An Industrial IoT Architecture for Autonomous Maintenance: Integrating Digital Twins, Distributed Control, and On-Demand Additive Manufacturing

Industrial Internet of Things (IIoT) technologies enable interconnected sensing, computation, and actuation across distributed industrial systems, but translating equipment health information into autonomous physical maintenance actions remains challenging. This work proposes and experimentally validates an IIoT-enabled Cyber–Physical Production System architecture that integrates equipment health estimation, Digital Twin-based prescriptive logic, distributed control, and on-demand Additive Manufacturing to close the loop between degradation assessment and maintenance execution. The architecture distributes sensing, health estimation, decision-making, and manufacturing functions across heterogeneous edge nodes using International Electrotechnical Commission (IEC) 61499 standard-compliant function blocks, DINASORE, and Message Queuing Telemetry Transport (MQTT)-based communication. A nine-node physical testbed comprising degrading robotic systems and a three-dimensional (3D) printer was developed to evaluate individual components and the complete information-to-action workflow. Experimental results demonstrate reliable distributed data acquisition and control, online degradation assessment, and autonomous triggering of replacement-part production. In three closed-loop trials on newly degraded components, the estimated health index crossed the prescribed threshold and automatically initiated fabrication, with the print started approximately 0.3 s after the order reached the manufacturing node and no human intervention in the decision path; the replacement part was produced before failure in all trials, although with margins of only 2 to 8 min. The results demonstrate how IIoT connectivity, distributed intelligence, and cyber–physical actuation can support autonomous maintenance workflows linking equipment health monitoring directly to physical manufacturing resources, and identify the timing of the maintenance decision, rather than the integration itself, as the critical factor for their reliable operation.