Mode Clustering Based Dynamic Equivalent Modeling of Wind Farm for Small-Signal Stability Analysis

TL;DR

This paper introduces a mode clustering method to create dynamic equivalent models of wind farms, improving small-signal stability analysis by accurately representing oscillation modes with fewer aggregated modes.

Contribution

It proposes a novel mode clustering approach based on modal participation factors to generate simplified yet accurate dynamic equivalents of wind farms for stability analysis.

Findings

The aggregated DEM accurately matches the frequency response of detailed models.

The method effectively evaluates oscillation mode similarities among wind turbines.

Time-domain simulations confirm the DEM's validity for stability assessment.

Abstract

Dynamic equivalent models (DEMs) are necessities for the small-signal stability (SSS) analysis of the power system with large-scale wind farms (WFs). This paper proposes a mode clustering based dynamic equivalent modeling method of WFs for the SSS analysis. It is deemed that a DEM can be used to represent the whole WF to evaluate its impact on the SSS of power systems, as long as the frequency response of the DEM adequately matches that of the detailed WF model around the frequency of oscillation modes of concern. Focusing on the concerned oscillation modes in the small-signal model of the whole WF, closely distributed modes of them on the complex plane are classified into a cluster and then represented by a single mode. By the linear superposition principle, the modal participation factor (MPF) regarding each of modal clusters are superimposed, generating a feature vector for each wind turbine (WT). Based on the feature vectors, the oscillation feature similarity of the WTs as well as their aggregation can be evaluated and performed easily. Taking the DC-link voltage control (DVC) mode of full-power WT converters as an example, the aggregated DEM is verified by frequency-domain analyses and time-domain simulations.