A computational intelligence framework for smart grid
With the recent development of electronics technique and distributed generations, such as wind power, solar power and electric vehicles (EVs), modern power system is advancing towards a critical and promising intelligent generation known as the smart grid. During the upgrade to this new generation, stability and security concerns have also been raised with complex communication and control challenges. Even worse, because of the new constraints placed by the environmental and economical concerns, the system planning and operation is toward maximum utilization of the existing infrastructure with tight operating and stability margins. The decreased system stability margin together with the increased penetration of renewable energy sources will bring new challenges to smart grid control, operation, stability and reliability. ^ Smart grid with conventional synchronous generators, renewable energy generation systems, flexible AC transmission system (FACTS) devices, and EVs are large-scale, nonlinear, nonstationary, stochastic and complex systems distributed over large geographic areas. The traditional control tools and techniques have limitations to control such complex systems to achieve an optimal performance. Therefore developing intelligent adaptive control and optimization systems for smart grid has become one of the critical research topics worldwide. Among many efforts toward this objective, machine learning and computational intelligence research provide the key technical innovations. Various aspects of intelligent and adaptive systems have been developed and improved in terms of learning and optimization capabilities based on reinforcement learning (RL), adaptive dynamic programming (ADP), and swarm intelligence. ^ To this end, this work focuses on the development of new architectures, frameworks and algorithms for smart grid optimal control and operation, such as energy storage based low-frequency damping control, islanded micro-grid frequency stability, doubly-fed induction generator (DFIG) low-voltage ride-though (LVRT) improvement, and optimal reserve scheduling in economic dispatch (ED) with wind power penetration. The proposed control and optimization methods are validated by simulation studies in Matlab/ Simulink. Results show that the significantly improved grid stability, reliability and dynamic performance.^
"A computational intelligence framework for smart grid"
Dissertations and Master's Theses (Campus Access).