Rule-Based Non-Intrusive Load Monitoring Using Steady-State Current Waveform Features

Area of Science:

• Electrical Engineering
• Energy Systems
• Artificial Intelligence

Background:

• Effective energy monitoring is crucial for reducing power consumption.
• Non-intrusive load monitoring (NILM) offers a cost-efficient approach to disaggregate energy usage from aggregate measurements.
• Current NILM techniques often demand extensive datasets or complex algorithms for high performance.

Purpose of the Study:

• To propose a novel NILM technique, CRuST (Current waveform features with Rule-based Set Theory), for accurate individual load identification.
• To develop a NILM system that requires fewer data and simpler algorithms compared to existing methods.
• To evaluate the performance and accuracy of the proposed CRuST NILM technique.

Main Methods:

• The CRuST NILM architecture involves event detection using current waveform features, signal preprocessing, feature extraction, and load identification.
• An event detection stage identifies changes in connected loads based on current waveform characteristics.
• Load identification utilizes six extracted current waveform features and an appliance model within a rule-based set theory framework.

Main Results:

• The CRuST NILM technique demonstrated over 96% accuracy across various test scenarios.
• The proposed method achieved superior performance compared to feed-forward back-propagation networks and other existing NILM approaches.
• CRuST NILM effectively identifies the event-causing load using extracted features and an appliance model.

Conclusions:

• The CRuST NILM technique provides a highly accurate and efficient solution for disaggregating electricity energy usage.
• This approach overcomes limitations of existing NILM methods by requiring less data and employing a more straightforward algorithm.
• CRuST NILM offers a promising advancement in smart energy monitoring and management systems.