Computational Intelligence in Smart Grid System: Application Cases of Particle Swarm Optimization in Renewable Energy System

• Main Authors: Yu-Shan Cheng, 鄭于珊
• Other Authors: Yi-Hua Liu
• Format: Others
• Language: en_US
• Published: 2017
• Online Access: http://ndltd.ncl.edu.tw/handle/39zg8h

博士 === 國立臺灣科技大學 === 電機工程系 === 105 === The transition from mitigating fossil fuel to clean and sustainable energy sources, such as solar power and wind power, involves the entire upgrade and the fundamental reengineering for current electricity grids. Smart grid is introduced under the growing consciousness of this grid evolution. Many challenges arise in terms of technology, economy and regulation when developing smart grid. Considering a high level of uncertainty and dynamism of smart grid system, computational intelligence is expected to solve problems in smart grid technology which are full of variables and complex scenarios. This dissertation aims to make a link between computational intelligence and smart grid system with two different application cases and figures out the advantages and effectiveness of the computational intelligence based approaches. For the given cases, one focuses on the optimum power dispatch within a microgrid, while the other proposes an optimized fuzzy controlled charging strategy for household photovoltaic (PV)-battery systems in a community. Both cases are based on particle swarm optimization (PSO), but they reveal different aspects of issues in smart grid system and the solutions have been developed accordingly. In the first case, to satisfy the power balance requirement which is interpreted as an equality constraint, a roulette wheel (RW) redistribution mechanism is integrated with PSO as a novel optimum power dispatch algorithm. In this way, the unbalanced power is able to be reallocated to more superior element in microgrid and the searching diversity can be preserved. On the other hand, in the second application case, it is assumed there are 74 houses installed PV-battery system individually. To cater distinct power profiles in each house within the community, a battery charging control strategy is designed to have adaptability to achieve minimum cost for houses without any meteorological or load forecasts. To sum up, it is anticipated that the introduction of the given cases can embody the computational intelligence based algorithms on the smart grid and present possibilities of a wide integration between intelligence and energy.