A Smart Autonomous Time- and Frequency-Domain Analysis Current Sensor-Based Power Meter Prototype Developed over...

Area of Science:

• Electrical Engineering
• Computer Science
• Artificial Intelligence

Background:

• Electrical energy management is crucial for smart grids, with the Internet of Things (IoT) enabling appliance monitoring.
• Current IoT energy management systems often rely on cloud analytics, but there's a growing need for edge-based solutions.
• Edge analytics in IoT devices offer potential for real-time processing and reduced latency in energy management.

Purpose of the Study:

• To develop a smart, autonomous IoT end-device for electrical energy management in smart grids.
• To integrate edge analytics capabilities within an Artificial Intelligence across IoT (AIoT) architecture.
• To demonstrate the feasibility of online load identification (auto-labeling) of electrical appliances at the network edge.

Main Methods:

• Development of a novel IoT end-device: a time and frequency analysis current sensor-based power meter prototype.
• Implementation of an edge analytics-based AIoT architecture combining cloud and fog-cloud analytics.
• Deployment of pre-trained AI models from the cloud to the edge device for autonomous onsite operation.

Main Results:

• Successful development of a functional hardware and software prototype for demand-side management (DSM).
• Experimental demonstration of auto-labeling (online load identification) of electrical appliances using the developed prototype.
• Validation of the proposed methodology as feasible and workable for edge-based energy management.

Conclusions:

• The developed edge analytics-based AIoT architecture and prototype are effective for smart grid energy management.
• Autonomous online load identification at the edge is achievable, enhancing the efficiency of demand-side management.
• The study presents a viable approach for next-generation IoT-enabled energy management systems.