Synchronization stability analysis and control of grid-connected wind farms
File(s)
Author(s)
Nagam, Sai Sowmya
Type
Thesis
Abstract
Wind farms (WF) are typically interfaced with the grid through power converters, where the controllers largely define the characteristics. The controllers are designed to ensure stable interconnection while meeting the set control objectives. The interconnection can be jeopardised by dynamic interactions within the WF and with the rest of the network. This work focuses on dynamic interactions in the context of synchronisation stability and, later, WF stability in general. A model-based approach using the port-Hamiltonian (pH) framework is employed for synchronisation stability analysis and enhancement. A model-free approach based on Dynamic-mode Decomposition (DMD) is employed for a more general assessment of wind farm stability. Synchronisation stability analysis of Synchronous Reference Frame – Phase-Locked Loop (SRF-PLL) and Double Second-Order Generalised Integrator (DSOGI) - PLL uses pH theory. It is observed that the timescale separation assumption employed in controller design results in optimistic stability predictions, masking the dynamic interactions of PLL with the rest of the system. Therefore, the stability assessment is carried out while accounting for converter current and voltage dynamics, which the controllers define. Conventional converter controllers restrict the PLL response time to avoid dynamic interactions. The Interconnection and Damping Assignment - Passivity-based Control (IDA-PBC) controller is proposed as an alternative strategy, allowing higher control bandwidth to enhance synchronisation stability. It is shown that the IDA-PBC current controller operates relatively independently of PLL dynamics. The suitability of the DMD algorithm for stability assessment of converter-based systems is discussed. Modal analysis using DMD is validated against the critical modes obtained from the detailed eigenvalue analysis of the WF test system. It is observed that DMD can reach comparable or better accuracy with just $10$\% of the model order. In contrast, a full-order model is necessary for detailed eigenvalue analysis. Thus, DMD opens an opportunity for real-time tracking and stability assessment.
Version
Open Access
Date Issued
2023-08
Date Awarded
2024-02
Copyright Statement
Creative Commons Attribution NonCommercial Licence
License URL
Advisor
Pal, Bikash
Sponsor
European Union
UK Research and Innovation
Grant Number
861398
EP/W005557/1
Publisher Department
Electrical and Electronic Engineering
Publisher Institution
Imperial College London
Qualification Level
Doctoral
Qualification Name
Doctor of Philosophy (PhD)