A stacking ensemble classification model for detection and classification of power quality disturbances in PV integrated power network

This paper proposes a stacking ensemble classification model to classify the different Power Quality Disturbances (PQDs) in Photovoltaic (PV) integrated power network. For this study, the network is developed in Matlab-Simulink with introduction of different PQDs for analysis. In pre-processing stage, Discrete Wavelet Transform technique is used to extract the features from different PQDs. The extracted features are used to train the base classifiers (Logistic Regression (LR), Naive Bayes, and J48 decision tree) at base level (level 0). The predictions from the base classifiers are used to learn the Meta classifier (LR) in next level (level 1) to get the final predictions. The proposed ensemble model attains higher classification accuracy than base classifiers under standard test condition (92.22%) and dynamic environmental condition of solar PV (91%), and addition of noise into the classifier (89.33%). Further, the proposed method offers superior performance than base classifiers in terms of evaluating performance indices.