Classification and identification of power system events using Hilbert Huang Transform

This work presents a mathematical tool applicable to the characterization and classification of power system events. Disturbances without a periodic pattern or with a nonlinear pattern require a more suitable tool than the Fourier series (Fast Fourier or Windowed Fourier Transforms). To overcome the difficulties, other tools have been broadly used in the past years, such as the Wavelet Transforms. However, these transforms have also some drawbacks that the Hilbert Huang Transform technique could mitigate. In the paper the technique is applied to create the input vector database suitable for using a neural network methodology.