Detection of Fault in Transmission Line During Power Swing Using Fuzzy C-Means Clustering

In normal condition distance relay's measured impedance is greater than set impedance. But during swing in power the impedance sensed by relay is less than relay's set impedance, Resulting trip of relay at the time of swing without any fault. Relay should be blocked & should work after a fault is there at the time of swing. The paper proposes new method to detect fault at the time of power swing and classify the fault with swing. In the technique process starts by collecting samples of current & voltage during swing & process it through fuzzy c-means clustering, then we get two clusters one of swing with fault & other of swing without fault. A set of training data is created & a data is passed through it, if data lies near to faulty cluster centroid then it is fault otherwise swing near to other centroid. A two parallel line circuit model is checked in PSCAD (TM)/EMTDC (TM).