A Neural Network Based System for Recognition of Power Quality Disturbances

Recognition of power quality (PQ) disturbances has become imperative for utilities as well as for consumers due to increasing cost burden of poor power quality and augmented use of sensitive electronic equipments. It is also important for identification and control of sources of PQ disturbances. But, recognition of PQ disturbances is often troublesome because it involves a broad range of disturbance categories or classes. This paper presents a recognition system for power quality (PQ) problems in electrical power systems. Most of the methods used for this purpose entail application of combination of signal processing and artificial intelligence techniques for feature extraction and recognition respectively. The work presented in this paper proposes and implements a novel algorithm based on continuous wavelet transform (CWT) for generating signatures of PQ disturbances and feed-forward neural network for recognition of these signatures. Almost 100% accuracy of recognition substantiates the effectiveness of the proposed system.