基于神经网络的矮塔斜拉桥近断层地震响应参数敏感性分析*

为研究近断层地震动对矮塔斜拉桥地震响应参数敏感性的影响，提出基于BP神经网络配合有限元分析的参数敏感性分析方法（FBSA法），以某铁路矮塔斜拉桥为研究对象，进行参数化有限元建模，运用BP神经网络计算近断层地震动作用下矮塔斜拉桥响应以及自振频率对各参数的敏感性系数，分析近断层脉冲地震动对地震响应敏感性的影响规律。研究结果表明:FBSA法可快速、精确地完成敏感性分析，为抗震设计与优化提供更全面的参考；矮塔斜拉桥几何布置和截面尺寸对振型刚度影响较大；近断层脉冲地震动作用下，桥梁位移和内力响应明显增大，各设计参数对桥塔位移、桥墩位移和内力以及主梁横向位移和弯矩的参数敏感性具有显著影响，对主塔内力和主梁竖向位移参数敏感性的影响不明显。

Sensitivity analysis on near-fault seismic response parameters of extradosed bridge based on neural network

In order to study the influence of near-fault ground motion on the sensitivity of extradosed bridge’s seismic responses parameters,a parametric sensitivity analysis method based on BP neural network combined with finite element analysis (FBSA method) was proposed.Taking a railway extradosed bridge as the research object,the parametric finite element modeling was carried out.The BP neural network was used to calculate the sensitivity coefficients of extradosed bridge’s response and natural frequency under near-fault ground motion to each parameter,and the influence law of near-fault pulse ground motion on the seismic response sensitivity was analyzed.The results showed that the FBSA method could complete the sensitivity analysis rapidly and accurately,and provide a more comprehensive reference for seismic design and optimization.The geometric layout and section size of extradosed bridge had significant impact on the modal stiffness.Under the near-fault pulse ground motion,the displacement and internal force response of the bridge increased obviously,and the design parameters had significant impact on the parameter sensitivity of the bridge tower displacement,displacement and internal force of bridge pier,and transverse displacement and bending moment of main girder,whereas the effect on the parameter sensitivity of the internal force of main tower and vertical displacement of main girder was unremarkable.