Premium
Study on output prediction system of wind power generation using complex‐valued neural network with multipoint GPV data
Author(s) -
Kitajima Takahiro,
Yasuno Takashi,
Sori Hitoshi
Publication year - 2013
Publication title -
ieej transactions on electrical and electronic engineering
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.254
H-Index - 30
eISSN - 1931-4981
pISSN - 1931-4973
DOI - 10.1002/tee.21788
Subject(s) - wind power , artificial neural network , wind speed , computer science , electric power system , power (physics) , grid , nonlinear system , renewable energy , meteorology , engineering , artificial intelligence , electrical engineering , mathematics , physics , geometry , quantum mechanics
In recent years, wind power generation has seen rapid growth as a solution to the depletion of fossil fuels and to global warming. However, the output of the wind power generator fluctuates widely depending on the wind speed, which affects the power grid. In this paper, we propose a prediction system of wind power generation ahead of 24 h by using a complex‐valued neural network (CVNN). To predict the output of wind power generation, it is necessary to predict the wind speed accurately. Generally, wind data is expressed as a vector, which has both magnitude and direction. Therefore, it is also possible to treat wind data by a complex number and then use it as input information of the CVNN, which is very useful for operation of a complex number and nonlinear data. In computer simulations, we use multipoint numerical weather prediction data that surround the wind prediction point to realize long‐time prediction. Thereby, we expect that the CVNN can take into account wind dynamics in two‐dimensional space. Several simulation results and t ‐tests demonstrate the effectiveness of the proposed prediction system. © 2012 Institute of Electrical Engineers of Japan. Published by John Wiley & Sons, Inc.