基于自适应模糊神经网络的风速软测量  

Wind Speed Soft Sensor Based on Adaptive Fuzzy Neural Network

作　　者：董海鹰[1] 魏占宏[1] 杨玺[1] 李晓青[1] 

机构地区：[1]兰州交通大学自动化与电气工程学院,兰州730070

出　　处：《电力系统及其自动化学报》2013年第1期60-65,共6页Proceedings of the CSU-EPSA

基　　金：甘肃省科技重大专项资助项目(0801GKDA058)

摘　　要：针对风力发电系统有效风速无法直接测量的问题,提出了基于自适应模糊神经网络的风速软测量方法,在该方法中利用优化的自适应模糊神经网络建立了风速软测量模型,以发电机功率、桨距角和风力机转速作为模型的输入,有效风速作为模型的输出,网络学习中采用可变的学习速率和可变的动量学习率。仿真结果表明,与传统的神经网络风速软测量模型相比,基于自适应模糊神经网络的风速软测量方法是有效的,风速的估计值较好地跟踪了有效风速的变化趋势,具有较高的准确性。For the effective wind speed of wind power generation systems can not be measured directly, wind speed soft sensor method based on adaptive fuzzy neural network was proposed, in which wind speed soft sensor modeling was established by adopting optimized adaptive fuzzy neural network. The model uses wind generator power,pitch angle and speed of wind turbine as the input, and effective wind speed as the output. Variable learning rate and mo- mentum rate are adopted in network learning. Simulation results show that compared with traditional neural network wind speed soft sensor model, the method of wind speed soft measurement based on adaptive fuzzy neural network is effective, and the trend of effective wind speed is better and more accurate followed by the estimates of effective wind speed,

关 键 词：风力发电 有效风速 软测量 自适应模糊神经网络 

分 类 号：TM614[电气工程—电力系统及自动化]