In-Situ Model Validation for Continuous Processes Using In-Network Computing

2024-05-12

Industrial Cyber-Physical Systems

AI Summary

Industrial data-driven models for continuous processes often suffer from long-term model-process mismatch due to dynamic process shifts, risking operational instability or equipment damage. To address this, we propose a lightweight, real-time online model validity verification method. Our approach innovatively offloads the validation task to an Intel Tofino-based programmable switch, establishing an In-Network Computing architecture that overcomes latency and bandwidth bottlenecks inherent in conventional cloud- or edge-based validation. By integrating prior process knowledge modeling with real-time variable comparison detection, the method achieves millisecond-level state identification and response. Experimental evaluation on a laboratory-scale water treatment platform demonstrates significant improvements in detection accuracy and real-time performance, validating both feasibility and engineering practicality in realistic industrial settings.

Technology Category

• Machine Learning: Hardware-aware ML
• Planning, Routing, and Scheduling: Model-Based Reasoning
• Cognitive Modeling & Cognitive Systems: Other Foundations of Cognitive Modeling & Systems

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Data management and stream processing for Web, mobile and wireless applications
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Economics, Online Markets and Human Computation: Incentives in network design for Web infrastructures and ecosystems

Abstract

The advancing industrial digitalization enables evolved process control schemes that rely on accurate models learned through data-driven approaches. While they provide high control performance and are robust to smaller deviations, a larger change in process behavior can pose significant challenges, in the worst case even leading to a damaged process plant. Hence, it is important to frequently assess the fit between the model and the actual process behavior. As the number of controlled processes and associated data volumes increase, the need for lightweight and fast reacting assessment solutions also increases. In this paper, we propose CIVIC, an in-network computing-based solution for Continuous In-situ Validation of Industrial Control models. In short, CIVIC monitors relevant process variables and detects different process states through comparison with a priori knowledge about the desired process behavior. This detection can then be leveraged to, e.g., shut down the process or trigger a reconfiguration. We prototype CIVIC on an Intel Tofino-based switch and apply it to a lab-scale water treatment plant. Our results show that we can achieve a high detection accuracy, proving that such monitoring systems are feasible and sensible.

Problem

• Validating model accuracy for continuous industrial processes
• Detecting process deviations using in-network computing
• Ensuring safety by monitoring and reconfiguring control models

Innovation

• In-network computing for model validation
• Continuous in-situ monitoring of process variables
• High accuracy detection using a priori knowledge