A Fault Detection and Faulty Section Identification Scheme for Micro-Grids

In recent years, microgrids have become more prominent in power systems, which means the existing conventional protection schemes must be approached with caution while using them for micro-grid protection. Presently available micro-grid protection techniques bank on features extracted from the extreme ends of the feeder to recognize any fault but do not provide any information regarding the status of in-line feeders. The scheme presented in this paper incorporates a light GBM classifier to identify the fault and give information about the faulty feeder section. Discrete Fourier Transforms (DFTs) have been used for previewing the time-synchronized measurements. For simplicity, the statistical features were computed from DFT coefficients and used as input features. A standard IEC 61850-7-420 microgrid incorporating renewable distributed generators (DGs) has been used to perform simulation studies by varying different operating conditions. The proposed methodology has been validated by the results illustrated.