An adaptive and reliable protection scheme for critical fault detection in IEC microgrid considering dissimilar AC faults and weather-based random scenarios

The rapid fault isolation is the necessity for the proficient operation of the microgrid because it can increase resiliency of the system by restoration of the power distribution system after isolation of the faulty area. In modern power system, many renewable and nonrenewable sources are integrated through different types of converters; therefore, it is too much tedious to develop a protection scheme which can easily detect and isolate faults under such unpredictable faulty conditions. Further, variation in weather conditions and the fault current level during such conditions is not predictable for traditional protection schemes and needs modification. In addition to above difficulties in the microgrid, distinct category of the AC faults makes protection task more difficult when fault resistance is varying due to change in grounding conditions. Motivated by the above challenges in the proposed microgrid, an ensemble of kNN-based protection scheme has been devised in this work to provide robustness to the microgrid. The tasks of the mode detection, fault detection/classification and faulty section identification under varying weather conditions have been considered in grid-connected and islanded modes as a class of the problem. Discrete wavelet transform has been used to pre-process the measured voltage and current signals retrieved from the appropriate bus. To validate the protection scheme, numerous cases of dissimilar operating scenarios have been considered under both of the operating modes. The results in Sect. 6 indicate that protection scheme is efficient and reliable to increase the robustness of the microgrid.