Skip to main content

Articles & Blogs

Harnessing the Data Dividend in the Process Industries with Hexagon's Smart Digital Reality™

In the ever-evolving landscape of artificial intelligence (AI), data reigns supreme. As generative AI gains prominence, the importance of data quality and quantity becomes even more pronounced. McKinsey’s concept of the “data dividend” aptly describes the competitive edge that can be gained through strategic data utilization. For process industries, this entails a dual approach, focusing on both contextual and informational data. Hexagon's Smart Digital Reality simplifies access to both in sufficient quality and quantity. 

 

Contextual Data: The Compass for AI Understanding 

Contextual data provides AI with the understanding necessary to decipher the relationships between physical items, processes and data points within process plants. This context is delivered through two related approaches: 

  1. Plant Topology: Although being very powerful, plant topology data is often underutilized beyond design and engineering. It represents an untapped potential, especially for AI. Plant topology provides insights into how a plant operates, how equipment components interact and how processes impact each other. By leveraging this data, AI models gain a deeper understanding of the plant’s intricacies, enabling more informed decision-making. Plant topology is typically derived from process schematics like piping and instrumentation diagrams (P&IDs). 

  1. Asset Linkages: At Hexagon, we recognize that data interconnectivity is essential. Our solutions, including HxGN SDx2, utilize AI to simplify the creation of connections between tags, documents and data. By automating these workflows, we ensure that data and its interconnections remain up to date. This interconnectedness resembles a knowledge graph, enhancing AI’s ability to navigate complex information landscapes—a principle echoed by Gartner regarding the value of knowledge graphs in amplifying large language models (LLMs). 

 In process industries, with its great degree of heterogeneity, it remains very challenging to build AI-based solutions at scale. Even if one intends to build rather simple solutions for only one equipment type to monitor performance, one faces different process conditions, different level of automation etc., making it almost impossible to apply one algorithm for every piece of equipment without significant data engineering efforts.

Implementing your own Smart Digital Reality can help with these challenges. Its contextual data enables you to easily answer questions like “Which equipment tags are impacted by a failure of equipment tag XYZ?” and “Which data is available for tag XYZ?” That is the foundation for AI workflows at scale and in depth.

Informational Data: The Substance of AI Decision-Making 

Informational data serves as a critical component of artificial intelligence, enabling it to analyze and interpret various aspects of the digital world. This category of data includes: 

  • Transactional records, which detail business interactions. 

  • Time-series data, which monitors temporal progressions. 

  • Unstructured data, encompassing documents and images such as reality scans. 

However, data silos can hinder the access and reliability of information. For example, different departments in an organization might have inconsistent copies of the same datasheet. How do you know which one is accurate? Also, data silos often stem from organizational silos. There is a lot of potential in combining engineering and asset management data, but they usually lack a common interface. Smart Digital Reality technology solves these problems by integrating data. With this efficient approach, AI engineers can focus on improving and fine-tuning AI models instead of searching through data. They have a single source of truth, which significantly reduces the time spent on finding and verifying data. 

 

Hexagon: Empowering AI at Scale 

Hexagon's Smart Digital Reality is a pivotal enabler for AI utilization beyond proof-of-concepts. Our suite of design and engineering solutions provides the contextual backbone for AI applications. These solutions ensure scalability and mitigate the risk of erroneous outputs in generative AI models. However, our solutions cover the entire asset lifecycle beyond design and engineering, containing valuable informational data.

Examples include: 

  • HxGN EAM: A comprehensive asset record system.

  • Hexagon’s Plant State Integrity (PSI): Delivers a standardized view of alarm and event data. 

  • Hexagon’s EcoSys: Acts as a central aggregator for project performance data. 

  • Hexagon’s j5 Shift Operations Management: Consolidates operational data for better system visibility and control. 

Combining data from these solutions into Smart Digital Reality will enable complex AI-enhanced workflows in the future. Imagine an alarm flood is noticed in PSI. To determine its root cause, plant topology is used to find causal relationships with other tags. Then, via asset linkages, relevant data connected to these tags is found in other entries in HxGN EAM and j5. These records ultimately reveal that improper maintenance is the real root cause.

Only with both contextual and informational data in high quality and quantity, can such use cases be done at scale due to tremendously reduced data engineering efforts. Ultimately, such a data foundation is needed to enable meaningful (generative) AI-enhanced applications and Smart Digital Reality is the key to unlocking the data dividend in the process industries.

About the Author

Martin joined Hexagon's Asset Lifecycle Intelligence division in 2022 as Lead Strategist for AI and Enabling Technologies. Previously, he led product management at Bilfinger Digital Next, part of Bilfinger group. He has an educational background in chemical engineering, business administration and data science.

Profile Photo of Martin Bergmann