The Top 5 Mistakes to Avoid When Building Your Digital Twin for Oil & Gas Asset Management

The landscape of asset management technology is exploding with the advent of ‘asset performance 4.0,’ and the continued technological advancements in IoT, big data, autonomous task execution, computer vision and augmented reality present both abundant opportunities and unrelenting challenges for operations and asset management teams. The biggest driver of all the above are digital twins – the virtual representation of a physical piece of equipment, system or entire production facility. 

Often, digital twins are synchronized with real-time data from either the well path or from key production equipment like gas compression systems and use machine learning to aid in decision making. Building and maintaining digital twins are no longer a nice-to-have, they are table stakes for high-performing asset management teams.

While the proliferation of digital twins has made them more commonplace, many teams still struggle in both their design and implementation, and fall short of their intended benefits – efficiency, cost savings and enhanced safety. Here are the five most common mistakes companies make when designing and building their digital twins:

1. Unclear objectives and goals. If the scope and the objectives of the MVP and long-term digital twin are not clearly defined at the onset, it can be difficult to determine what data and models to include, leading to a digital twin that is either too broad or too narrow in scope to be useful. Before building it out, consider and prioritize the most important use cases for the new digital twin.

This could include improved operational efficiency, in which case the key requirements is the ability to simulate and optimize processes, or the proactive identification of inefficiencies. If the twin should be used for enhancing its reliability and availability, then real-time condition monitoring and asset performance capabilities must be well documented. If the objective is to improve technician training and safety, then high-fidelity 3D rendering and integration with interactive augmented reality applications, like Unreal of Unity, will be necessary.

2. Missing necessary data. You must accurately represent the physical systems, processes or objects defined in the twin’s objectives. A digital twin is only as good as the data and models that are used to create it, and if there’s inaccuracies or the data is incomplete, then the twin will provide invalid insights and put any decisions that leverage the twin’s insights at risk.

Beyond exhaustively identifying sensors and other monitoring equipment, it is important to use the design and engineering documents. Blueprints and schematics can provide important information on the structure and the function of the physical object or system, data on the materials, dimensions, and other physical characteristics. Don’t underutilize historical data, as digital twins need data on past operating conditions, production rates, and maintenance and repair data. Not only is it important to identify the types of data necessary, but also the technology you need to collect, store and analyze this data. For example, what data acquisition and managements systems do you plan to use, and will they collect data in regular intervals or reactively for specific events or conditions?

3. Insufficient integration. A lack of integration can cause implementations and designs to fall short. There’s significant effort to fully integrate a digital twin with SCADA systems, automation systems and enterprise resource planning platforms. Digital twins that drive value are well-integrated with not only the critical data sources, but also the end-user outputs.

• If the primary objective of your digital twin is for predictive maintenance in addition to all your sensor data and analytics/computer vision platforms, you must be able to deliver the outputs to your CMMS and work management applications.
• If the twin is for real-time condition monitoring and failure recognition, then it must integrate with your current operations dashboards and asset performance platforms.

Once all necessary integrations are identified, teams should focus on scalability and flexibility. The digital twin must be able to adapt and evolve as the production facility grows and changes over time. This requires scalable and flexible architecture that can accommodate both new integrations, as well as updates to the overall data model as new equipment and processes are added.

4. Unintuitive and complex user experience. The digital twin must also be intuitive and easy to use for the end users, it must support effective collaboration and decision making among your various operations, engineering, maintenance and reliability teams. Don’t assume that the software and technology infrastructure selected during the design phase will be user-friendly. Consider the end user’s expertise with digital twins, as well as the goals and objectives for this user when interacting with the technology. In many cases, it may be more cost effective to build a custom user interface layer for your digital twin that provides just the information your users need in a seamless and efficient manner.

5. No plan for security and privacy. Don’t leave the security and privacy of your digital twin as an afterthought, as encryption and access should be thought of as core design elements. Digital twins must be protected against cyber threats and other forms of unauthorized access to safeguard sensitive operation data and protect the integrity of the production facility. This may require the use of advanced encryption and other security measures as well as regular monitoring and testing to identify and address potential vulnerabilities. Understanding and defining the security requirements needed to ensure the model is not vulnerable to unauthorized access or manipulation can save the team significant time and effort down the road. Some of the worst news a project team can receive is that their designed architecture, data model and integrations don’t pass security review and need to be re-worked.

Once teams have well-defined goals and objectives, a full list of data sources and planned integrations, and considerations for user interactivity and security, it’s time to implement and build. Long-term success is dependent upon governing the accuracy and reliability of your data, which should involve regularly checking and validating the data as well as implementing measures to prevent errors or inconsistencies. In addition, a strong sustainment process will be needed to ensure the digital twin is kept up to date and, as modifications and improvements are made to your productions systems and equipment, those changes are accurately captured in your twin. Digital twins by nature can be complex systems. It is important to have a clear understanding of the technology strategy behind the different platforms, how they are maintained and updated, and how to sustain both the flexibility and scalability of your design.