上海大学学报(自然科学版) ›› 2019, Vol. 25 ›› Issue (6): 870-878.doi: 10.12066/j.issn.1007-2861.2016

• 研究论文 • 上一篇    下一篇

叶片疲劳寿命神经网络近似计算模型数值实验

王雷1,2, 陆金桂1(), 李乐为2   

  1. 1. 南京工业大学 机械与动力工程学院, 南京 211800
    2. 南瑞集团有限公司, 南京 210061
  • 收稿日期:2018-01-24 出版日期:2019-12-30 发布日期:2019-12-31
  • 通讯作者: 陆金桂 E-mail:13852285790@qq.com
  • 基金资助:
    国家自然科学基金资助项目(50975133)

Neural network for fatigue life of wind turbine blade numerical calculation of approximate model

Lei WANG1,2, Jingui LU1(), Lewei LI2   

  1. 1. College of Mechanical and Power Engineering, Nanjing Tech University, Nanjing 211800, China
    2. NARI Group Corporation, Nanjing 210061, China
  • Received:2018-01-24 Online:2019-12-30 Published:2019-12-31
  • Contact: Jingui LU E-mail:13852285790@qq.com

摘要:

借助结构近似分析, 建立叶片疲劳寿命神经网络近似计算模型, 近似计算风力机叶片疲劳寿命性能. 进一步通过风力机叶片疲劳寿命神经网络近似计算模型的数值实验, 得到学习样本数、神经网络隐含层节点数、学习精度对风力机叶片疲劳寿命近似计算结果的影响关系, 以提高风力机叶片疲劳寿命值近似计算的准确率. 基于神经网络的风力机叶片疲劳寿命近似计算方法, 为风力机叶片疲劳寿命性能计算提供新的计算手段.

关键词: 叶片疲劳寿命, 结构近似分析, 神经网络模型

Abstract:

By means of structural approximate analysis based on neural network, approximate calculation model of neural network for fatigue life of wind turbine blade is established. And, the numerical experiments of different parameters for constructing the model of neural network for fatigue life approximation computation of wind turbine are made. The results show the influence of the number of learning samples, the number of hidden layers elements of the neural network, and the learning accuracy on the approximate calculation results of the fatigue life of wind turbine blades. It is helpful for improving the accuracy in calculating the fatigue life value of wind turbine blade based on the approximate model of the wind turbine blade fatigue life neural network. The approximate calculation method of wind turbine blade fatigue life based on neural network provides a new calculation method for calculating the fatigue life performance of wind turbine blades.

Key words: blade fatigue life, structural approximation analysis, neural network model

中图分类号: