10 best practices for digital twins in the energy sector

Best practice guide

May 17, 2023

Mindsharing at Unity’s Energy Day

May 17, 2023 – Unity’s Energy Day, Houston

This event, hosted by Unity and Capgemini, brought together energy industry leaders to share the challenges they’re facing related to building and scaling digital twin initiatives, and to brainstorm ways that those challenges can be addressed.

The discussions spanned the full scope of the subject, from defining digital twins and preparing data to build one to their multiple use cases and efficiently growing a digital twin program.

Here we summarize the 10 best practices identified by the group.

Unity and Capgemini

A global alliance partnership to help organizations leverage the business value of immersive and metaverse experiences. Together, we play a key role in designing and scaling tailored metaverse and digital twin solutions for clients, accelerating the convergence of the physical and digital worlds.

Executive summary

This event, cohosted by Unity and Capgemini, was an example of digital twin experts across the energy sector collaborating and sharing ideas about their programs.

Contributors from organizations such as SLB, bp, Exxon, and others came together to discuss the challenges they’ve encountered and to share their top 10 tips on setting up digital twins for success in order to drive the energy transition.

The best practices discussed during the event cover these three focus areas:

The importance of a comprehensive digital twin business case tied to an end-user need
Participants talked through the crucial step of ensuring that your digital twin plan is built around solving a problem that already exists for a group of users. In addition, the group emphasized having carefully considered objectives relating to measuring success, be that ROI, ESG goals, or others.
Driving for consistent and repeatable optimization methods for complex data
Our contributors shared the pain of collating data from different sources, including CAD, BIM, IoT, 2D data, and even non-digital sources. All these inputs are critical for an informative and effective digital twin. The group also discussed the need to prepare data in a way that optimizes its application in the digital twin.
Ensuring future development and scalability of the digital twin
The pain of ‘black box’ solutions was discussed, along with the need to plan ahead with your digital twin to ensure that it is not only interoperable with other systems/digital twins, but also that you are in a position to maintain it and scale it up.

As companies focus on achieving sustainability targets, increasing productivity and operations efficiency, software spend (with a focus on digital twins) is projected to double in the next five years. According to this McKinsey report, renewable energy will account for 80% to 90% of power generation globally by 2050. Many companies are leveraging digital twin technologies to improve collaboration and decision-making as they manage this transition.

The energy sector is at the forefront of digital twin adoption, with pioneers such as SLB, bp, Enbridge, and Exxon paving the way. From slb’s virtual twin of a gas processing facility to bp’s production optimization digital twin to Enbridge’s system optimization program to Exxon’s XR training program, digital twins are already proving successful for multiple use cases. The path forward will be built on addressing the challenges raised at this event to ensure a connected and all-encompassing digital twin program. Keeping an eye on incorporating new technologies such as AI will also be important. As one contributor said, “We can go fast alone, or far together." In this rapidly developing area, collaboration is the key.

Focus area 1: The business case

“The industrial metaverse, it's where you could get all of your IoT data, all of your systems from ERP, SAP, all the different pieces together. Putting that into a visualization format, it doesn't matter if it's 3D or 2D. The point is, you're getting all that data together and focusing on the collaborative impact of it.”

Organizations often jump into new technology, such as digital twins, without a clear strategy in place. This can quickly lead to problematic situations, such as siloed ‘black box’ digital twins that aren’t interoperable, lack of clarity on the objectives, and/or solutions that are not truly offering real improvements.

Our contributors had a few words of advice on this topic:

Tip #1 – Solve for the user
The digital twin must work for the end user. Start with a small target group (e.g., a facilities team), understand their challenges, and build a solution that works for them. As you extend the digital twin, always return to this principle to ensure you avoid the potential pitfall of investing in a digital twin program that is underutilized because it doesn’t meet the end users’ needs.

Tip #2 – Set realistic expectations
A frequently asked question, especially by the C-suite, is about the return on investment (ROI) of a digital twin. Be clear from the outset what it is that you would like to achieve and how you will measure that. Our group shared their pain points around this, with a specific discussion around cost avoidance as a measurement factor:

“If you have a piece of equipment that fails – that causes a catastrophic explosion or an oil leak, or even something as simple as a biofuels leak or something like that – how do you measure that? It will cost hundreds of thousands or millions of dollars to prepare, clean up, etc., when preventative maintenance could’ve cost hundreds.”

The consensus is that highlighting the potential economic gain will have the strongest resonance with decision makers.

For example, predictive monitoring and maintenance alerts can minimize downtime by providing staff with specific data about the potential issue, which components require attention, etc. This value can be presented to decision makers in a cost-per-day amount, comparing the total cost of previous asset downtime timelines against the scenario of the digital twin providing real-time, accurate, preemptive instructions to minimize or possibly eliminate downtime.

Tip #3 – Determine environment, sustainability, and governance (ESG) targets
Similarly, the key here is to be clear about what the expectations for the digital twin are. For example, how are you anticipating it will impact your carbon footprint, and how will you measure that?

Tip #4 – Allow for in-house upskilling and consider cross-industry talent acquisition
Plan for building, maintaining, and scaling your digital twin by ensuring you have the appropriate skills in-house. You want to avoid the costly solution of contracting in third parties whenever you want to make changes, fix errors, or upscale your digital twin.

If you need to onboard additional resources, don’t rule out the option of investigating outside your normal talent acquisition pool; developers working in other industries like gaming can easily transfer their existing skills.

“The chance that any one person is going to have what it takes to be that single point of digital twin knowledge is very rare. So when you look at the competency of the digital twin core team, really what they are is integrators. They're the group that draws those different people together. They don't have to know in depth how to do every single piece of the puzzle, but they have to know how to talk to the different parties and bring them together.”

Tip #5 – Educate about digital twin maturity levels
Educating stakeholders about the maturity levels of digital twins (see diagram) ensures that expectations are clear about what the application will and will not be capable of.

Most often, the initial implementation is a Level 1 Virtual Twin, which is defined as a physically accurate and spatially aware 3D digital representation of an asset or facility. Twins at this level are primarily a visualization of complex data (such as BIM, CAD, or GIS), and are best suited to enabling collaboration between cross-functional teams.

The levels range from 1 to 5 with increasing capabilities as you go up the scale, including functionality, intelligence, and interconnectivity.

A chart of the Digital Twin Maturity model — The digital twin maturity model describes the different levels of complexity that a digital twin can fulfill.

Focus area 2: Data preparation

“In each aspect of a digital twin, whether it's simulation, data, aggregation, or visualization, for us to build scalable digital twins, and maintain them for the lifecycle, we have to have a perfect collaboration of all those things together. And it all starts with data. Every bit of it is data.”

The group shared a general consensus about the headaches caused by data wrangling, and yet without addressing this essential step, creating a digital twin is impossible. But what are the specific issues that our contributors are facing?

Tip #6 – Getting data sources ready for digital twins
The datasets available vary widely, from decades-old engineering drawings, to CAD files of various types and complexity, to there being no data available in some cases. In the latter examples, data is created through a scanning method such as Lidar or photogrammetry.

The challenge lies in bringing these various data sources together and ensuring that they are compatible not only with your chosen digital twin platform but also with each other.

This is where a data preparation and optimization tool like Unity Pixyz comes into play. Adopting the right tool at this stage ensures data sources are optimized for your digital twin. This can eliminate the common obstacle of complex, disparate data slowing down deployment and ultimately the performance of a digital twin.

Tip #7 – Make the most of existing assets
Ensure that you seek out the data you already have and make the most of it. For example, if you have access to BIM data for a facility, make use of that data in your digital twin. This is a more efficient approach than seeking to duplicate efforts, for example by commissioning photogrammetry to capture information about that same building.

A screen rendering of a chemical plant design — One of the functions of an operational digital twin is to monitor equipment status and provide alerts for maintenance requirements.

Focus area 3: Interoperability and scaling

“We focused on componentizing – on making sure that we can put all the puzzle pieces together to create a digital twin. So, for example, if you go to a digital twin of a rig or refiner, you can put parts X, Y, and Z together.”

Tip #8 – Avoid black box solutions
The issue of ‘black box’ solutions was raised. In these situations, a digital twin has been created – often by a third party outside of the organization using software with limited customization capabilities – and handed over to the organization. It operates just fine by itself. However, this poses several challenges, including a lack of access to the core platform to make changes and updates. The nature of black box solutions also makes interaction between digital twins within the organization problematic, as they may not have common structures, or may not easily interface with any other data sources.

Tip #9 – How to scale effectively
Relating to the point around interoperability, an eye on the future is critical. The group concurred that any digital twin program should allow for the potential of scaling at a future date, even if no such plans exist for the immediate future. This could mean scaling an existing twin to a higher level of maturity, expanding to incorporate additional data sources, or connecting to other digital twins. In order to ensure efficient and effective scaling, this is something that should be taken into consideration from the start of the program to avoid future pain and additional costs.

Tip #10 – Share your vision with program partners
To ensure your digital twin program has scalability built in from day one, it is best to share your long-term vision with your third-party partners. When all parties are aware of both the short- and long-term goals of your program, you can develop digital twins that are truly interoperable, thereby avoiding unscalable black box solutions.

A digital rendering of a large plant from the outside — An operational digital twin is an accurate representation of a physical site, which allows for analysis using current and accurate data.

What’s next?

“What it is about is really building huge datasets that allow machines to understand what they see…. It’s linking technology together.”

Future technologies

→ Artificial intelligence (AI) for digital twins

With the rapid development of AI, there are clear applications for digital twins. As can be seen in the Maturity Model, AI and machine learning play an increasingly important role from level 2 onwards.

In a digital twin context, AI is most powerful when deployed to build, analyze, and interpret very large datasets. AI tools analyze vast amounts of data collected by the digital twin and generate predictions about future performance. These analytics can identify areas where improvements are needed, suggest alternative strategies, or provide recommendations on how best to manage different scenarios.

In the most mature digital twins, AI tools and machine learning will enable correct predictive actions to be taken without the need for human intervention, thus minimizing errors, streamlining costs, and maximizing productivity.

Ultimately, the more data you supply into a Level 2+ digital twin, the greater its autonomy in data-driven decision making. An AI-empowered digital twin, fed sufficient data, can be trained to understand desired outcomes. This is the foundation for enabling an autonomous digital twin (Level 5 on the maturity scale), which is able to predict and take action on the most effective decisions.

Our contributors are already investigating this technology:

“Looking to the future, integrating with our geospatial system, so that we can have twins and geospatial and geolocation all playing together in the same space. This goes back to that whole data principle that we have an integration, that system of systems, and making sure that our systems are an entire landscape.”

“We've been looking at no-code and low-code solutions. How can we put the building blocks of the digital twin together in a way that requires a minimal amount of customization? You don't want to build a new solution for every single asset. You want to be able to customize and change it. Building for future extensibility, composability, and customization.”

“If your metadata is correct, you can classify things. If you see a pylon, and you have a set of stairs, or walkway, or scaffolding, those things shouldn't intersect. The model should know that. If you have metadata in the digital twin, you can plug AI into it. You can apply engineering knowledge, and the digital twin is able to determine these things just by knowing what's intersecting each other and infer what impact that has.”

A digital rendering of an oil platform at sea — A digital twin of an operational site can serve many functions including remote training, predictive maintenance, monitoring, and reporting.

Get started with digital twins

Create it yourself

Get the creation tools and enterprise-level support you need to transform your CAD and 3D data into immersive digital twin applications available on any device, anywhere.

Buy now

Learn more

Build it with our experts

Looking for expert assistance to get started? Check out our Operations Starter Package and get help from our team of world-class designers, developers, and energy industry specialists to help you start your digital twin program.

Not sure where to start? Contact our team to find the solution that’s right for you.