When a fault happens in a distribut

When a fault happens in a distribution feeder, it causes to produce traveling waves (TWs) that travel between fault location and substation. Hence, frequency of the TW measured at the substation is directly related to the fault distance from substation. By increasing fault distance, period of TW increase and so their frequency goes down. However, for a distribution system that comprise of a main feeder and many laterals, in addition to the reflected waves from the fault location (the main components), some reflected waves from junctions also reach to substation (measuring point). These waves (which contain less energy in compared to the main components) also generate components in frequency spectrum of faulted voltages that in this paper are called subordinate components. Using Clark transformation, the faulted voltages are first transformed from phase domain into modal domain. Then frequency analysis is performed on aerial mode voltage (V1) and ground mode voltage (V0) to identify both main and subordinate components. Comparing the results obtained from analysis with earlier generated database, the proposed algorithm determines faulted section and distance of fault uniquely. Because of delta connection of the secondary winding of distribution transformers, amplitude of single line to ground (SLG) fault current with respect to other fault types is lower. Hence, the amplitude of TW produced by SLG faults in comparison to other fault types is smaller considerably. In this circumstance distinction between the main and the subordinate components cannot be done precisely. Accordingly, faults are divided into two groups in this paper. The SLG faults and the other fault types. Two individual algorithms are developed for detecting the section and location of each group of fault.