ANN-based Estimation of MMC and Synchronous Machine Fast Active Power Response
G. van Putten (TU Delft - Electrical Engineering, Mathematics and Computer Science)
J.L. Rueda Torres – Mentor (TU Delft - Electrical Engineering, Mathematics and Computer Science)
M. Sharma – Mentor (TU Delft - Electrical Engineering, Mathematics and Computer Science)
A. Shekhar – Graduation committee member (TU Delft - Electrical Engineering, Mathematics and Computer Science)
P.A. Procel Moya – Graduation committee member (TU Delft - Electrical Engineering, Mathematics and Computer Science)
More Info
expand_more
Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.
Abstract
The increasing penetration of converter-interfaced renewable energy sources (RES) in modern power systems has significantly reduced synchronous inertia, posing challenges to short-term frequency stability. As a result, fast active power response (FAPR) from power electronic converters, such as modular multilevel converters (MMCs) connected to offshore wind power plants (WPPs), has become increasingly important. However, transmission system operators (TSOs) often lack access to proprietary converter control settings, making it difficult to assess the available frequency support capability of these resources in real time.
This thesis proposes a signal-record-based estimation method using an artificial neural network (ANN) to quantify the fast active power response of a mixed generation system consisting of synchronous generators and MMC-interfaced wind power plants. The method relies exclusively on measurable system signals, such as system frequency, rate of change of frequency (RoCoF), and pre-disturbance operating conditions, without requiring explicit knowledge of converter control strategies.
A comprehensive synthetic dataset is generated using detailed RSCAD-RTDS simulations of a multi-terminal offshore HVDC network connected to a reduced onshore AC system. The dataset captures a wide range of operating conditions by systematically varying key system parameters, including
Synchronous generator inertia, initial loading levels, and wind speeds at multiple offshore WPPs. Controlled load-step disturbances are applied to excite system frequency dynamics and corresponding fast active power responses. An ANN is trained to estimate the active power response trajectories of both the MMC and the synchronous generator following a disturbance. The results demonstrate that the proposed approach can accurately infer fast active power response characteristics from frequency measurements alone. This work provides a practical estimation tool that supports TSOs in assessing frequency support capability in converter-dominated power systems under uncertain and time-varying conditions.