论文部分内容阅读
本文研究了基于不同结构动力参数的径向基网络损伤辨识方法,并结合简支梁损伤辨识进行了应用。研究发现:(1)径向基网络的输入参数选择对结果有较大影响;(2)使用模态曲率变化作为输入参数的网络辨识效果优于采用频率变化率的辨识效果;(3)综合使用频率变化率和模态曲率变化的网络辨识效果优于单独使用频率变化率或模态曲率的效果。结果表明,基于动力参数和径向基神经网络的结构损伤辨识方法能够准确地辨识结构损伤。
In this paper, the damage identification of radial foundation network based on the dynamic parameters of different structures is studied, and combined with the damage identification of simply supported beams. The results show that: (1) the choice of input parameters of radial basis network has a great influence on the result; (2) the network identification effect using modal curvature change as input parameter is better than that of using frequency change rate; (3) The effect of network identification using frequency change rate and modal curvature change is superior to that using frequency change rate alone or modal curvature alone. The results show that structural damage identification method based on dynamic parameters and RBF neural network can accurately identify structural damage.