Skip to content
Open access

Robust Grid Impedance Estimation Based on Dynamically Weighted Likelihood Function and Adaptive Bandwidth Optimization

Oct 2026 · Electricity · 0 citations · 24 references

Abstract

Online identification of grid-side wideband impedance is of great significance for stability assessment and the formulation of coordinated suppression strategies. Bayesian estimation mitigates the influence of background harmonics on grid-side wideband impedance estimation, thereby providing an effective means for its identification. However, anomalous data interference at the point of common coupling (PCC) can cause a shift in the posterior distribution of traditional Bayesian estimation methods. Moreover, the bandwidth coefficients in conventional Kernel Density Estimation (KDE) are typically selected based on empirical values, making it difficult to adapt to varying practical measurement conditions. This paper proposes a robust non-parametric Bayesian impedance estimation method with adaptive bandwidth optimization. The proposed method introduces a weighted likelihood function, where sample weights are guided by physical constraint residuals, granting the likelihood function inherent robustness against anomalous signals. To address the challenge of parameter determination, the Likelihood Cross-Validation (LCV) criterion is integrated to achieve a closed-loop search for the optimal bandwidth by maximizing the predictive likelihood function within a prescribed candidate range. Simulation results demonstrate improved estimation accuracy and robustness under strong interference conditions, while field measurements at multiple harmonic frequencies further support the practical applicability and estimation stability of the proposed method.

Read PDF

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.