Time-Domain Differential Protection of Transmission Lines Based on Bayesian Inference

This paper presents a time-domain differential protection method based on the Bayes’ Theorem and on the current samples to determine the fault occurrence probability as a function of time. It also describes the theory involved in developing the proposed protection method when considering the differential protection of transmission lines. The authors performed fault simulations using an existing electrical system modeled in EMTP varying the fault type, phases involved, fault resistance, inception angle, and fault location, and carried out a sensitivity analysis regarding high impedance faults, current transformer saturation, noisy data, misalignments between terminal samples and power swings. The authors also tested the proposed method against actual fault events from four existing transmission lines of the Brazilian interconnected power system. All results indicate that the proposed solution is fast, accurate, secure, and dependable.