Power quality disturbance identification using morphological pattern spectrum and probabilistic neural network

This paper proposes a method for identification of power quality (PQ) disturbances using morphological pattern spectrum (MPS) and probabilistic neural network (PNN). The PQ disturbance signals are decomposed by a three-order MPS to extract a number of features which are used for disturbance identification. These features compose a feature vector to train PNN classifier. The trained PNN is employed to classify PQ disturbances signals. The proposed method is tested by 760 PQ disturbance signals with additive noise, including sag, swell, interruption, harmonics, notching, oscillatory and fluctuation, which are simulated according to the IEEE 1159-2009 standard. The results demonstrate that the features extracted are effective and the PNN classifies disturbances with high accuracy rate.