Power Quality Disturbance Detection in Hybrid Wind-SPV Plants using Deep learning approaches

The surge in renewable energy adoption, notably hybrid wind-solar photovoltaic (Wind-SPV) plants, necessitates optimal operation and power quality assurance. This paper introduces a novel deep learning approach to identify electricity Quality Disturbances (PQDs) in hybrid Wind-SPV plants that are connected to electricity networks. The approach use Continuous Wavelet Transform (CWT) to analyze Power Quality Disturbances (PQDs) and produce picture files of CWT scalograms. The hybrid deep learning technique integrates a deep neural network (DNN) based on MATLAB/Simulink with Neighbor Component Analysis (NCA) and Support Vector Machine (SVM). Image data are processed using ResNet and VGG-Net architectures in the MATLAB/Simulink environment. NCA is used to extract distinguishing features from the final dropout layer. SVM classification detects anomalies in the hybrid Wind-SPV system. MATLAB/Simulink simulations evaluate the method using real PQD data from modified IEEE 9-bus test systems, achieving high accuracy (98.54%), precision (97.57%), and recall (95.21%) for the IEEE 9-bus system. This hybrid deep learning approach proves effective in addressing power quality issues associated with Wind-SPV integration, contributing to the development of reliable and efficient renewable energy systems for a sustainable future.