Discrimination between Inrush current and Internal Faults using Pattern Recognition Approach

This paper presents a new approach to distinguish between inrush current and internal faults of power transformer using pattern recognition approach. The HS-transform (hyperbolic S-transform) is used to extract patterns of inrush current and internal faults from the captured transformer current. HS-transform is a very powerful tool for non-stationary signal analysis giving the information of transient currents both in time and frequency domain. The spectral energy and standard deviation are calculated to distinguish between inrush current and internal fault. Classification of internal faults and inrush current is done through fuzzy C-means clustering.