DWTGB-Based Power Quality Disturbance Detection in Electric Vehicle Charging Station

In response to the growing necessity for ways to reduce reliance on fossil fuels and decrease carbon emissions, there has been a recent increase in the production and use of electric vehicles (EV). This rise in EV adoption has driven the expansion of fast-charging stations across the power grid, which can lead to challenges in power quality (PQ), protection, and stability both in the power grid and at the charging stations. To this end, this paper utilizes ML methods for the detection and classification of power quality disturbances (PQD) at EV charging stations. Charging stations for EVs can not only cause power quality disturbances in the grid, including, but not limited to, harmonics, sags, flicker, and interruptions, but they can also be impacted by disturbances originating from the grid. Accurate detection of PQD is crucial to prevent equipment damage and mitigate more severe problems. For this purpose, wavelet analysis is used for feature extraction from signals, followed by machine learning techniques for event detection and classification. The effectiveness of these methods has been evaluated, and the best performance has been achieved through the DWTGB, which is the combination of discrete wavelet-based feature extraction and the Gradient Boost (GB) method, resulting in excellent and reliable detection accuracy.