Lecture Notes: Enhancing SCADA with Built-in AI

Introduction

• Presenters:
  • Brian Thichen, Chief Revenue Officer at Sorba
  • Aldo Fronte, CIO and Co-founder at Sorba
  • Travis Cox, Chief Technology Evangelist at Inductive Automation
• Objective: Discuss how AI can be integrated into SCADA systems to drive efficiency and performance.

Role of AI in Industrial Settings

• Impact Areas:
  • Provides insights from various data types (process, quality, MES-related, etc.).
  • Enables predictive maintenance, optimizes production, and improves quality control.
• Challenges:
  • Cultural and political barriers within organizations.
  • Importance of leadership in driving AI adoption.
  • Need for foundational data integration and modeling.

Key Challenges in AI Adoption

• Importance of having a clear use case and starting small.
• Cultural shift needed from leadership down to operational teams.
• Integration between OT and IT is crucial for seamless implementation.
• Need for scalable and integrated systems.

Critical Operations Impacted by AI

• Use Cases:
  • Predictive maintenance to minimize downtime.
  • Enhancing operator efficiency and decision-making through data augmentation.
  • Potential for AI to act as an assistant.

Integrating AI with SCADA

• SCADA systems serve as the foundational layer for operations.
• Importance of modeling data for unified access and integration.
• SCADA is crucial for digital transformation due to its fundamental role in data management.

Overview of Companies

Inductive Automation

• Founded in 2003, focuses on SCADA and industrial integration.
• Product: Ignition, a platform for building diverse industrial solutions.
  • Supports scalability, open standards, and direct integrations.

Sorba

• Founded by industry veterans, focuses on AI solutions for industrial automation.
• Offers a no-code environment for building AI models.
• Solutions deployed across various industries for enhancing operational efficiency.

Sorba and Ignition Integration

• Sorba's Platform:

  • DataOps for data integration and cleansing.
  • MLOps for building and managing machine learning models.
  • AutoML for automating the model building process.
  • Supports deployment across diverse hardware and environments.

• Integration with Ignition:

  • Leverages a module for seamless AI integration into Ignition.
  • SCADA remains the central hub for monitoring and controlling industrial processes.

AI Model Building with Sorba

• Steps:

  • Define assets and configure data channels.
  • Utilize AutoML for model training and deployment.
  • Supports a variety of AI and ML models (e.g., clustering, regression, digital twin).

• Operationalizing AI:

  • Deploy models in real-time within SCADA systems.
  • Enable auto-retraining to maintain model performance.
  • Real-time monitoring through Ignition's interface.

Closing Thoughts

• AI integration into SCADA is critical for future industrial solutions.
• Combining platforms like Sorba and Ignition can lead to significant operational improvements.
• Continuous learning and adaptation are key for successful AI deployment.

Q&A Highlights

• Handling imbalanced data and adjusting decision thresholds using AutoML.
• Typical sample rates for building effective ML models vary by application.
• Possibility of closed-loop control for automated response to anomalies.

---

Contact Information:

• Reach out to Sorba at [email protected] for further inquiries.
• Sorba University and Inductive Automation provide extensive training resources.