摘要
针对风速大于额定风速时风速波动引起风电机组的功率波动及变桨距系统频繁启停的问题,提出基于无迹卡尔曼滤波(unscented Kalman filter,简称UKF)与动态模糊神经网络(dynamic fuzzy neural network,简称DFNN)相结合的变桨距控制策略.为了消除传统变桨距控制中风速作为输入信号时产生的时延,将风电机组转速及输出功率作为反馈输入量.利用UKF对反馈输入量进行实时滤波处理,且将滤波后的数据用DFNN动态调整其权重,得到精确的桨距角指令值.采用Matlab/Simulink构建仿真模型,将UKF-DFNN控制与模糊PID、径向基函数(radial basis function,简称RBF)神经网络控制进行对比分析.仿真结果表明:所提策略能提高风速波动时系统的鲁棒性、抑制桨距角的波动范围、输出稳定的功率.
Based on unscented Kalman filter(UKF)and dynamic fuzzy neural network(DFNN),this paper proposed a variable pitch control strategy in terms of the power fluctuation and the frequent start and stop of pitch system problems,which were caused by wind speed fluctuations.In order to eliminate the time delay caused when wind speed was used as an input signal in traditional pitch control,the wind turbine speed and output power were introduced as feedback input.UKF was used to filter the feedback in real time,and the filtered data was dynamically adjusted with DFNN to adjust the parameter weights to generate accurate pitch angle command values.Finally,the simulation model was constructed in Matlab/Simulink.UKF-DFNN control was analyzed and compared with fuzzy PID and radial basis function(RBF)neural network control.The simulation results showed that the proposed strategy could improve the system robustness and suppress the range of pitch angle fluctuations and stabilize output power when wind speed changed suddenly.
作者
王江江
王维庆
王海云
萨妮耶·麦合木提
WANG Jiangjiang;WANG Weiqing;WANG Haiyun;MAIHEMUTI Saniye(School of Electrical Engineering, Xinjiang University, Urumqi 830047, China;Engineering Research Center of Ministry of Education for Renewable Energy Generation and Grid Connection Technology, Xinjiang University, Urumqi 830047, China)
出处
《安徽大学学报(自然科学版)》
CAS
北大核心
2021年第2期55-62,共8页
Journal of Anhui University(Natural Science Edition)
基金
国家自然科学基金资助项目(51667020)
教育部创新团队项目(IRT_16R63)
新疆维吾尔自治区重点实验室开放课题(2018D04005)
新疆维吾尔自治区高校科研计划自然科学重点项目(XJEDU2019I009)。
关键词
风力发电
变桨距控制
UKF
DFNN
wind power
variable pitch control
unscented Kalman filtering
dynamic fuzzy neural network