Machine Learning Based Intelligent Gateway for Collaboration System of Virtual Green Power Generation

• Main Authors: Tang, Sheng-Hsien, 唐盛賢
• Other Authors: Huang, Chung-Chi
• Format: Others
• Language: zh-TW
• Published: 2017
• Online Access: http://ndltd.ncl.edu.tw/handle/63v4x3

碩士 === 遠東科技大學 === 機械工程研究所 === 105 === In the thesis, a machine learning based intelligent gateway for the collabration system of virtual green power generation is proposed. It works for power dispatching of virtual green power plants. The intelligent gateway for the collabration system is developed by using supervising machine learning method. It can reduce the risk of power supply instability caused by poor loading distribution of green power plants. So it will effectively improve the performance of power dispatching. The intelligent gateway for the collaboration system contains: (1) Communication Module (2) Coordination Module (3) Cooperation Module (4) Shared Creation Module. Multiclass decision jungle algorithm is choosed for machine learning in the communication module and the coordination module. And supervised neural network regression algorithm is choosed for machine learning in the cooperation module and the shared creation module. These modules are integrated into the intelligent gateway for the collaboration system. Finally, the design of the intelligent gateway for the collabration system is verified by the results of simulation and implementation. The experimental results of the system showed the good performance and feasibility of power dispatching for virtual green power plants.