论文部分内容阅读
随着大跨度桥梁数量及重要性的增加,桥梁损伤识别成为学界的研究热点。文中应用小波分析提取桥面的一阶振型,推导了梁式结构由不等间距位移模态向曲率模态转化的乘子矩阵DTOC,由损伤前后曲率模态的变异特征构造了梁式结构单损伤情况的普适概率神经网络,使之可以识别无训练样本时的损伤位置。通过数值仿真计算了某斜拉桥有限元模型桥面单元刚度折减前后在模拟脉动风荷载下的桥面板关键点的振动时程,讨论了小波参数识别技术和该普适概率神经网络的联合应用在不同级别的测量噪声影响下的斜拉桥桥面板损伤识别效果,说明在振动测量信号的信噪比达到一定要求时该方法具有实际效果。
With the increase of the number and importance of long-span bridges, bridge damage identification has become a hot research topic in academia. In this paper, wavelet analysis is used to extract the first-order mode of bridge deck. The multiplier matrix DTOC is derived from the modal transition from unequal-spacing mode to curvature mode. The beam structure is constructed from the variation of curvature modes before and after damage The universal probabilistic neural network of single damage conditions makes it possible to identify the damage location when no training sample is available. Through the numerical simulation, the vibration time history of the key points of bridge deck under simulated pulsating wind load before and after the deck element stiffness reduction of a cable-stayed bridge finite element model was calculated. The combination of wavelet parameter identification technology and the pervasive probabilistic neural network The effect of different levels of measurement noise on the cable bridge deck damage identification results show that this method has practical effect when the signal-to-noise ratio of the vibration measurement signal reaches certain requirements.