Answers are grounded in this paper and cite their sources.
The manufacturing process of magnesia-carbon bricks involves complex equipment and dense workflows, leading to frequent production anomalies that are hard to diagnose. Traditional modelling methods struggle with multi-level anomaly detection and process awareness. This paper proposes a digital twin-based multi-layer modelling framework for anomaly detection in magnesia-carbon brick production. The framework unifies equipment behaviour and workshop logic through three layers: equipment twin, workshop twin, and system decision-making. It addresses modelling gaps by dividing equipment into five component types, creating a three-level twin structure (unit, chain, and equipment) to map structure-action-state relationships. For workshop logic, a complex event processing method models process events using behaviour correlation matrices and directed graph adjacency matrices, capturing time evolution and path dependencies. A prototype system was developed, integrating data collection, state mapping, anomaly detection, and visualisation. Initial tests in a pilot workshop confirmed its feasibility and scalability. This research offers a structured, multi-level modelling approach for anomaly detection in complex manufacturing, enhancing digital twin expressiveness and workshop intelligence.
No takes yet. Share an insight, caveat, or question.
Yi et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: