A probabilistic inference-based harmonic source localization method considering imprecise network parameters

Fan, Yu

doi:10.1016/j.epsr.2026.113217

Return

A probabilistic inference-based harmonic source localization method considering imprecise network parameters

2026-05-01

PRE

DOI：10.1016/j.epsr.2026.113217

Original

Original request for help

Save

Abstract

En 中文

The increasing integration of distributed renewable energy sources and power electronic devices has exacerbated harmonic pollution in power systems, making accurate localisation of harmonic sources crucial for effective mitigation. However, existing methods often depend on precise network parameters or impose restrictive statistical assumptions about harmonic sources, limiting their practical applicability. To address these challenges, this paper proposes a probabilistic inference-based localisation method. This method introduces a unified framework that models admittance parameter uncertainties as bounded random variables and employs Monte Carlo sampling combined with correlation analysis to achieve source localisation. A key advantage of the proposed method is its ability to maintain accurate localisation despite significant parameter errors, without requiring restrictive assumptions about the harmonic sources. Under conditions of 30% network parameter error, the proposed method achieves correct localisation rates exceeding 92% in the IEEE 14-bus system and 84% in the IEEE 33-bus system, outperforming representative benchmark methods (BOMP, ICA, Inter-ICA, and FAJD) by more than 15% under the same conditions. This advancement offers an effective and practical solution for monitoring harmonic pollution in networks with imprecise parameters.

Keywords:

Power system harmonic

Harmonic analysis

Harmonic source localisation

Imprecise parameters

Monte Carlo sampling