Automation of an intelligent SCADA system for production line monitoring and supervision

Abstract

Modern industrial environments face significant challenges due to unexpected downtime caused by mechanical and electrical faults. This Master’s thesis presents an Intelligent SCADA System designed for real-time monitoring and automated fault diagnosis across multi-stage production lines. The architecture integrates a robust multi-protocol network linking TIA Portal, Factory I/O, Node-RED, and MATLAB, utilizing a Multi-Layer Perceptron to classify equipment states. To ensure operational continuity and high system availability, it deploys two completely independent diagnostic frameworks: an Online AI-Assisted layer powered by the Gemini LLM for complex web-based root-cause analysis, and a local Offline Weight-Based layer on the WinCC HMI.