Secured protection of transmission line by distance relay using data mining approach

Abstract

Distance relay is one of the most important protection element of a transmission line used in protection schemes. Relay may malfunction if it is not able to distinguish faults from system stressed conditions. This work mainly focusses on enhancing the performance of the distance relay in a secured manner based on data mining approach which uses two phases of classification. Level 1 classifier identifies the system conditions like normal, fault, and power swing and level 2 classifier gets initiated when there is a power swing and distinguishes between the persistence of power swing condition and a three-phase fault. In both the phases, the protection scheme in the respective zone where the fault occurred gets activated. The proposed methodology is tested for an IEEE 9-bus system wherein the data is collected from phasor measurement units placed in optimal locations. Optimal PMU placement is economical since it overcomes issues like cost, communication infrastructure issues, maintenance and complexity. The results proved that the proposed method is effective with good efficiency and higher accuracy with less number of PMUs.

Early fault identification for operating circuit breaker based on classifier model system

“…Tail spring adjustment is sometimes used to hasten the trip on a slow CB, but this neglects the main reason for the slow breaker which may be due to faulty lubrication. The analysis can also detect trip coils that are close to failure, as well as inappropriate/dirty auxiliary switch contacts [21], [22].…”

A security enabled real time fault detection and classification of power system conditions

“…Kiruthika and Bindu [30], carried out a study for the classification of system conditions like normal, fault, and power swing and also to discriminate symmetrical fault during power swing by placing PMUs at all bus locations in IEEE-9 Bus system using data mining algorithms. Kiruthika and Bindu [31], proposed a methodology by considering data collected from the PMU's placed at optimal locations of the IEEE -9 Bus system for the classification of system conditions like normal, fault, and power swing and also to discriminate symmetrical fault during power swing and zone protection using data mining algorithms. ML/data mining classifiers considered for the above analysis are k-nearest neighbours (KNN), naïve bayes (NB), DT, and convolutional neural network (CNN).…”