AAU Energy
PhD Defence by Wentao Liu
Pon 101 - 1.001/online
23.05.2024 13:00 - 16:00
English
Hybrid
Pon 101 - 1.001/online
23.05.2024 13:00 - 16:0023.05.2024 13:00 - 16:00
English
Hybrid
AAU Energy
PhD Defence by Wentao Liu
Pon 101 - 1.001/online
23.05.2024 13:00 - 16:00
English
Hybrid
Pon 101 - 1.001/online
23.05.2024 13:00 - 16:0023.05.2024 13:00 - 16:00
English
Hybrid
Supervisor:
Remus Teodorescu
Co-Supervisor:
Tomislav Dragicevic
Tamas Kerekes
Assessment Committee:
Associate Professor Szymon Bęczkowski, AAU (Chair)
Giovanni De Carne, Karlsruhe Institute of Technology
Masssimo Bongiorno, Chalmers University of Technology
Moderator:
Florin Iov
Abstract:
With the large-scale penetration of renewable energy sources interfaced with power electronic equipment, grid-forming (GFM) control emerges as a crucial component in providing grid support and ensuring power system stability by mimicking the behaviour of synchronous generators. Meanwhile, modular multilevel converters (MMCs) with excellent power quality, high efficiency, adaptability, and scalability are indispensable converter topologies in high-voltage conditions. Therefore, the application of GFM on MMC is the core unit to effectively maintain voltage and frequency stability in the future converter-dominate power system (CDPS). However, there are still some stable operation challenges of GFM-MMC that hinder the CDPS evolution.
In pursuit of enhancing the stability of GFM-MMC-connected power grids across various operational scenarios, the small signal analysis of the dominant factors impacting the stability is conducted by establishing the state space model of the GFM-MMC system. It reveals that factors such as small grid impedance and coupling effects of submodule capacitor voltages adversely affect GFM control, leading to undesirable low-frequency power oscillations. Building upon this understanding, this thesis initially presents the stability enhancement strategies for GFM-MMC, categorized into adaptive control and advanced control, depending on whether leveraging the stability estimation capability of AI or the GFM control structure is upgraded to account for MMC internal coupling.
To enable adaptive tuning control parameters without altering the original GFM controller structure, a neural network model is employed to assess the multiple-state of stability (SOS) defined based on the dominant eigenvalues calculated from the theoretical model. In this study, a three-level SOS is adopted, namely general stable, marginally stable and unstable. The assessment results aid in roughly determining the system stability range, and assist in adjusting control parameters to maintain stability and improve dynamics under various grid strengths.
In further simplifying the GFM-MMC operations, and eliminating the need for frequent control gain adjustments, the significant impact of multi-harmonic capacitor voltages on the GFM control is thoroughly analyzed using vector frames and DQ frames. For the vector-based GFM controller improvement, a correction method based on ABC-frame voltage reference is proposed, preserving the natural energy balance capability and effectively decoupling MMC internal circuit nonlinearity for stable operation with consistent control parameters across a wide range of grid strength. Additionally, to achieve real-time output voltage tracking without static errors while ensuring stability, the single voltage closed-loop approach that feeds forward only the fundamental component of the capacitor voltage in a DQ framework is introduced and its excellent dynamic performance is verified in the experiments with different operation conditions. Furthermore, in conjunction with inner current control, the achieved accurate voltage tracking control obtains one more degree of freedom to control output current, as verified through experiments.
Keywords: modular multilevel converters, grid stability assessment, small signal model analysis, modulation compensation, submodule capacitor voltage, short circuit ratio.