throughput and the quality of the physical assets it is manufacturing. Finally, this
digital data helps us to capture the birth record of the manufacturing process,
contributing to the digital thread of the asset. Such service leads to the possibility
of Robots as a Service (RaaS) for providers such as ABB or Kuka.
A Digital Twin of the human heart and the smart pacemaker is a similar
application if you think of the human body as a biological factory. Just like a
smart robot can augment and improve factory operations, a pacemaker can
improve the human body when the heart weakens. The pacemaker tries to
replicate the human heart both mechanically and electrically. A Digital Twin of
the pacemaker helps to personalize the physical asset – in this case, the
pacemaker – to the human's heart, who is using this specific pacemaker. The
signals from the pacemaker, which have been digitally captured and modeled,
help improve the delivery of care to the patient. (Please refer to
https://www.reuters.com/article/us-healthcare-medical-
10. technology-ai-insi/medtech-firms-get-personal-with-digital-twins-
idUSKCN1LG0S0.)
In this scenario, the Digital Twin provides the following possible
outcomes:
The manufacturer of the pacemaker can gain insights into their product
to improve its design over time.
The manufacturer can provide additional data and analytics services
around the product to the physicians or the patient, as the understanding
of the data and its patterns increase over time.
This allows care providers, such as physicians, to improve the monitoring
and care of the patient including when to replace the battery in the
pacemaker, which is an invasive process.
This allows the patient to self-monitor their own heart-related activities
with the help of a smartphone- or tablet-based application.
In the previous sections, we explored various scenarios, such as optimizing the
assets in discrete manufacturing, the optimization of process manufacturing, and
the smart factory. All of these areas can drive additional business outcomes by
embracing the industrial Digital Twin. Figure 2.6 summarizes, in a simple
visualization, how the Digital Twin adds value by providing insights into the
intervention of the physical asset.
This invention of the real-world object does not need to be automatic and can
include a human in the loop in the early generations of the solution:
Figure 2.6 – A feedback loop of the physical asset and Digital Twin
The information regarding the physical object is a combination of the real-time
sensor data and the historical knowledge of the asset. In a more advanced scenario,
other sources of information might include third-party data, such as weather data
or macroeconomics data, along with enterprise data from the IT systems at the
enterprise.
11. Next, let's take a look at the industrial Digital Twin of systems, that is, Supply Chain
Management (SCM).
Supply chain management
SCM connects raw material providers to manufacturers. Then, on the distribution
side, it connects the manufacturers to the business (B2B) or end user consumers
(B2C). Figure 2.7 shows the relationships between these different entities in a
supply chain process:
Figure 2.7 – The supply chain process
NOTE
Image source: https://geobrava.wordpress.com/2019/04/16/how-ai-innovation-
transforms-supply-chain-planning/
In Figure 1.5 of Chapter 1, Introduction to Digital Twin, we discussed the
relationship between the Digital Twin and the digital thread from the perspective
of a physical asset. The Digital Twin of the supply chain focuses on efficiency from
the system perspective and is independent of a specific physical asset being
produced. In Gartner's Hype Cycle for Supply Chain Strategy, published in August
12. 2020, you can see the digital supply chain twin is in the innovation trigger phase.
(Please refer to https://www.supplychainquarterly.com/articles/3877-gartner-
says- iot-technology-is-two-to-five-years-from-transformational-impact.) This
positioning of a Digital Twin for a supply chain suggests that it will reach its
"plateau of productivity" within 5+ years; however, we believe that in certain
segments of the industry, it will provide substantial business value much earlier.
In the previous sections, we looked at different scenarios in which quantifiable
business outcomes can be delivered via industrial Digital Twins. We looked at the
potential beneficiaries of the outcomes, which is key for decision-makers who are
either investing in the Digital Twin initiatives or are paying for the services and
benefits as consumers. The business outcome expected from the industrial Digital
Twin is well supported by prominent analyst firms such as Gartner. They projected
that Digital Twins will be used by approximately half of the large industrial
enterprises by 2021. This will make these organizations 10% more effective in their
business endeavors. (Please
to https://www.gartner.com/smarterwithgartner/prepare-for-the-impact-
of- digital-twins/.)
refer
In the next section, we will examine the prerequisites for the Digital Twin.
Prerequisites for the Digital Twin
To justify investment in the Digital Twin, we should frame the business problem in
such a way that the applicability of the Digital Twin is clearly understood. Let's take
the example of fuel costs in the trucking industry. Now, let's try to figure out the
prerequisites in this context:
Business problem: The high cost of fuel for the trucking fleet
Business objective: To reduce the cost of fuel without any adverse
impact on trucking operations
Proposed solution: To build a Digital Twin of the truck(s) or trucking
operations and optimize the fuel cost
Based on the simplified statement of the business problem and proposed solution
involving the Digital Twin, let's take a look at the prerequisites:
13. Model: As per Chapter 1, Introduction to Digital Twin, a physics-based or
analytics-based model would be needed to create the Digital Twin of the
truck. Either such a model should exist or should be easy to create based
on the data and knowledge of the physical asset and the operations.
Framework: The framework refers to the software or the system that can
be used to instantiate this model and apply it to the asset, which, in our
case, is the truck. This framework should be able to ingest the sensor data
and other contextual data to create the Digital Twin or keep it current.
Application: In order to achieve the business objective, such as reducing
the fuel cost for the trucking fleet, an application on top of the framework
is required to provide the actionable steps to the business user. In this
case, this application can guide the truck driver periodically or in near
real time.
In a nutshell, the model, framework, and application are the key prerequisites for
the successful adoption of a Digital Twin within the organization. There are
various viewpoints from which to observe the building blocks of the Digital Twin.
Notably, Futurithmic defined it as the three critical components of the Digital Twin.
(Please refer to https://www.futurithmic.com/2020/04/14/how-digital-
twins- driving-future-of-engineering/.) The three components are as follows:
1. The data model
2. Algorithms or analytics
3. Executive controls
In this view, the data model will also correspond to the asset model of the physical
entity. Simply put, if an asset has sensors for temperature, pressure, and vibration,
the data model or the asset model will provide information about the metadata of
the sensor data. This will help us to decide which stream of time-series data is the
value of the temperature and the corresponding engineering units (in Celsius or
Fahrenheit). The algorithm will tell us about the importance of the sensor data and
its correlation to the health of the asset. A simple example would be if the asset
temperature increases by 5 degrees in 10 minutes, along with an increase in
vibration levels, then it creates an alarm. The executive controls would refer to the
orchestration of the actions such as triggering a human action due to rapid
increase in temperature and vibration, in this case. The interventions in response,
could be to shutdown the asset thus responding to change in its behavior measured
via the sensor data namely temperature and pressure and rate of changes in these
attributes.
14. A paper, titled Digital twin requirements in the context of Industry 4.0, by Durao,
Haag, Schutzer, and Zancul has a more exhaustive list of the prerequisites or
requirements for Digital Twins based on the survey of related literature. The list is
in descending order of the number of occurrences of these prerequisites for Digital
Twins, within the surveyed literature:
Real-time data from the asset
Integration
Fidelity
Interaction
Communication
Convergence
Automatically updated
Autonomy
Connectivity
Data acquisition
Data capture
Data quality
Data security
Data warehousing
Efficiency
Expansibility
Globally available in real time
Independently expanded
Interoperability
Modularity
Process planning
Real-time location
Scalability/scalable
Stable data acquisition
Stable operation
Based on this viewpoint, the most important requirement for the Digital Twin is its
ability to handle real-time data followed by integration and fidelity. Let's analyze
these further. To generate data, the asset has to be instrumented with appropriate
sensors. These sensors can be part of the asset or retrofitted onto the surface or
within the surroundings of the asset. The real-time data requirement refers to the
ability to collect this information at a rate that makes sense for the application
served by the Digital Twin. The real-time data ensures the current view of the
asset's behavior and allows us to devise a timely intervention. For instance, an
15. aircraft sends snapshot or summary information during a flight of its critical
systems, such as the jet engine. This real-time information is processed in a timely
manner so that a critical decision can be made about the health of the aircraft for
the next flight. Hence, here, the term "real time" is relative. In this case, the time
horizon of the decision based on applying the aircraft data to its Digital Twin is
from minutes to the order of one hour. However, in the case of a pacemaker
augmenting a human heart, the time grain could be much finer.
The second requirement of data integration refers to the stitching of data from
different subsystems of the twin or, in the case of a fleet of assets, from different
assets. A commercial aircraft can have two or four engines. In most cases, an
aircraft is designed to operate safely when one engine fails mid-flight. By design,
the in-flight shutdown of one engine is supposed to be harmless to the aircraft and
not even noticeable by the passengers. In such scenarios, the data integration
between the different components of the composite asset, that is, the aircraft and
its engines, is critical. Not only do we need data integration but near real-time data.
This is so that the sister engine(s) on the aircraft can be corrected to increase the
level of thrust generated to maintain the aircraft posture with one less engine.
A rapid decline in engine oil pressure might be the leading indicator of engine
failure mid-flight. To fly safely with one engine or one less engine, the pilot might
adjust the airspeed or the altitude.
As discussed in Chapter 1, Introduction to Digital Twin, Digital Twin fidelity is often
the result of a model's sophistication. The degree of fidelity makes the industrial
Digital Twin resemble the physical asset. While a higher fidelity Digital Twin can
be used for more complex applications of the Digital Twin, it also increases the
computational complexity and, hence, the costs of managing the whole process.
Now we have a better understanding of the prerequisites and requirements of
Digital Twins. In the next two sections, we will take a look at the organizational
and technological requirements of the Digital Twin initiatives.
Next, we will take a look at the organizational and cultural factors.
Organizational factors
16. The adoption of Digital Twins by industrial giants will drive a key digital
transformation in these companies. They will become software companies of some
sort. This phenomenon can already be seen in many large industrial giants in the
last 5–7 years. Let's take a look at the examples of companies such as Honeywell,
GE, Siemens, ABB, Hitachi, Bosch, Schneider Electric, and more:
Honeywell: In July 2019, Darius Adamczyk, the CEO of Honeywell, said
that the company is moving toward "a premier software-industrial
company, with connected software sales continuing to grow at a double-
digit rate organically." It created the Honeywell Connected
Enterprise (HCE) unit to focus on such digital technologies as Industrial
IoT (IIoT) software solutions.
GE: GE decided to become a digital industrial company in the mid-2010s.
GE's vision of a digital industrial company can be viewed in an
infographic, which can be
downloaded
at https://www.ge.com/digital/sites/default/files/download_assets/What-
is-a-digital-industrial-company-infographic.pdf. GE created GE Digital
under the Chief Digital Officer (CDO) to act on this vision. GE Digital was
a software company-like business unit. The goal was to create an
industrial internet platform that can be used to create and maintain the
Digital Twins of industrial assets manufactured by GE. The same platform
could be used to apply generic twins to similar assets of other
manufacturers or allow the building of custom Digital Twins for third-
party assets used by GE's customers. An example of a third-party asset
could be a de-icing machine used by an airline that uses an aircraft
powered by GE's jet engine. GE Digital also created an extensive
ecosystem of partners around this and has played a key role in
the emergence of the Industrial Internet Consortium (IIC) in 2014.
Siemens: Along similar lines, Siemens expressed its vision for a
comprehensive Digital Twin, saying "We blur the boundaries between
industry domains by integrating the virtual and physical, hardware and
software, design and manufacturing worlds." (Please refer
to https://www.sw.siemens.com/.)
Emerging technologies such as IoT and Digital Twins require a number of changes
in the organization to take full advantage of them. We discussed some examples of
large industrial companies such as Honeywell, GE, and Siemens going through
some similar changes. We will group these into factors, as follows:
17. Digital technologies and talent: This can include technical skills in
areas such as IoT, simulation, cloud computing, and more.
Ecosystems and alliances: To take full advantage of the benefits of the
industrial Digital Twin, enterprises need to collaborate across the
ecosystem they operate in and create partnerships and alliances as
needed.
Organizational structure and culture: Companies that are agile and
can change their culture and organizational structure to align with the
new initiatives are more likely to succeed in benefitting from an
industrial Digital Twin.
Digital technologies and talent
A Digital Twin could involve a combination of emerging technologies such as IoT
platforms to manage the sensor data from connected devices or operations.
Additionally, the building of a Digital Twin might involve additional software
capabilities for the modeling and visualization of the twin. Often, the IoT platform
might consist of an IoT Core within the cloud and edge computing environment
located close to the assets or the operations. Likewise, conceptually, the Digital
Twin of the asset could reside in the cloud or on the edge, depending on the use
case. The organization will need the corresponding digital talent to identify, build,
and maintain the technology requirements.
Ecosystems and alliances
Often, organizations have to work with both the internal resources and the players
in the ecosystem. For example, earlier, we looked at an aircraft, which includes a
complex assembly of components from suppliers. For instance, Boeing might buy
the necessary jet engines from GE or Rolls-Royce. In such cases, the Digital Twin of
the entire aircraft would heavily depend on the participation of the entire
ecosystem. Historically, different players in the supply chain often competed with
each other and did not have many incentives to cooperate. This made it harder for
the industry segments to foster collaboration across their value chain.
In recent years, we have witnessed the growth of industry consortiums such as the
IIC and Digital Twin Consortium (DTC). Such organizations as IIC and DTC have
been able to bring stakeholders from the industry together to work on common
frameworks and guiding principles, to accelerate the adoption of emerging
technologies such as IoT and industrial Digital Twins.
18. Organizational structure and culture
In our opinion, an industrial Digital Twin requires close coordination between
the subject matter experts (SMEs) and the technologists. Hence, the Digital Twin
might not be wholly owned by the IT organization under the Chief Information
Officer (CIO). Likewise, it cannot be fully owned by a Line-of-Business (LoB)
leader such as the VP of Manufacturing. Instead, we have seen the emergence of
new organizations and roles in large enterprises that own such techno-functional
responsibilities. One such role is the CDO. Often, divisions such as the CDO's
division are tasked with the success of initiatives, such as an industrial Digital
Twin. In such settings, the CDO's group is then responsible for identifying the
associated technology and the development of the digital talent pool around it.
They might also foster the participation of the company in relevant consortiums
and create alliances and partnerships to accelerate the development and value
creation from Digital Twins. This process may lead to rapid adoption of emerging
technologies. In this book we explain how Digital Twin and innovation in
renewable energy come together.
The culture of innovation and experimentation is key when adopting emerging
technologies. This also requires technologists and functional SMEs to be able to
interact and work together across departmental barriers. Sometimes, this is
achieved by collocating such cross-functional professionals under one roof such as
in a Center of Excellence (CoE). In other cases, such groups might be visual but
with a higher degree of communication and collaboration.
The outcomes of organizational agility
Indeed, to validate that such organizational and cultural changes help drive value,
let's take a look at some of the quick success stories of some of the companies
discussed earlier. The Honeywell asset Digital Twins capability has been used by
Lundin in Oslo, Norway. This oil and gas company operating in the North Sea uses
Honeywell Forge's Enterprise Performance Management software to monitor its
offshore oil platform's processes and equipment with the goal of maximizing the
productivity of people, processes, and assets.
Lundin is using the Honeywell asset Digital Twins to create "energy loss" reports,
which help it to calculate CO2 emissions. The Digital Twin of the energy generation
19. process, helps Ludwin to do a full energy accounting, at the energy generation asset
level.
Now, let's take a look at the technology needs of the industrial Digital Twin.
Technological needs
In the preceding sections, we comprehensively discussed the prerequisites and
requirements around the industrial Digital Twin. Now, let's take a look at the
technological needs arising out of those requirements. Some of the recent
emerging and digital technologies will come in handy here. We will examine areas
such as the following:
1. The framework and the model: IIoT systems and cloud versus on-
premises
2. Connectivity: From the asset to the edge, the edge to the cloud, and the
cloud to the cloud
3. Data capture and storage
4. Edge computing
5. Algorithms and analytics: The Central Processing Unit
(CPU) and the Graphics Processing Unit (GPU)
6. Platforms and applications
7. Visualization: The dashboard, alerts, and Augmented
Reality (AR) or Virtual Reality (VR)
8. Insights and actions: A human in the loop and field services
9. Feedback: Product feedback, processes/operations, and training
10. Software development paradigms and low code
The preceding list is not meant to be exhaustive. Additional considerations include
the sensors required to collect the data, the placement of the sensors in the asset,
the power source, and the battery life.
The framework and the model
Here, we will take a look at the software system that is used to manage the different
aspects that are needed to understand the context of the Digital Twin and manage
the metadata of the assets, processes, or systems involved in the Digital Twin. In
general, we will find that there is a significant overlap between IIoT platforms and
20. systems for Digital Twins. Often, Digital Twin systems are part of the IIoT platform.
Some examples of commonly discussed Digital Twin systems include the following:
The Oracle IoT Digital Twin Framework (
https://docs.oracle.com/en/cloud/paas/iot-cloud/iotgs/iot-digital-twin-
framework.html)
The Azure DigitalTwin (https://azure.microsoft.com/en-
us/services/digital-twins/)
IBM Digital Twin Exchange (
https://www.ibm.com/internet-of- things/trending/digital-twin/ or
https://digitaltwinexchange.ibm.com/)
Ansys Twin Builder (
https://www.ansys.com/products/systems/ansys- twin-builder)
GE's Predix Platform (
https://www.ge.com/digital/applications/digital- twin)
Of the preceding list, GE, Microsoft, Oracle, and IBM are also well known for their
offerings around the IIoT platforms. Here, we will not go into further details about
these technical systems; however, we will do a deep dive in later chapters.
Connectivity
Connectivity can be broadly classified into the following categories:
Connectivity between the sensors and the asset: Sensors can be built into
the physical asset, retrofitted during servicing, or retrofitted aftermarket,
such as on the surface of the asset. In all of these cases, the sensor needs
to communicate with one central system either per asset or per location.
Not all sensors can have wired connectivity. Additionally, some might
have their own batteries as not all physical assets have their own power
source. Remember, an aircraft on the ground without its engines on uses
the ground power unit. Additional capabilities such as protocol
conversions might be needed here.
From assets or sensors to the gateway/edge device: The asset might be
wired to the aggregator or the gateway device at the location. In such
cases, it will need wireless communication such as Bluetooth Low
Energy (BLE) or Wi-Fi. This setup allows one gateway device (or a
minimal number of devices) to manage the connectivity for all the assets
from one location to the IoT Core/Digital Twin system, which might be in
21. the cloud or on a remote data center. These devices might also aggregate
or process the data.
From the edge/gateway to the cloud: Assuming that the IoT or the digital
system is in a public cloud or on a remote data center, the edge devices
need to connect to the core and send the data in a secure manner. In some
cases, the edge device might process or store the data to some extent. The
emergence of 5G technologies could help to centrally view or maintain
the Digital Twin of the remote field assets in one central location. Nokia
and Bosch are working toward such initiatives. The actuation of systems
in the asset, as a result of the insights from such a connected Digital Twin
system, is a possibility in the near future. Often, assets in factories or
power plants are connected via a wired network to cloud-based systems.
From cloud to cloud: In some cases, the data collection and data
processing systems could be different and might be on different clouds.
In such scenarios, we might need communication between the clouds.
Cybersecurity becomes important when data is stored off-premises.
In the previous section, we looked at the different scenarios of data connectivity
requirements. These connectivity scenarios are required to facilitate different
architecture paradigms of the Digital Twin framework.
Data capture and storage
The asset data needs to be ingested, stored, and organized in the Digital Twin
system. The most common format of data from sensors is time-series data, and
often, these are stored in Historians or time-series databases. Other forms of
unstructured data, such as video or sound files, which are often referred to as big
data, might also have to be stored in such systems. Often, Hadoop systems might fit
in here. The metadata or asset data along with the enterprise data could be stored
in relational databases. In summary, the data store requirements could be met by
a combination of technologies such as SQL, NoSQL, and Hadoop technologies.
Edge computing
Often, data originates in the sensors and assets and traverses to the core via the
edge. However, in many cases, the edge might have an important role in data
shaping or in preprocessing, analyzing, storing, and communicating the data.
Depending on the requirement, the Digital Twin system components might be
distributed between the edge system and the core system. The Digital Twins of an
22. entire fleet of geographically dispersed assets can only exist in the core or the
central location. The edge can be used to deploy the twin of a single asset on the
fleet.
Algorithms and analytics
The edge might be able to run a limited form of algorithms and analytics in near
real time for a given asset. Hence, in a generic Digital Twin system, the algorithms
and analytics models at the asset level should, ideally, be written once and be
allowed to be deployed either at the edge or within the core. Depending on the
volume, nature, and speed of computation required in the edge, it might use a CPU
and a GPU. A typical example of the necessity of a GPU is when dealing with video
data and the need for processing it on the edge. Even in the core, the system
running the IoT or Digital Twin platform might use a combination of virtual
machines, bare-metal servers, or high-performance computing (HPC), which is
often equipped with GPUs.
Additionally, GPUs are deployed when learning with complex Artificial
Intelligence (AI) algorithms, including deep learning. More details regarding this
will be covered in later chapters of this book.
Platforms and applications
The generic capabilities of the building blocks of the Digital Twin system are often
referred to as the platform. The platform prevents the rebuilding of the same
collection of generic capabilities over and over again. GE's Predix Platform or the
Microsoft Azure platform could fall into that category for IoT platforms. These
serve a wide variety of use cases across multiple industries. However, the
applications built on top of these platforms could serve a very specific purpose.
The same platform might be able to manage the Digital Twin of a pacemaker,
aircraft, or automobile. However, the application and objectives of these
applications could be very different. Sometimes, a layered approach is taken
where, on top of the IoT platform, an industry-specific or functionality-specific
(asset monitoring versus manufacturing) application family might be developed,
such as for the aviation or healthcare industry. This layer might try to generalize
the common application needs that are often seen in that industry along with the
industry-specific security and compliance requirements. Then, another layer of the
23. application might be for a specific set of assets such as the jet engine in the aviation
industry or a pacemaker in the healthcare/medical devices segment.
Visualization
The platform might provide basic visualization capabilities such as the
visualization of the Digital Twin, an asset monitoring dashboard with alerting
capabilities, or a fleet view of the asset twins. The application users can reuse or
customize these capabilities or build their own Digital Twin visualizations with
specific outcomes in mind. In more sophisticated solutions, AR or VR might be used
for enhanced interaction with the asset Digital Twins.
Insights and actions – a human in the loop and field
services
The broad set of capabilities that allow human operators to gain insights into the
assets via the Digital Twin, and take appropriate actions, would fall in this category.
A field service professional might use AR/VR to augment their interaction with the
asset in the field setting, such as when dealing with unscheduled maintenance.
Feedback – product feedback, processes/operations, and
training
The overall solution for Digital Twins must provide a feedback loop. This is so that
insights gained from the twins can be captured via knowledge management
systems. The product designer and engineers should be able to mine that
information to improve future products or, in the case of software-defined
products (such as a Tesla car), provide the future revisions of the product in the
field. Tesla uses over-the-air (OTA) updates to its software in the car to improve
the current product over its lifetime; please refer to Figure 2.8:
24. Figure 2.8 – OTA updates to the car
Software development paradigms and low code
Finally, the software development framework should be robust and functionally
rich to allow agile and rapid development. A few commonly used terms include the
following:
A cloud-native or microservices framework
A low-code development framework
A Software Development Kit (SDK) and Application
Programming
Interface (API) to allow the collaboration between different teams and
companies in the entire ecosystem.
In the preceding section, we looked at the technological requirements for
industrial Digital Twin systems. In future chapters, we will dive deeper into some
of these aspects, as we begin to decide how to select these technological
components for a specific problem that can be solved with the Digital Twin.
Summary
In this chapter, we looked at the planning process for the industrial Digital Twin.
Additionally, we looked at the key criteria based on the nature of the Digital Twin
application and the desired expected outcomes. We examined the technical and
non-technical prerequisites for the success of the Digital Twin in an enterprise. We
looked at examples from different industries in order to apply this decision
process, such as in the aviation industry, the oil and gas industry, and the medical
devices industry. Hence, the general framework of the Digital Twin, developed in
this chapter is agnostic of the industry domain, it is applied to. We want to consider
and look at the business justification for the industrial Digital Twin.
Chapter 1, Introduction to Digital Twin and Chapter 2, Planning Your Digital
Twin wrap up the part of the book where we focused on the "what" and "why" of
the industrial Digital Twin. Chapter 1 covered the background and definition of the
Digital Twin. Then, this chapter built on that information and set the stage for the
evaluation and assessment of the prerequisites of the Digital Twin.
25. In Section 2 of this book, we will focus on identifying, planning, and building the
Digital Twin. In Chapter 3, Identifying the First Digital Twin, we will discuss how to
go about evaluating the right candidates for the Digital Twin. We will explore the
different roles and responsibilities in this context. This evaluation process will help
to narrow down the finalist for building a prototype of an industrial Digital Twin
in this book.
Questions
Here is a list of questions to test your understanding of this chapter:
1. What are some of the expected business outcomes of an industrial Digital
Twin?
2. Are Digital Twins applicable to the process manufacturing industry?
3. What organizational factors can contribute to the success of the digital
twin initiative?
4. What is the role of cloud computing in a Digital Twin?
