Title :
Pitch angle control based on renforcement learning
Author :
Qin Bin ; Li Pengcheng ; Wang Xin ; Zhu Wanli
Author_Institution :
Acad. of Electr. & Inf. Eng., Hunan Univ. of Technol., Zhuzhou, China
fDate :
May 31 2014-June 2 2014
Abstract :
According to the random characteristics of external wind speed, time-varying of the internal unit parameters and nonlinearity of wind turbine system, a pitch angle control strategy based on reinforcement learning algorithm for wind turbine is proposed in this paper. The framework of Actor-Critic is adopted in this algorithm and RBF neural network is used to process continuous input and output space. With this algorithm the system can optimize its control parameter in time varying environment. The simulation results of wind power generation system show that the algorithm can quickly converge to the optimal value and has a good dynamic response and strong anti-disturbance.
Keywords :
learning (artificial intelligence); power engineering computing; power generation control; radial basis function networks; time-varying systems; wind power; wind turbines; RBF neural network; actor-critic framework; continuous input space; external wind speed random characteristics; output space; pitch angle control; reinforcement learning; time-varying internal unit parameters; wind power generation system; wind turbine system nonlinearity; Decision support systems; Quality function deployment; pitch angle control; reinforcement learning; wind turbine system;
Conference_Titel :
Control and Decision Conference (2014 CCDC), The 26th Chinese
Conference_Location :
Changsha
Print_ISBN :
978-1-4799-3707-3
DOI :
10.1109/CCDC.2014.6852110
