A New Signal Processing Based Solution for PQ Disturbance Classification

Software and hardware for automatic classification of power quality disturbances are highly needed for both utilities and commercial customers. Existing automatic recognition methods need much improvement in terms of their capability, reliability, and accuracy. This paper presents a new signal processing based approach for discriminating power quality events. A classification-driven time-frequency representation (TFR) is designed and used for feature extraction. A neural network with feed-forward structure i s chosen as the classifier. The long-term applications of this technique include: enhancement of real-time power system protection, statistical analysis of power quality problems, and incipient fault detection. This algorithm is successfully demonstrated with a large number of simulated test data. An algorithm demonstration software is developed using Matlab GUI. This paper also introduces the data source for this project. The current and future research work of the SEAL lab on PQ monitoring is also discussed.