Demystifying digital twins

Demystifying digital twins

All models are wrong, but some are useful

The 'digital twin' is now a recognised core component of the Industry 4.0 journey, helping organisations understand their complex processes, resources and data to provide insight into their business and help optimise their operations.

But while 'digital twins' can sometimes mean different things to different people, we at BMT define them as "precise, virtual copies of machines or systems". These digital models receive regular data input from the relevant physical system, thus providing intelligent support for operational decisions. Twins are, first and foremost, learning systems, driven by data that’s collected from sensors in real time. This means that sets of complex digital models can adapt to mirror every element of a product, process or service. The efficiencies of an asset digital twins’ product life-cycle management can typically offer powerful benefits in four areas:

1. Operations optimisation
   

   Using variables like weather, fleet size, energy costs or performance factors, models are triggered to run hundreds or thousands of "what-if" simulations to evaluate readiness or make necessary adjustments to current system set-points. This enables system operations to be optimised or controlled during operation to mitigate risk, reduce cost or gain any number of system efficiencies.

2. Predictive maintenance

   In industry 4.0 applications, models can determine the remaining useful life of an item of equipment and advise operators on the best time to service or replace it.

3. Anomaly detection

   The model runs in parallel to the relevant real asset and immediately flags any operational behaviour that deviates from expected (simulated) behaviour. For example, a petroleum company may stream sensor data from offshore oil rigs that operate continuously and the digital twin model will look for anomalies in the operational behaviour to help avoid catastrophic damage.

4. Fault isolation

   Anomalies can trigger a battery of simulations to isolate the fault and identify its root cause, so that engineers, or the system, can take appropriate action.

A good deal of progress has been made across several sectors since the original inception of the twin concept in 2002, which initially began by gaining momentum within high-value product manufacturing industries, such as mining, automotive and aerospace.

Typically, specific elements of a twin were developed to help achieve a specific aim for a specific system or sub-system. Since then, the digital twin concept has proliferated, taking on many interpretations and variations and has found increasing relevance across many other sectors.

However, there’s still much to be done before the benefits of twins can be realised across industry on a wider scale. For instance, there are no standards in place for such tools, and many challenges yet remain around open data capture, data storage, and performing effective analysis to support decisions with confidence.

Indeed, a standard definition of a digital twin has not yet been established and there are many variations in the accepted scope of a twin across sectors. To add further complexity, what began as a niche engineering model is increasingly being expanded in scope and aligned to wider digital transformation initiatives.

The complexity of twins

To provide benefits across an entire enterprise, a digital twin must be more than a single entity. Fully-featured digital twins will be realised as a family of twins, each of which matures and grows in depth and value through its life.

The models used to support asset performance may differ from those used in design and development, but there will be clear data linkages across all of these model sets that need to be managed. Twins will both consume and create high volumes of data throughout their lives.

Several coherent twins will be created in the design phase, from basic 3D geometric and design information through to complex high order simulation models and wider design information management. This "Design Twin" set will mature through the build phase to create a "Manufacturing Twin" - a key enabler for the future digital shipyard.

This evolution will continue through to the in-service phase, where sensor inputs from the physical asset will help realise the final "Digital Twin." Further models will need to be created at this point, typically fast surrogate models that include a measure of their own uncertainty to effectively provide predictive and prognostic outputs in response to real-time input data.

The creation of a twin is therefore a complex process and relies upon a continuous digital thread throughout the lifecycle.

Although IoT is a key enabling technology for digital twins, twins are much more than a simple IoT implementation, which may integrate an intelligent sensor with a data platform.

Twins bring together on-platform sensing with other data sources (e.g. environmental data, historical maintenance data) and match this data with a set of models, applying analytics and machine learning to enable predictive and prognostic insights. These insights provide effective decision support in an intelligent asset management context.

Due to the complexity and volume of data, the effective visualisation of twin outputs is critical to enable usability and layered consumption of data.

Digital twins: Why now?

The concept of a digital twin isn’t a new one - the term was first used in 2002 and the original information concepts date back to the 1980s. However, there are several enabling factors that are coming together to accelerate the implementation of twins.

The first of these is cheap, high bandwidth sensing availability - IOT is a key enabler for twins. The second is ‘big data’, there is now an abundance of cheap data storage and associated advances in data management and mining.

The third is the increased accessibility of the high-performance computing (HPC) required to manage the computing burden of a functioning twin.

Finally, advances in integrated design and analysis environments are creating key digital threads at the start of the lifecycle, enabling the later creation of a twin.

These design digital threads link 3D information and multi-physics with integrated data management to create a configuration managed environment that can be maintained and augmented through production.

Challenges to adoption

Within the defence maritime enterprise, many of the challenges to adoption do not just come from the technical domain, but also as products of the commercial and supply chain environment.

Defence maritime has traditionally contracted separately for design, build and maintenance contracts. These contracts are focused on providing physical equipment rather than data and models and this approach typically requires the involvement of multiple prime contractors for each asset through life, along with a correspondingly complex supply chain network.

There is currently no mechanism to allow data and digital models to flow freely and securely through this enterprise whilst protecting Original Equipment Manufacturer’s intellectual property. It is likely that a future digital twin for defence maritime will span multiple solution platforms, multiple companies, and multiple solution platform vendors.

BMT’s focus

BMT sees digital twins as a key element of industry 4.0, having seen the benefits and savings they are bringing to our customers. BMT’s Whole-of-Mine Digital Twin is just one of many bespoke examples.

Built on our Remote Operations Automation and Robotics (ROAR) software platform and using lidar mapping capabilities and expertise in 3D algorithmic decision making, our solution integrates precise terrain data, machine data, health data and production metrics into a single 3D immersive and intuitive interface.

This allows production teams and planners to measure, monitor, report and optimise mining operations in real-time, reducing costs and risks.

BMT continues to develop and expand our digital twin technology into other submarkets, with ongoing projects that are delivering Autonomous TLO (Train Load Out) Digital Twins, Hopper Surface Mapping and Wagon Surface Mapping solutions for train unloading rail operations, for example.

Other technologies are in development for Stockyard Mapping and Ship Hold Surface Mapping for bulk materials handling customers in ports.

Future maritime defence digital twins

To support future maritime defence digital twins, we’re taking a multi-threaded approach through a range of internal strategic initiatives and R&D activities, as follows:

1. Digital threads in design, enabling the creation of the twin We’re mapping out our design toolset of the future for ship design, creating an integrated digital design twin with advanced integrated simulation and analysis capabilities that allow the left shift of risk from physical testing back into the design phase.
2. Integrity, security End-to-end (sensor to decision) secure data transmission, management and verification; digital twin cyber threat detection, protection, response and recovery. A digital twin ship or fleet is susceptible to threats that are similar to those seen by physical ships.
3. Interoperability, supply chain engagement Federated twin high-level architectures and APIs that enable prime contractors and OEMS to support a digital ship with ‘digital equipment’ whilst protecting I.P. and ensuring the interoperability of ‘digital assets’.
4. Digital twin return on investment Research into future intelligent asset management - a digital twin creates value when embedded into a holistic intelligent asset management system, supporting decision-making and business model optimisation at portfolio and organisational levels.

These solutions not only highlight our abilities and expertise in developing bespoke customer solutions, but are also are testament to our commitment to Digital Transformation programmes in the industries we serve.