基于DIC方法的混凝土结构裂缝开展全局识别与重构  

Detection and Measurement Method for Full-field Crack Propagation of Concrete Structure Based on DIC Technology

作　　者：晏班夫[1,2] 寇宇航 秦筵越 罗磊 李寿科 李水生[4] YAN Ban-fu;KOU Yu-hang;QIN Yan-yue;LUO Lei;LI Shou-ke;LI Shui-sheng(College of Civil Engineering,Hunan University,Changsha 410082,Hunan,China;School of Civil Engineering and Architecture,Guangxi University,Nanning 530004,Guangxi,China;School of Civil Engineering,Hunan University of Science and Technology,Xiangtan 411201,Hunan,China;China Construction Fifth Engineering Division Co.Ltd.,Changsha 410011,Hunan,China)

机构地区：[1]湖南大学土木工程学院,湖南长沙410082 [2]广西大学土木建筑工程学院,广西南宁530004 [3]湖南科技大学土木工程学院,湖南湘潭411201 [4]中国建筑第五工程局有限公司,湖南长沙410011

出　　处：《中国公路学报》2024年第3期283-297,共15页China Journal of Highway and Transport

基　　金：湖南省高新技术产业科技创新引领计划项目(2020GK2026);湖南省重点领域研发计划项目(2021GK2005);国家自然科学基金项目项目(51578227);广西区重点研发计划项目(2022AB23045)。

摘　　要：裂缝的全局动态识别对揭示混凝土结构的破坏机理十分重要。为解决传统单点位移及应变测试手段难以实现全局裂缝识别的问题,提出了一种基于数字图像相关(Digital Image Correlation, DIC)对混凝土结构裂缝开展全局识别与重构的方法。通过设置裂缝识别阈值,在DIC位移场中按水平方向逐行逐点识别表征开裂位置的位移突变点,随后对开裂位置两侧位移向量作差得到裂缝水平缝宽,再通过局部最小二乘拟合得到裂缝开展走向,根据水平缝宽与裂缝走向关系得到开裂处实际宽度。对位移场内所有目标点执行上述操作即可实现对裂缝形态的全局识别与动态重构。针对图像噪声问题,通过模拟试验验证了位移场降噪的必要性;采用均方根误差、信噪比与平滑度指标量化分析了不同降噪方法的去噪效果。针对实际工程中最优裂缝识别阈值的选取问题,引入信息熵与相关系数,基于缝宽分布熵值曲线与识别损失相关系数曲线提出了最优阈值的确定方法。最后通过超高性能混凝土加固混凝土梁加载试验进行验证。结果表明:该方法实现了混凝土结构裂缝开展过程的自动测量与可视化展示,与裂缝观测仪的对比测试结果表明缝宽测量误差在0.01 mm以内,满足工程要求。该研究为工程结构加载试验提供了一种非接触、可视化、精确的裂缝全局测量手段。The full-field dynamic identification of cracks is very important to reveal the failure mechanism of engineering structures.To address the problems that the traditional single-point displacement and strain measurement methods are difficult to achieve full-field crack detection,this paper proposes a method for full-field detection and reconstruction of concrete structure cracks based on digital image correlation(DIC).By setting the crack detection threshold,the displacement bouncing change point representing the crack location is detected line by line and point by point in the horizontal direction in the DIC displacement field,then the horizontal crack width is obtained by the difference of displacement vectors on both sides of the crack,and then the crack direction is obtained by local least squares fitting,and the crack width is obtained according to the relationship between the horizontal crack width and the crack direction.Global detection and dynamic reconstruction of crack geometry can be realized by global operation of the displacement field.Aiming at noise contamination to image,the necessity of noise reduction in the displacement field is verified by simulation tests.Mean square root error,signal-to-noise ratio and smoothness index are used to quantitatively analyze the noise removal effects of different methods.Aiming at the selection of the optimal crack detection threshold in practical engineering,the information entropy theory and correlation coefficient are introduced,and the determination method of the optimal threshold is proposed through the curve of crack width distribution entropy and the curve of identification loss correlation coefficient.Finally,the loading test of concrete beams strengthened with ultra-high performance concrete is carried out to verify the results.The results show that the proposed method realizes the automatic measurement and visual display of the crack development process of concrete structures.The comparison results with the crack observation instrument show that the meas

关 键 词：桥梁工程 混凝土结构 裂缝识别 数字图像相关 裂缝重构 位移场 降噪 

分 类 号：U446.2[建筑科学—桥梁与隧道工程]