基于径向基函数神经网络的风力机结构适应性评估

FLEXIBILITY EVALUATION OF WIND TURBINE STRUCTURE USING RADIAL BASIS FUNCTION NEURAL NETWORK

提出一种基于径向基函数神经网络(radial basis function,RBF)的风力机结构适应性评估方法,综合考虑多个载荷分量对风力机结构应力响应的影响。该新型适应性评估方法与当前采用的单变量、线性插值预测方法相比精度更高更可靠,与有限元方法相比更高效。采用该评估方法,1 min内可完成风力机机头部件及关键连接螺栓强度评估,最大应力预测误差不超过1%。

A new wind turbine structure flexibility evaluation method based on the radial basis function(RBF)neural network is presented in this paper.The effect of multiple load components on the stracture stress of the wind turbine is comprehensively considered to wind stress into account.Compared with the current single variable and linear interpolation prediction methods,this new flexibility evaluation method has higher precision and is more reliable.Compared with the finite element method,it is no doubt a more efficient way.The strength evaluation of wind head parts and key connection bolt can be completed in one minute by using this evaluation method,and the maximum stress prediction error is less than 1%.