An Intelligent Particle Filter with Neural Network for Fault Location and Classification in Microgrid

Microgrid concept is initiated due to increasing involvement of distributed generation resources with the utility grid. Microgrid provide reliable and sustainable power but the protection of microgrid become challenging due to bidirectional power flow, dual mode of operation (grid connected and islanded mode). Faults in the microgrid reduce its stability and efficiency. Identification, classification, and location of faults are critical for rapid restoration and microgrid protection. This research proposes a neural network-based intelligent particle filter for microgrid fault detection and classification. Even with low fault current, which is typical of inverter-based DGs, the suggested method seeks to precisely identify fault kinds, locations, and directions. The features are extracted from data using S-Transform, then extracted features are estimated using particle filter. A neural network is then used for classification and finalisation of location. The proposed scheme provides extremely precise fault detection, ensuring that the classification and location of the fault are promptly identified for effective protection and service restoration.