配筋混凝土缺陷无线超声波阵列聚焦成像与智能识别研究

Common defects of reinforced concrete (e.g. hole et al.) have adverse effects on the strength and durability of concrete structures, which would lead to a potential risk of structural failure under extreme circumstances. Therefore, it is critical to obtain comprehensive information regarding defects in the reinforced concrete, which is also critical for the accurate quality evaluation of concrete structures. Based on this, this proposal intends to carry out the research from the aspects of effective defect detection, synthetic aperture focusing imaging, intelligent defect recognition and quantitative estimation by the digital image method. Details of this study are structured as follows. First of all, the propagation mechanism of ultrasonic wave in reinforced concrete under coupled effects of multi-factors will be studied at a micro-scale level, and this study will be conducted by both experimental techniques and numerical simulation. Based on the mechanism study above, a collaborative optimization algorithm of key parameters of ultrasonic excitation and receiving will be established to achieve effective defect detection. Secondly, the mechanism study of synthetic aperture focusing imaging will be carried out, and the influence mechanism of ultrasound echo signal characteristics on the imaging resolution will be revealed. Based on the research, a real-time imaging resolution optimization at low wireless transmission rate will be carried out on the basis of the targeted optimization of ultrasonic signals collected by an array of sensors. In addition, the reliability and accuracy of the algorithm will be verified through experiments with the array of wireless ultrasonic sensors. In the end, an intelligent defect recognition method will be established using the convolution neural network. In addition, the image characteristics of defects will be extracted automatically through designed algorithm, and a quantitative evaluation of defects based on will be achieved based on image characterization parameters. This study will lay theory foundation for further study of concrete quality control and life-cycle health monitoring of structures. In addition, it can also play an important constructive role in engineering practice.

孔洞等配筋混凝土常见缺陷对结构强度与耐久性产生不利影响，带来极端情况下结构失效的潜在风险。因此获得配筋混凝土内部全面缺陷信息，为结构质量评估提供准确依据显得尤为重要。本项目拟从缺陷有效感知、聚焦成像、智能识别与图像定量评价入手，开展以下三方面研究：(1)实验与数值模拟结合，从细观层面研究多因素耦合作用下钢筋混凝土介质中超声波传播机理，通过协同优化超声波信号激励与采集布置核心参数，实现缺陷有效感知；(2)开展合成孔径聚焦成像机理研究，揭示超声回波信号特征对成像分辨率影响机制，通过针对性优化阵列采集信号，提高无线低传输速率下实时成像分辨率，并采用无线超声波阵列传感器试验验证；(3)基于深度学习卷积神经网络，研究超声波缺陷成像智能识别与特征提取方法，实现基于图像表征参数的配筋混凝土缺陷量化评价。研究成果为混凝土质量科学控制和结构全寿命健康监测进一步奠定理论基础，对工程实践具有重要的指导作用。

孔洞等常见缺陷对土木工程结构的强度与耐久性产生不利影响，带来结构在极端情况下失效的潜在风险。因此定量评价损伤的尺寸，为评估提供准确依据显得尤为重要。本项目从损伤有效超声波感知、超声波与缺陷相互作用机理和定量评价三方面入手，开展以下研究工作：（1）缺陷作用下超声波传播规律研究：研发无线超声波传感系统，采用实验方法研究不同尺寸缺陷对超声波传播规律的影响，结合数值仿真方法，建立对应有限元分析模型，开展超声波传播规律研究，对不同到达时间的采集超声波子波信号归属进行分析；（2）无线采集信号预处理方法研究：开展了基于sinc方程的波形增采样重建方法和超声波采集信号带通滤波方法研究，研究结果表明，提出的方法可有效提高超声波采集信号的信噪比；（3）缺陷定量识别方法研究：结合不同到达时间超声波到达子波的来源分析，确定需要分析的关键子波组分；开展损伤敏感因子提取方法研究，通过基于离散希尔伯特—黄变换的超声波波形信号包络检波，获取更精准的子波幅值信息；结合损伤尺寸信息，建立基于超声波采集信号包络线特征参数的损伤定量模型，通过在不同类似的疲劳损伤试件中应用，验证了方法的正确性。研究成果可为严酷环境下混凝土结构全寿命周期性能智慧感知与劣化预警相关研究提供技术支撑。