A Fault Location Method Based on Polynomial Chaos Expansion for Non-uniform Power Transmission Lines With Uncertainty Parameters

Abstract

This article introduces a novel fault location method utilizing polynomial-chaos expansion (PCE) designed specifically for non-uniform transmission lines affected by uncertain parameters. It considers the uncertain parameters arising from height and ground conductivity in transmission lines, examining their impact on conventional fault location methods, such as natural frequency and full-transient analysis approach. These uncertainties lead to considerable location errors, particularly magnified with increasing fault distances. To address this issue, we propose a fault location approach based on PCE and correlation estimation. Simulations cover fault distances ranging from tens to hundreds of kilometers, considering variations in non-uniform line section-lengths, and examining scenarios with single and multiple conductors. Results demonstrate that the proposed method exhibits robustness across different degrees of uncertainty parameters in non-uniform settings, reducing the relative location error to below 1%. In terms of computational efficiency, the PCE method can accelerate calculations by up to 12 times compared to the Monte Carlo method. Furthermore, the PCE method has been validated using fault transient data from an actual 220 kV power line achieving a location error of 2.41%, which demonstrates its practical applicability in real-world power grid scenarios.