Power quality disturbance classification employing modular wavelet network

The wavelet network (WN), which integrates wavelet decomposition concept and neural network (NN) training method, is employed for classification of power quality disturbances. Unlike earlier approaches of feature extraction with wavelet transform and then classification by NN, this method classifies the disturbances directly by the WN. The modular concept incorporated in this approach solves a relatively complex problem by decomposing it into simpler subtasks which are easier to manage and then assembles the solution from the results of the subtasks. An extended Kalman filter (EKF) based network-design is also developed for the purpose. Results show that the method provides high classification accuracy