现役钢-混凝土组合结构界面损伤超声导波无基线识别及评价

Interface damage is the most serious damage type in reinforcd concrete composite structures. It has some obvious characteristics, for example,high hiding,strong random,clear localization.Ultrasonic guided wave(UGW) techniques are very suitable for the interface damage monitoring, however,conventional UGW methods compare signals obtained from the pristine condition of a structure (baseline signals) with those from the current state, and relate certain changes in the signal characteristics to damage.In this project the baseline-free interface damage dectection methods were proposed.First,corresponding numerical models based on rapid speed and high precision wavelet spectral element method were developed. Second,a fully automated baseline-free technique for interface damage detection in reniforced concrete was established,developing signal processing techniques to experimentally estimate dispersion curves in-situ. Third, due to the wave propagation environment being dispersive, multi-modal, and multi-path,a compensate and cancellation method and of dispersion ,multi-mode separation techniques were proposed. A data-driven framework for effective and robust analysis of guided wave signals is devised to automatically detect and localize interface damage.The sensed ultrasonic data were characterized and mapped onto a high dimensional feature space.Different wavefield images were processed to detect defects in reinforced concrete interface, and a nonlinear guided wave technique was applied to monitor interface early damage.Finally, the effectiveness and correctness of these methods in the project would be verified through experiment.

界面损伤是钢-混凝土组合结构最严重的破坏形式，它具有隐蔽性比较高、随机性强、局部化突出等特征，超声导波技术非常适合界面损伤监测，但是现有的超声导波大部分都依赖健康状态下的基线信号去识别结构损伤，很难真正用于现役工程。本项目提出无基线超声导波监测技术识别钢-混凝土界面损伤并进行评价。首先提出一种计算速度快、精度高的小波谱单元模拟钢-混凝土界面损伤波导传播。然后，建立钢-混凝土界面损伤超声导波无基线监测数学模型、提出基于实验方法实时获取结构频散曲线。其次，针对试验中导波信号频散、多模态、非线性等特点，发展频散补偿和移除方法、多模态分离技术。再次，在上述处理后的超声导波信号基础上，运用数据驱动技术提取损伤特征信号，构建特征数据库，自动识别界面损伤；并发展界面不同损伤类型超声导波成像方法、界面早期损伤非线性超声导波监测。最后，通过试验验证本项目提出的方法有效性和正确性。

界面损伤是钢—混凝土组合结构最严重的破坏形式，它具有隐蔽性比较高、随机性强、局部化突出等特征，超声导波技术非常适合界面损伤监测，但是现有的超声导波大部分都依赖健康状态下的基线信号去识别结构损伤，很难真正用于现役工程。本项目提出了基于波束成型的Lamb波无基线钢—混凝土界面损伤结构损伤识别，实现了基于概率的Lamb波无基线板状结构损伤识别，并开发了双曲线概率成像算法。为了克服小波有限元需要满足严格的时间步长的缺点，发展一种基于快速傅里叶变换（FFT）的BSWI有限元频域谱方法。该方法在空间域上利用紧支小波尺度函数近似理论波解，计算与频率相关的近似动力刚度矩阵。基于单元层面的动力缩聚技术被用于消除小波单元的内部节点自由度，从而得到矩阵尺寸非常小的等效动力刚度矩阵进行整体动力刚度矩阵的组装。提出基于实验方法实时获取结构频散曲线；利用频散和非频散波的传播方程以及频散波重构方程，发展频散补偿和频散移除方法；对多模态信号进行分解；并在建立的理论模型上，研究适用于导波信号的信号处理方法。提出了强噪声环境中提取有效导波信号及去噪方法；发展多模态导波信号分解方法，提取不同模态特征参数，建立特征参数与模态特征的对应关系；分析不同导波模态检测不同类型缺陷的适用性和有效性。开展三种不同类型界面损伤试验，验证了本项目所提方法的有效性和实用性。