【摘要】 对简支梁进行损伤分析,研究不同损伤工况下的频率变化率和模态振型曲率变化,并采用径向基神经网络对结构进行损伤识别研究。研究中分别采用频率变化率、第1阶模态曲率变化、综合使用前3阶频率变化率和模态曲率变化3种方案。结果表明,基于动力参数和径向基网络的结构损伤识别方法能够准确识别结构的损伤程度;神经网络的输入参数选择对结果有较大影响,综合使用频率变化率和模态曲率变化方案的识别效果最好。更多还原

【Abstract】 Damage analysis of simply supported beam is presented,involving rate of frequency change and change of mode shape curvature at different kinds of scenarios.RBF neural networks are adopted to identify the structural damage.Three schemes,which choose different parameters as input of networks,were used in the damage identification process: rate of frequency change,the first order mode shape curvature change and the combination of the first three orders of rate of frequency change and change of mode shape curvature.Results show that the structural damage can be identified by using RBF networks and vibration characters,and the input parameters of networks do influence the identification effect.The combination of rate of frequency change and mode shape curvature change makes the best results.更多还原