MODEL-FREE CONTROL DESIGN FOR PERMANENT MAGNET SYNCHRONOUS GENERATOR IN WIND ENERGY CONVERSION SYSTEMS
Loading...
Date
2023-11-30
Authors
Zholtayev, Darkhan
Journal Title
Journal ISSN
Volume Title
Publisher
School of Engineering and Digital Sciences
Abstract
This research, conducted as part of a PhD program at Nazarbayev University, investigates advanced control methods for electric engines, with a focus on Wind Energy Conversion Systems (WECS), in pursuit of sustainable and efficient energy solutions. Electric machines permeate many facets of our daily lives, ranging from household appliances to industrial applications, making their efficient management a necessity. WECS, which are characterized by inherent nonlinearity, present challenges that necessitate sophisticated control strategies for efficient power harvesting, assuring power quality, and minimizing system wear and tear.
The investigation of an advanced model-based control method, Super-Twisting Adaptive Sliding Mode Control (ST-ASMC), for Permanent Magnet Synchronous Generator (PMSG)-based WECS is the focus of this research. ST-ASMC is presented as an effective solution, preserving the robustness characteristics of conventional sliding mode control while reducing chattering via gain adaptation and the generation of second-order sliding modes. ST-ASMC facilitates optimal power acquisition by resolving a nonlinear multi-input multi-output tracking control issue. The proposed control method outperforms other sliding mode control techniques in the presence of variations in stator resistance, stator inductance, and magnetic flux linkage, as determined by comparative simulation studies employing actual wind speed data.
In addition, this thesis introduces a novel context for implementing a model-free control method based on the Twin Delayed Deep Deterministic Gradient Descent (TD3) technique within the context of PMSG-based WECS. This method is effective for adapting to dynamic uncertainties and disturbances in the WECS. TD3 is effective, stable, and robust, leveraging deep neural networks and reinforcement learning to incrementally improve decision-making. Notably, this model-free control method requires minimal understanding of the wind power conversion system, relying only on wind speed, turbine diameter, tip-speed ratio, electromagnetic torque, and stator direct current. In other words, for the first time in the context of WECS deep reinforcement learning (DRL) was implemented as the sole controller for the machine side. Through exhaustive evaluations, the TD3-based Maximum Power Point Tracking (MPPT) algorithm demonstrates adaptability and efficiency under varying conditions and system parameter fluctuations, validating its potential to enhance the performance of WECS for sustainable development and, in certain metrics, outperforming the Linear Quadratic Regulator (LQR) model.
In conclusion, this thesis advances the field of control methods for electric drives in WECS by proposing and validating advanced control strategies that can considerably contribute to the efficiency and sustainability of wind energy harvesting and conversion systems.
Description
Keywords
Research Subject Categories::INTERDISCIPLINARY RESEARCH AREAS, Type of access: Embargo
Citation
Zholtayev, D. (2023). Model-free control design for permanent magnet synchronous generator in wind energy conversion systems. School of Engineering and Digital Sciences