Detection of High Impedance Faults in Primary Distribution Grid using Support Vector Machines Classification

The occurrence of high impedance faults (HIF) in primary distribution grid poses a danger to the safety of people, equipment, and animals. However, protection devices along the distribution network are not capable of being sensitized by this type of defect, most of the time. This work presents an integrated strategy of HIF classification and detection, based on the use of Support Vector Machines. An improved fault model was used to emulate randomness behaviors, and especially intermittence of high impedance faults. The residual current is monitored, and the extraction of its characteristics is performed with Short-Time Fourier Transform. A logic of temporal consistency has been applied to the detection stage. The presented algorithm´s operation was achieved throughout several simulations in a 20 kV distribution system.