Abstract
Power quality disturbances (PQDs) pose significant challenges in modern energy power plants-based systems (MEPPBS), especially with the increasing integration of renewable energy sources (RESs) such as wind and solar photovoltaic (PV) systems. The intermittent nature of these sources introduces voltage fluctuations, harmonics, and transient disturbances, affecting grid stability and reliability. This paper presents a novel dual algorithm-based protection approach for detecting, classifying, and mitigating PQDs in grid-connected MEPPBS. The proposed method utilized an advanced adaptive median filter (AMF) as a signal processing-state observer and a support vector machine (SVM)-based scheme to accurately identify disturbances, including voltage sags, swells, harmonics, interruptions, and transients. The proposed dual-algorithm approach, combining AMF and SVM, offers a novel solution that enhances PQD detection accuracy and speed compared to existing methods. Furthermore, the proposed scheme tests five individual PQDs and ten combined disturbances-based datasets, including voltage sags, swells, harmonics, interruptions, and transients were used to train the proposed SVM classifier. Then, SVM-based residuals were calculated by SVM algorithms from estimated AMF data, SVMBR index reveals the detection of PQDs quickly. Simulations were performed in MATLAB® R2022b (Version 9.13) to evaluate the effectiveness of the suggested approach under various operating conditions. The results demonstrate high detection accuracy of 97%, fast response times of less than 15 milliseconds, & robustness in discriminating different PQDs when trained by just 50% of the dataset under a signal-to-noise ratio (SNR) of 20dB. The proposed method achieves 96% precision, 94% recall, and a 0.04 false positive rate, demonstrating high accuracy and reliability in PQDs detection. The findings highlight the potential of the presented method to enhance power system resilience and ensure reliable operation in renewable energy-integrated grids.