【作者】 李德春；

【导师】 练继建； 赵冲久；

【作者基本信息】 天津大学 ， 水利水电工程， 2013， 博士

【摘要】 恶劣的外部环境和复杂的作用条件,使港航水工结构在服役期间不可避免的产生损伤,损伤经过累积必然引起结构使用性能降低甚至发生破坏,威胁生命和财产安全。通过恰当的手段全面了解港口码头和航电枢纽等结构的“健康”状况,及时发现危及结构安全的隐蔽损伤已经势在必行。本文将结构无损动态检测技术和局部细化无损检测技术结合,研究基于港航水工结构荷载特性激励的模态参数识别方法,对港航水工结构无损检测新技术展开了较为系统的研究。(1)大部分港航水工结构在工作状态下均承受明显的水流脉动压力荷载激励,水流脉动压力在时间和空间上具有随机性,可用概率方法和统计函数来描述,其频率结构可用谱密度描述。港航水工结构水流脉动压力谱密度一般为多种噪声谱的组合,可近似为白噪声输入激励。(2)港航水工结构可简化为多自由度系统,可利用系统的状态空间模型和可控性、可观性条件,进行模态参数识别。在特征系统实现算法(ERA法)的基础上引入熵的概念,利用奇异熵技术解决了系统特征矩阵定阶难的问题。在频域分解法的基础上引入模态一致性函数,比较准确地确定了奇异值曲线的峰值,同时确定该阶模态起主要作用的优势频域带宽。基于奇异熵的ERA法和功率谱峰值法在高桩码头的模态参数识别中效果较好。(3)提出了一种综合保证测点应变模态能量和目标模态线性无关性的方法,采用逐步累加的方式求解改进的Fisher信息阵,得到同时满足模态可测性和强抗噪能力的传感器优化布置方案。基于QR分解和MAC准则优化传感器布置,并利用整数编码遗传算法使MAC矩阵的最大非对角元迅速降低。(4)对基于模态置信度、模态柔度、刚度变化、曲率模态、应变模态、变形曲率的结构损伤诊断方法理论和损伤敏感特征因子的计算方法进行介绍与分析,在此基础上进行了基于模态参数识别的高桩码头结构损伤评估研究。研究了基于LS-SVM与模态参数识别相结合的损伤诊断方法,并应用于大型枢纽工程导墙结构的损伤定位分析中。(5)提出一种基于整体无损动态检测与局部细化检测相结合的检测新技术。介绍了一种适用于港航水工结构的整体无损动态检测系统。系统研究了基于声波CT的结构损伤局部静态无损检测方法,并将声波CT检测技术应用到了新建码头的质量检测中。更多还原

【Abstract】 Under the abominable external environment and complex conditions, hydraulic structures in ports and navigation-power junction were damaged inevitably in their service period. With accumulation of damage, the structure performance was deteriorating, even cause destruction. The life and property safety were threatened. A comprehensive understanding the health conditions and timely discovered of hidden damage of hydraulic structures in ports and navigation-power junction through appropriate means has been already imperative. In this paper, the new non-destructive testing technology of hydraulic structures in ports and navigation-power junction was studied on the base of two technologies integrate: the non-destructive dynamic testing technology and the local-refinement non-destructive testing technology. The focus of the study is structural modal parameter identification under fluctuating pressures.(1) Under working state, most of hydraulic structures in ports and navigation-power junction beard significant fluctuating pressures. Fluctuating pressures was random in time and space, which could be described by the probability method and statistical functions. The frequency structure could be described by spectral density. The natural excitation load was generally combination of aforesaid noise, which could be approximated as white noise excitation.(2) Hydraulic structures in ports and navigation-power junction could be simplied to multi-degree of freedom system. The modal parameters could be identified on the base of the system state space model and the conditions of controllability and observability. The idea of entropy was introduce into Eigensystem Realization Algorithm(ERA). The singular entropy technology was used to solve the order determination problem of system characteristic matrix. The modal consistency function was introduced into Frequency Domain Decomposition(FDD). The peak of singular value curve and the dominant frequency domain bandwidth played a major role was accurately determined. The ERA and Peak Picking(PP) method showed good effect on the modal parameter identification of high-piled wharf.(3) A new optimal sensor placement method, which aims on the combination of modal strain energy and linear independence, has been investigated. The improved Fisher information matrix for effective independence method is solved by stepwise accumulation style, and the sensor placement scheme can ensure the modal measurability and anti-noise ability at the same time. Based on QR decomposition and Modal Assurance Criterion(MAC) and taking advantage of integer-coded genetic algorithm, the maximum off diagonal element of the Modal Assurance Criterion(MAC) matrix decreased rapidly.(4) The method theory of structural damage detection and damage sensitive feature were introduced and analyzed on the base of modal assurance criterion, modal flexibility, stiffness change, curvature mode, strain mode, deformation curvature. An experimental study on high-piled wharf structural damage assessment was conducted based on modal parameters identification theory. The damage diagnosis method is studied based on LS-SVM and modal parameter identification, and then applied to damage localization of a guide wall structure in a large hub project.(5) A new technique, which combines the global nondestructive dynamic detection and the local refinement detection, is proposed. The paper introduces a global nondestructive dynamic detection system for hydraulic structures in ports and navigation-power junction, systematically studies a structural damage local static nondestructive detection method based on sound wave CT, and applies detection technique of sound wave CT to the quality detection of a new wharf.更多还原