A Smart way to Detect Islanding and Fault in Overhead Power Lines Using Magnetic Measurements

Power system is large network consisting different components. In addition to the conventional system, when the solar system is added in Power System, the complexities increase. The system undergoes different modes of operations. To carryout efficient operation, it is necessary to operate the breaker that switches the operation of conventional system into solar system and provide continuous supply of power. The methodology is based on detection of different modes such as normal operation, fault scenario and islanding mode which will be beneficial in recognition of kind of issue through which the power system has gone. For that purpose, both Simulink tool of MATLAB and Machine learning is employed. The methodology proves to be beneficial for every situation, environment and issues for the power system. The results have been generated for normal operation, fault scenario and islanding. The accuracy of the results have been obtained by different methods of machine learning including neural network and decision tree. The system has been trained for the patterns and can accurately differentiate among three cases. The results provide efficient way of detection through a contactless method which is based on utilization of a single-phase sensor and training of machine for easily recognizing among different cases and then operated according to the situations of the trained files.