Edge-to-Cloud IIoT for Condition Monitoring in Manufacturing Systems with Ubiquitous Smart Sensors.

Area of Science:

• Industrial Internet of Things (IIoT)
• Machine Learning
• Predictive Maintenance

Background:

• Data-driven condition monitoring in manufacturing faces challenges with network congestion due to massive data transmission.
• Limited labeled anomaly data hinders supervised learning models for fault detection in industrial systems.

Purpose of the Study:

• To propose an Industrial Internet of Things (IIoT)-based condition monitoring system with an edge-to-cloud architecture.
• To reduce network traffic overhead and accurately detect anomalies in industrial settings.

Main Methods:

• Implemented an edge-to-cloud architecture for IIoT condition monitoring.
• Computed relative wavelet energy as feature vectors on the edge layer to minimize data transmission.
• Utilized an unsupervised deep long short-term memory (LSTM) network for anomaly detection.

Main Results:

• The proposed system effectively reduced network traffic overhead in the IIoT environment.
• Experimental results demonstrated accurate anomaly detection capabilities for manufacturing equipment.

Conclusions:

• The developed IIoT condition monitoring system offers an effective solution for intelligent manufacturing.
• The approach successfully addresses network congestion and data imbalance issues in anomaly detection.