Abstract
This research addresses the critical need to develop enhanced Automatic Voltage Regulation (AVR) in traditional coal-fired power plants in the context of South Africa's ongoing electricity crisis. Traditional Proportional-Integral-Derivative (PID) controllers are widely used in AVR systems. These controllers rely on fixed gain parameters, resulting in inadequate response times, poor voltage recovery under non-linear conditions and significant voltage oscillations. Enhanced PID controllers offer potential improvements but raise concerns regarding their complexity, ease of parameterisation, and computational demands. Optimisation algorithms and auxiliary controllers, such as state and disturbance observers, are integrated into the AVR control framework to improve the effectiveness and robustness of AVR controllers. While optimisation improves voltage regulation, it can face challenges such as becoming trapped in local optima, slow convergence rates, and inefficiencies in managing non-linearities and complex dynamics during extreme load fluctuations and disturbances. Additionally, complex optimisation algorithms increase the computational load of the control framework and are not practically feasible. Auxiliary control may further complicate and burden the AVR framework if not designed and implemented correctly.
This study identifies these gaps and proposes a novel control framework integrating an Extended Proportional-Integral-Derivative with Acceleration (ePIDA) controller with a three-degree-of-freedom (3DOF) structure, coupled with an Extended State Observer-based Disturbance Observer (ESO-DOB). In addition, this research introduces the novel Integral Square Time Absolute Error (ISTAE) objective function to guide the optimisation process. The study investigates two advanced optimisation techniques, the Salp Swarm Optimiser (SSO) and the newly developed Snake Optimiser (SO). These novel control strategies were rigorously validated through extensive performance and sensitivity simulation analyses under various operational conditions, including extreme load fluctuations and parameter uncertainties.
Key findings from the research demonstrate that the ePIDA controller, with its 3DOF structure, significantly enhances voltage regulation by effectively managing load fluctuations, disturbances, non-linearities, and system uncertainties. Integrating the ESO-DOB provides real-time disturbance estimation, enabling pre-emptive control actions that improve system stability and robustness. The ISTAE objective function refines the control strategy by prioritising rapid error correction and
vi
maintaining steady-state accuracy. Comparative analyses reveal that the SSO, while simple and effective under normal conditions, shows sensitivity to extreme system dynamics, potentially limiting its practical application.
In contrast, the SO-ePIDA strategy outperforms all other tested methods, offering superior speed, accuracy, and resilience under challenging conditions, including +30% voltage variations and ±50% parameter fluctuations. The SO-ePIDA framework, combined with the ESO-DOB, provides a comprehensive and robust solution for AVR applications, ensuring stable and precise voltage regulation in coal-fired power plants facing severe operational challenges.
This research contributes significant theoretical and practical advancements in AVR systems. It offers a robust and efficient control strategy that addresses the complexities of modern power systems, thereby providing a viable solution for enhancing the stability and reliability of power generation in South Africa, which may ease the current energy crisis.