When troubleshooting problems in a process plant, access to trusted data and information is vital. Too often, however, we don’t know when data can be trusted. A digital twin can help establish a solid foundation for trusted data. It facilitates collaboration between departments, bringing together all sources of engineering, maintenance and operations data. By providing visibility to analytics and insights, a digital twin aids operators and maintenance staff in making fast and accurate decisions to improve the production process, asset integrity, safety, reliability and overall efficiency.
Experience tells us that when data are suspect, out-of-date or inaccurate, engineers spend 30–50% of their time searching for or validating data before accepting the data as accurate and trustworthy enough for critical decisions such as those affecting safety, compliance and risk.
Troubleshooting plant problems becomes a challenge when the data needed for decision-making reside in disconnected, inaccessible siloed systems or are out-of-date — these are so-called “dark data.” Most plants have multiple sources of data, including engineering models handed over by the engineering/procurement/construction (EPC) firm that designed and built the plant and the numerous information and operational systems used by specialists — and usually limit access to specific systems to particular groups or departments. The inability to access data from these systems as well of those of supply chain partners, customers and other stakeholders often is a great source of pain.
Do You Have A Dark Data Problem?
Ask yourself the following questions to determine if your plant has a dark data problem.
1. Do I know what assets I have and where they are located — and can I prove it? Finding basic specifications about an asset, including where it’s located, can be the most-frustrating problem because it seems simple. However, if your plant lacks a central engineering data, document and information repository, keeping up-to-date with all assets is quite difficult.
A digital twin provides a visual, geo-referenced and federated information index that supports easy accessibility and maintenance for all relevant engineering, asset operation and maintenance information, such as tags, schematics, specifications, data sheets and machine-type settings. By using a digital twin, plant teams can visualize information from a wide range of sources and contextualize that information within a digitalized layout of the plant (Figure 1).
Figure 1. Digital twin enables not only viewing information from a wide variety of sources but also contextualizing that information.
Within a digital twin, an operations team can navigate through assets, systems or the entire plant with graphical data, picture-in-picture or interconnected tabular data. The team no longer must rely on a mental picture of an asset to understand location and proximity. This digital visualization supports better decisions and clearer communication of the plant layout.
2. How do I know what assets should be there? Plant engineering documents, such as piping and instrumentation diagrams (P&IDs), isometrics, schematics and 3D models supplied by the EPC company, form the design bases for the plant. Today, owners commissioning new plants should mandate the handover of a digital twin of the plant as part of the EPC contract. Progressive EPCs are ramping up plants 80–90% faster with digital delivery. Complete digital project delivery workflows, including an aggregated view of all related asset information, eliminate paper deliverables, improve procurement and logistics, enhance quality of design, minimize construction rework, reduce ramp up time and boost performance on commissioned systems. For example, Hatch used a digital twin to decrease production ramp-up time to one week from six months on the delivery of a processing plant in the Democratic Republic of the Congo in 2019, realizing significant savings on the project and operation efficiencies.
A digital twin allows engineers to review design bases, 3D models and associated engineering information delivered with the plant. This solid data foundation enables engineering, operations and maintenance to collaborate and share data. Digital twins shine the light on the dark data once only accessible as PDF documents.
3. Have my assets changed? Alterations often result in dark data. Over time, plant documents and data go missing, become obsolete or get stored in myriad formats in multiple locations. When a troubleshooting investigation begins, accessible data might not reflect the asset’s changes. This can pose great safety risks to the operation.
It’s not that uncommon for an organization to upgrade, modify or add to an asset without properly updating the engineering information models. Even when changes to the documents and models are made, they’re often hand drawn, marked up, confusing and delayed. Frequently, engineering documentation or models take six months or longer to get updated despite regulations requiring facilities to keep such documentation up-to-date. This noncompliance poses the risk of a significant fine. Worse yet, use of inaccurate documents could lead to an incident resulting in damage or injury.
Figure 2. A digital twin enables a plant to keep up with all the data it’s generating.
A digital twin is designed to be easily updated. Moreover, it maintains a timeline of alterations to the asset — capturing how and when the engineering data changed as well as who made the modification. By viewing change history graphically or in a tabular format, you can see the impact of any alteration on ongoing operations.
4. What is the current performance of the asset and how does it compare across the plant, fleet or industry? Equipment performance requirements often aren’t documented or accessible — and can change over time. Engineering staff frequently don’t have access to the asset-performance-monitoring system. The amount of raw data being generated in plants from Industrial Internet of Things (IIoT) sources is growing exponentially (Figure 2); many organizations can’t keep up. These sources of dark data make current asset status and performance less certain.
A digital twin provides a federated source of data for all related systems, including those for operational and asset performance. These views incorporate the ability to compare assets within a fleet, visibility of real-time equipment health and condition degradation trends. Any out-of-tolerance asset can appear as a flashing alarm on the 3D model of the plant or health dashboard. A digital twin also can provide a history of how often and why any asset, system or component failed.
5. When do I need to act and what should I do? Engineers might lack access to reliability and integrity management programs that can enhance their understanding of the likelihood, possible causes and consequences of an asset failure. They also may face hurdles in accessing maintenance records or risk assessments documented and maintained by a different department. Such data silos can obscure how to respond.
A digital twin gives a unified view of asset health to ensure plant production is predictable, safe, compliant and efficient. This single federated source of truth ensures everyone in the organization has accurate information for audits and course-corrections as well as a complete and consistent view of plant data for collaboration and problem-solving.
Engineers quickly can identify potential operational problems, troubleshoot and determine solutions with confidence the data are trustworthy (Figure 3). Operators can review as-operated and historical data to understand all field changes and engineering decisions to gain a big-picture view for fixing problems and optimizing production. Maintenance and reliability teams can monitor assets remotely, manage equipment health and prepare maintenance plans as well as easily spot trends and bad actors before traveling to the asset. A digital twin visualizes the history of failures, secondary damage, work performed and parts installed and removed. All data from the enterprise-asset-management and asset-performance-management systems are visible in the context of a 3D model of the plant, making working with the data intuitive and easy.
Figure 3. Plant staff get immediate access to trusted engineering and asset information.
Building A Digital Twin
If you have plant data, you can build a digital twin. You don’t need a 3D model. However, if you want the visualization aspect and lack a 3D model, you easily can develop one. Capture images of plant equipment using digital photography or laser scanning, then use reality modeling software, such as ContextCapture by Bentley, to create an accurate as-operated reality mesh of the plant. The software can identify components and systems with machine learning. The realistic representation also is geo-referenced. If you use 3D models as your starting point, ensure that the digital twin software you are using accepts data from multiple sources. For example, with PlantSight (see sidebar), you can aggregate models from multiple sources such as AVEVA (PDMS/E3D/P&ID) and Hexagon (Smart3D, SmartPlant P&ID).
Next, combine engineering information from multiple sources including spreadsheets, schematics, and piping and instrumentation diagrams. Lastly, connect and align your operational, asset-performance and maintenance systems. The digital twin will enable all departments to view, for instance, the asset registry in SAP, and will consolidate, analyze and make visible SCADA and historian data, e.g., from OSIsoft PI. This federated approach allows the best possible use of existing sources of data by indexing and surfacing rather than replacing or duplicating.
Achieving Success
The path to a digital twin will depend on your systems and processes, business priorities, market needs and overall goals. Regardless, developing and maintaining a good digital twin poses challenges. A number of under-appreciated issues and common mistakes can complicate efforts. So, let’s look at some challenges you might encounter and ways to troubleshoot them.
• You are building your digital twin from multiple sources and suspect data discrepancies due to out-of-synch asset tags, asset changes, modifications, revamps or additions over time.
What to do: Use the conflict resolution capabilities of the digital twin software to help identify, troubleshoot, compare and resolve the data discrepancies as you build the digital twin.
• Incompatible engineering data formats result in data in separate siloed sources.
What to do: Ensure your digital twin technology is open and interoperable. It should accept diverse formats from multiple sources. For example, the foundation of Bentley’s digital twin technology, iTwin, incorporates different types of data repositories — drawings, specifications, documents, analytical models, photos, reality meshes, IIoT feeds, and enterprise resource and enterprise asset management data — into a living digital twin. iModel.js is a good (and free) open source library for digital twins that’s available at www.imodeljs.org/.
• The digital twin you created uncovers problems related to the control of underlying engineering information.
What to do: Remember that visualization alone doesn’t equal a good digital twin. Use asset-lifecycle information management best practices to ensure the accuracy and up-to-dateness of foundational engineering information, the master asset information repository or asset attribute metadata.
• You have unstructured and untagged data and documents in different data storage locations or scanned P&IDs that are “dumb.”
What to do: Use the digital twin to aggregate, validate and contextualize these unstructured or dumb sources of data. Your digital twin application should include artificial intelligence tools to recognize scanned documents like P&IDs and then keep them up-to-date. For proof of compliance, digital twin software embeds workflows, turning unintelligent documents into key components of an intelligent digital twin.
Lastly, one of the most under-appreciated issues in managing plant, asset and operational performance is employee performance. Because a digital twin makes following maintenance and operational processes and responding to alarms easier, it will help your team stick with processes and enhance the visibility of leading and lagging key performance indicators. Often, a digital twin will encourage strong responses to weak signals and prevent complacency.
Take Advantage Of A Digital Twin
A digital twin affords visibility and contextual understanding of engineering information relationships. These important relationships provide the backbone for insightful change-impact analysis and enable context-relevant analysis and decision-making, reducing the time engineers spend on finding information. A digital twin pulls together data in context so you can see relationships better than ever and ensure any changes are captured and followed through to align all information. All relationships created during the engineering phase are carried over to operations to form a digital twin that acts as a common information backbone, giving engineers trusted information when and where needed.
A digital twin enhances collaboration among engineering, maintenance and operations by allowing them access to the same information while providing peace of mind they always are working with good-quality up-to-date data. A digital twin enables engineers to spot and troubleshoot equipment problems that otherwise might go unnoticed until it’s too late.
Sidebar: Collaboration Leads To Digital-Twin Cloud Service
The digital twin specifications outlined here were developed jointly by Bentley Systems and Siemens. PlantSight, a set of cloud services, uses Bentley’s design, engineering and construction experience along with Siemens’ extensive operational and maintenance knowledge to solve the growing dark-data problem faced by process plants.
PlantSight provides continually updated, trusted plant information. It visualizes the complete plant information model and reliability programs for all assets with an easy-to-use web-based portal view. Engineers gain digital line-of-sight working in an immersive digital mode of collaboration within the plant.
Besides giving an accurate, evergreen reflection of current plant conditions, the digital twin maintains a timeline that captures the how, when and initiator of engineering data changes.
PlantSight utilizes an open, connected data environment to access, validate and manage information; this provides the common link between project, plant, equipment, operational technologies and enterprise information.
SANDRA DIMATTEO is Burlington, Ontario-based director, digital twin solutions for asset and network performance, for Bentley Systems. Email her at [email protected].
