基于MQ-WaveNet的智慧新能源大规模风力发电智能控制
The intelligent control of smart new energy large scale wind power generation based on MQ-WaveNet
王家坤 1司化涛 1王希转 1张媛 1赵伟平2
作者信息
- 1. 山东国华时代投资发展有限公司,山东 济南 250000
- 2. 北京金风慧能技术有限公司,北京 100176
- 折叠
摘要
为了降低外部干扰,确保电力运行安全稳定,文章提出了基于多视界分位数和小波神经网络(MQ-WaveNet)的智慧新能源大规模风力发电智能控制方法.通过构建智慧新能源大规模风力发电机组模型,计算捕获的风能和叶尖速数值,调整发电机的速度,获得最佳功率系数.将气压、风向、风速等参数输入小波神经网络,根据层与层之间的权重,得到隐含层与输出层功率值;结合多视界分位数构成MQ-WaveNet模型,计算每一分位点的发电预测结果,明确风力发电的时序特征.利用李雅普诺夫函数估计,计算风力发电滑模面变换和控制矢量,在多分位点范围内达到滑模面,实现风力发电状态智能稳定控制.通过实验证明,文章提出的模型能够提高风力发电机组抗干扰能力,保证设备智能稳定运行.
Abstract
In order to reduce external interference and ensure safe and stable power operation,a research on intelligent control of large-scale wind power generation based on MQ-WaveNet for smart new energy is proposed.By constructing a smart new energy large-scale wind turbine model,calculating the captured wind energy and blade tip speed values,adjusting the speed of the generator,and obtaining the optimal power coefficient.Input parameters such as air pressure,wind direction,and wind speed into a wavelet neural network,and obtain power values for the hidden layer and output layer based on the weights between layers;Combining multi view quantiles to form an MQ-WaveNet model,calculate the power generation prediction results for each quantile and clarify the temporal characteristics of wind power generation.Using Lyapunov function estimation,calculate the transformation and control vector of the sliding mode surface for wind power generation,reach the sliding mode surface within the range of multiple quantiles,and achieve intelligent and stable control of the wind power generation state.Through experiments,it has been proven that the studied model can improve the anti-interference ability of wind turbines and ensure the intelligent and stable operation of equipment.
关键词
小波神经网络/智慧新能源/大规模风力发电/发电机组控制/多分位点Key words
wavelet neural network/smart new energy/large scale wind power generation/generator set control/multiple quantiles引用本文复制引用
基金项目
国家能源投资集团有限公司创新创效项目成果转化信息化项目(100000371F210008)
出版年
2024
国家科技期刊平台
NSTL
万方数据