An Improved Fault Identification and Location Prediction Method for AC Microgrid

The idea of a microgrid is used to actively encourage the usage of distributed energy resources (DERs) to reduce the load on the power grid that relies on fossil fuels. However, the dynamic behaviour of microgrid systems under faulty conditions makes adaptive protection a general requirement for reliable microgrid operation. Therefore, identification and exact fault location prediction are essential for the system to isolate and recover from faults quickly, ensuring smooth microgrid control and operation. This paper presents an improved fault identification and fault location prediction method for AC microgrids. First, a Support Vector Machine(SVM) Classifier is used in the developed microgrid to detect the fault class and faulted line section. Subsequently, Support Vector Regression (SVR) is used to predict the precise fault point in the detected line section. Experiments with different fault conditions validate the efficacy of the proposed methodology at various locations in an IEEE 9-bus meshed AC microgrid developed in the SIMULINK platform.