Local Distributed Node for Power Quality Event Detection Based on Multi-Sine Fitting Algorithm.

Area of Science:

• Electrical Engineering
• Power Systems Analysis
• Signal Processing

Background:

• Modern power grids face increasing complexity due to new generation systems, smart grid technologies, and a higher number of connected equipment.
• Effective Power Quality (PQ) evaluation is crucial for timely grid intervention and maintenance.
• Existing monitoring solutions are often prohibitively expensive due to the extensive infrastructure required.

Purpose of the Study:

• To propose a novel, cost-effective architecture for an Automatic PQ Events Classifier (APQEC).
• To develop a low-resource PQ event detection and segmentation method suitable for low-cost distributed nodes.
• To reduce the overall cost of PQ monitoring infrastructure while maintaining effectiveness.

Main Methods:

• A new APQEC architecture comprising Locally Distributed Nodes (LDNs) and a Central Classification Unit (CCU).
• LDNs perform signal acquisition, PQ event detection, and segmentation using a sliding observation window and multi-sine fitting algorithm.
• The CCU receives and classifies potential PQ events from LDNs, discarding irrelevant data to minimize communication and processing load.

Main Results:

• The proposed method effectively detects and segments PQ events with low computational requirements.
• Numerical tests validated the method's performance across various PQ events and signal-to-noise ratios (SNR).
• Experimental results demonstrated the effectiveness and reduced execution time compared to Fast Fourier Transform (FFT)-based methods.

Conclusions:

• The proposed LDN-CCU architecture offers a viable and cost-effective solution for widespread PQ monitoring.
• The low-resource detection and segmentation method enables efficient processing at the edge.
• The system successfully reduces monitoring costs and execution time, enhancing grid management capabilities.