The stochastic modeling of fault-induced transients

Abstract

A study on the stochastic characterization of fault-induced transients in HV and EHV transmission lines is performed, considering time dependence for both variance and autocorrelation. The aim is the definition of noise models for the design and the evaluation of digital relaying algorithms. A number of realistic network structures are considered for 400 and 220 kV lines, including shunt reactors and capacitors. Frequency-dependent models for machine windings are applied and proved to be important for the long-term behavior of the noise. The capacity of plants and network sparsity near the sending bus are investigated and shown to have a major role in the transients magnitude. Analog prefiltering is also assessed on a stochastic basis. The sampled variance and autocorrelation provide the basis for general conclusions on the modal dependence of the noise on network structures and for state modeling. >