Generative AI Use Cases for OPCUA

This blog explores potential areas where OPCUA could benefit from the use of generative AI.

Generative AI has a growing list of capabilities that include generating and manipulating documents, generating and executing code, and interacting with systems through APIs. AI can be applied to OPCUA client/server development including device and system configuration and testing in a variety of industrial contexts.

Depending upon perspective, there are many ways to look at an industrial enterprise. One way is through the lens of a system framework. Several relevant industrial system frameworks have been developed such as ISA 95, GWAC interoperability framework, and RAMI 4.0.

Using these frameworks, one possible approach is to categorize AI/UA use cases according to the primary life cycle phases of 1) design and development, and 2) operations and maintenance. Using these as a basis, the following might provide a use case landscape for further brainstorming:

Design and Development

Product engineering - developing commercial systems

• Device engineering
• Systems engineering

Project engineering - adapting products to specific processes

• Integration engineering
• Application engineering
• User interface engineering
• Control engineering

Operations and Maintenance

• Historical Data Analysis (offline)
• Trend Data Analysis (online)
• Real-time Data Analysis (streaming)
• Alarms and events analysis (streaming)
• Process modeling and simulation (offline)
• Process forecasting and optimization (online)
• Fault detection and isolation (online/offline)
  • Root cause analysis

Potential "Low Hanging" Use Cases

Product engineering

• Simplifying OPC UA client and server development
  • Using UA and companion specifications

Project engineering

• Simplifying integration of OPC UA servers into existing plant operations and maintenance systems
  • Using UA, companion specifications and browse spaces
  • Using an AI enabled editor (e.g. Cursor)
• Creating project specific user interfaces
  • Using an AI enabled editor (e.g. Cursor)

Historical and trend data analysis

• Simplifying integration of OPC UA servers into AI advisory systems
  • Using an assistant API (e.g. OpenAI Assistant API)
  • Using an AI enabled editor (e.g. Cursor)

Some More Thoughts

AI will evolve as accuracy and trust in AI increases

• Design-time will typically precede run-time
• Historical data analysis will typically precede real-time data analysis
• Engineering assistants will typically precede operations/maintenance assistants
• AI advisory systems will extend preconfigured systems for ad-hoc analysis and display
• AI advisory systems will evolve into AI control systems

OPCUA documents relevant for AI use cases include: UA specifications, Companion specifications, Node Ids, Node Types, Node Set, and Browse Space.

Having Fun With AI,

Dave Hardin

For further information about OPCUA https://opcfoundation.org/