基于振动特性的隐蔽工程结构损伤识别方法研究

发布时间：2018-04-20 00:12

本文选题：隐蔽工程结构 + 损伤识别方法　；参考：《上海大学》2014年博士论文

【摘要】：土木工程结构一般可分为两大部分：一部分是可以直接接触和观测的部分,称之为暴露工程结构；另一部分是无法直接接触和观测的部分,称之为隐蔽工程结构。隐蔽工程结构是大型土木工程结构的重要组成部分,在整个结构中具有极其重要的地位。但由于隐蔽工程结构位置比较隐蔽,其损伤不易被识别,已成为土木工程结构不可忽视的潜伏安全隐患。如果能够及早识别出隐蔽工程结构的损伤(特别是早期损伤),就可以对整个工程结构进行比较准确的安全评估,从而可以避免巨额基础设施蒙受重大的损失。因此,如何对隐蔽工程结构进行损伤诊断,包括确定隐蔽工程结构是否存在损伤、结构的损伤位置、损伤程度和损伤变化趋势,成为大型工程结构健康监测和安全评估系统最重要的问题。为此,本文在浙江省教育厅科研项目“环境激励下的结构损伤与荷载同步识别方法研究”(项目批准号：Y201223039)和桥梁工程安全控制技术与装备湖南省工程技术研究中心开放基金项目“基于支持向量机法桥梁冲刷损伤识别方法及其应用研究”(项目批准号：13KC05)的资助下对基于振动特性的隐蔽工程结构的损伤识别方法开展了相关探索性的研究,主要对隐蔽工程结构损伤识别方法的基本理论进行了研究,重点应用高效模态应变能、小波分析、子结构、时变自回归滑动平均模型、支持向量机和自适应卡尔曼滤波等对隐蔽工程结构损伤诊断方法进行了探索性的研究。论文的主要研究工作如下： (1)建立了基于高效模态应变能的隐蔽工程结构损伤识别方法。该方法摒弃了使用连续低阶模态对结构进行损伤识别的传统方法,而选用可不连续的高效模态进行结构损伤识别。首先,详细推导了根据结构损伤前后特征值变化率来选择高效模态的理论方程,选择的高效模态具有对损伤前后结构的单元应变能变化灵敏度较高的特点。在此基础上,建立高效模态下隐蔽工程结构的损伤判别指标和损伤识别方程；最后,对高效模态的损伤识别灵敏度和鲁棒性进行了研究。为验证该方法的有效性,对一海洋结构中隐蔽工程结构单一损伤和多处损伤在不同噪音工况下的损伤诊断效果进行了深入研究。 (2)建立了基于小波分析的隐蔽结构损伤识别方法。该方法把小波分析技术与高效模态有机结合起来对结构小损伤进行诊断,弥补了单一损伤动力指纹对结构损伤识别可能产生漏判和精度不能保证的不足。主要应用单元模态应变能对结构小损伤的敏感性、小波分析技术高效的局部分析能力对结构小损伤位置进行识别。在此基础上,应用损伤识别方程对隐蔽结构的损伤程度进行诊断。为了验证该方法的有效性,应用数值模拟和工程实例对该方法的识别程序、识别效果、影响因素等进行了深入研究。 (3)建立了基于相对子结构界面参数的隐蔽结构损伤识别方法,该方法主要利用相对子结构界面参数对结构损伤的敏感性来识别隐蔽工程结构的损伤。该部分重点对结构界面参数识别理论、隐蔽结构损伤和子结构界面参数变化的对应关系、隐蔽工程结构损伤位置和程度识别、识别方法的鲁棒性等进行了深入研究。 (4)建立了基于ARMA和SVM的结构损伤识别方法。该方法结合了时变自回归滑动平均模型(ARMA)与支持向量机(SVM)优点,解决了环境激励下结构健康检测信号存在非平稳性和损伤样本非常有限的问题。重点对ARMA与SVM相结合的损伤识别理论、大型工程结构在环境激励下隐蔽工程结构损伤识别程序、识别精度和该方法的鲁棒性等进行了深入研究。 (5)建立了基于自适应卡尔曼滤波的子结构损伤识别方法。该方法结合了自适应卡尔曼滤波和子结构的优点,解决了环境激励下单独应用相对子结构界面参数无法对隐蔽工程结构和暴露工程结构同时识别的问题。主要对自适应卡尔曼滤波算法理论、利用自适应卡尔曼滤波对子结构损伤诊断方法、遗忘因子确定方法、用该方法进行损伤诊断的识别精度和鲁棒性等问题开展了探索性研究。
[Abstract]:A civil engineering structure can be divided into two parts: part of which can be directly exposed and observed, which is known as exposure engineering structure, and the other is a part that can not be directly contacted and observed, which is called a concealed engineering structure. But because the location of the concealed engineering structure is relatively concealed and its damage is not easily recognized, it has become a latent security hidden danger in civil engineering structure. If it can identify the damage of the hidden engineering structure (especially early damage), it can make a more accurate safety assessment of the whole engineering structure. Therefore, it is the most important problem for the health monitoring and safety assessment system of large engineering structures to diagnose the damage of the hidden engineering structure, including the determination of the damage of the concealed engineering structure, the damage position of the structure, the degree of damage and the trend of damage change. "Study on the method of structural damage and load synchronization identification under the environment inspired by environmental incentives" (project approval number: Y201223039) and bridge engineering safety control technology and equipment of the Hunan Province Engineering Technology Research Center open fund project "bridge scour identification method and its application based on support vector machine method" in Zhejiang Province education department research project Research on the damage identification method of concealed engineering structure based on vibration characteristics (project approval number: 13KC05), the basic theory of the damage identification method of concealed engineering structure is mainly studied. The emphasis is on the application of high efficient modal stress, wavelet analysis, substructure, and time-varying autoregressive sliding. The average model, support vector machine (SVM) and adaptive Calman filter (adaptive) and other methods have been studied. The main research work of this paper is as follows:
(1) the damage identification method of concealed engineering structure based on high efficient mode strain energy is established. The method abandons the traditional method of using continuous low order mode to identify the damage of structure, and selects structural damage identification with discontinuous efficient modes. First, the selection of the change rate of eigenvalues before and after the structural damage is derived in detail. The theoretical equation of high efficiency mode and the high efficient mode selected have the characteristic of high sensitivity to the change of element strain energy before and after the damage. On this basis, the damage identification index and damage identification equation of the concealed engineering structure under high efficiency mode are established. Finally, the sensitivity and robustness of the damage identification of the high efficiency mode is studied. In order to verify the effectiveness of this method, the damage diagnosis effect of the single damage and multiple damage of the hidden engineering structure in a marine structure under different noise conditions is studied in depth.
(2) a method of identifying the damage of hidden structures based on wavelet analysis is established. This method combines the wavelet analysis technique with the efficient mode to diagnose the small structural damage, and makes up for the deficiency of the single damage dynamic fingerprint on the structural damage identification. The sensitivity of small damage, the local analysis ability of the wavelet analysis technique to identify the small damage position of the structure. On this basis, the damage identification equation is used to diagnose the damage degree of the hidden structure. In order to verify the effectiveness of the method, the recognition program of the method is applied to the recognition program of the method and the engineering example. Fruit, influencing factors and so on have been studied in depth.
(3) a hidden structure damage identification method based on the relative substructure interface parameters is established. The method mainly uses the sensitivity of the relative substructure interface parameters to the structural damage to identify the damage of the hidden engineering structure. This part focuses on the theory of the identification of the structural interface parameters, the damage of the hidden structure and the corresponding changes in the interface parameters of the substructure. The relationship between concealed engineering structure, location and degree of damage, and robustness of identification methods are deeply studied.
(4) a structural damage identification method based on ARMA and SVM is established. This method combines the advantages of time-varying autoregressive moving average model (ARMA) and support vector machine (SVM). This method solves the problem of non-stationary and very limited damage samples of structural health detection signals under environmental excitation. The emphasis is on the theory of damage identification combined with ARMA and SVM. The damage identification procedure, recognition accuracy and robustness of the engineering structure under Hidden environment are studied in depth.
(5) a substructure damage identification method based on adaptive Calman filter is established. This method combines the advantages of adaptive Calman filter and substructure, and solves the problem of identifying the hidden engineering structure and the exposed engineering structure at the same time by using the relative substructure interface parameters under the environment excitation. The main problem is adaptive Calman. The theory of filtering algorithm, using adaptive Calman filter to diagnose substructure damage, and the method of forgetting factor determination, carry out exploratory research on the recognition accuracy and robustness of damage diagnosis by this method.

【学位授予单位】：上海大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TU317

【参考文献】