Application of Model Predictive Control

Abstract

This thesis explores the application of advanced control strategies to enhance the stability and performance of a three-area power system. Specifically, we investigate the efficacy of Linear Model Predictive Control (LMPC), Neural Network Predictive Control (NNPC), and Adaptive Model Predictive Control (AMPC) in addressing the challenges associated with multi- area power systems. The LMPC method is employed to provide a baseline control performance by utilizing a linear model of the power system. This approach is benchmarked against more sophisticated techniques to highlight its advantages and limitations. Following this, NNPC is introduced, leveraging the capabilities of neural networks to capture the nonlinear dynamics of the power system more accurately. The predictive control strategy is enhanced through the use of a trained neural network model, enabling improved handling of system uncertainties and nonlinearity. Finally, the AMPC approach is implemented, which dynamically adjusts the model and control parameters in response to real-time system changes. This adaptive strategy is particularly effective in dealing with variations in system dynamics and operating conditions, ensuring robust and resilient performance. Through extensive simulation studies and comparative analysis, the results demonstrate that while LMPC provides a solid foundation, NNPC and AMPC significantly outperform in terms of accuracy, adaptability, and overall system stability. The findings underscore the potential of neural network-based and adaptive control strategies in modern power system management, paving the way for their integration into practical applications. This thesis contributes to the field by offering a comprehensive analysis of these advanced predictive control techniques, providing valuable insights into their practical implementation and benefits for enhancing multi-area power system operations