《框架结构失效元分析及损伤检测的综合研究》由会员分享，可在线阅读，更多相关《框架结构失效元分析及损伤检测的综合研究（134页珍藏版）》请在金锄头文库上搜索。
1、上海交通大学 博士学位论文 框架结构失效元分析及损伤检测的综合研究 姓名：王建民 申请学位级别：博士 专业：固体力学 指导教师：陈龙珠 20040801 上海交通大学博士学位论文 摘 要 I 摘 要 本论文针对框架结构系统主要失效模式分析中失效元的确定及结构损伤检测等相关问题 进行了分析和研究研究结果具有一定的理论和工程应用参考价值 本论文主要包括以下几方面 1. 第一章对结构可靠性分析及损伤检测的研究现状及关键问题进行了综述并简要介绍了 本论文研究的创新性和主要研究内容 2. 载荷增量法加载方式的改进第二章从理论上分析了结构系统静强度可靠性分析中利 用载荷增量法确定结构主要失效模式和极限承载
2、能力问题时原加载方式造成失效元漂移的可 能性及原因所在在此基础上对原载荷增量法的加载方式进行了合理改进 3. 基于矩阵摄动理论的频率灵敏度参数的分析第三章针对结构特征参数识别中的灵敏度分 析从矩阵摄动理论角度出发首先阐明结构各特征参数关于相关物理参数的灵敏度(即偏微分) 存在的前提条件然后对结构固有频率灵敏度参数进行了详细的理论分析和推导并针对结 构内的单构件损伤和多构件损伤提出相关的损伤检测方法 4. 框架的层水平位移变化增量参数在结构损伤检测中的应用第四章通过对框架结构的水平 柔度进行简化分析和量化表示给出相应的结构简化模型将构件损伤所引起的结构水平位移 变化特征进行局部化处理得到结构的层
3、水平位移变化增量(IOLDC)作为损伤识别参数同时 总结出结构内不同类型构件的损伤对结构水平位移变化产生不同的影响特征根据这些特征 通过 BP 神经网络实现对结构的损伤检测针对平面框架给出具体的检测识别方法 5. 基于模糊推理系统的损伤检测方法第五章根据上章分析所得到的框架结构内不同类型 构件的损伤与结构各层的水平位移变化增量之间典型的对应特征关系以框架的层水平位移变 化增量为损伤识别参数针对结构损伤中从损伤症状来看易混淆和模糊的损伤模式通过设 计模糊推理系统提出相关损伤检测方法 6. 基于结构位移变化特征规律的损伤检测分析第六章以空间框架为研究对象以结构的层 水平位移变化增量为损伤特征参数分
4、析了结构局部构件损伤与框架水平正交两个方向的层水 平位移变化增量之间的特征规律以水平正交两方向的各层水平位移变化增量为 BP 网络输入向 量对基于框架层水平基础上的构件损伤进行了检测识别然后再运用不同类型构件损伤与 层水平位移变化增量之间对应的特征规律对网络识别结果作进一步修正和完善 7. 损伤子结构频率变化特征及组合神经网络在结构损伤检测中的应用第七章将子结构模态 综合法应用到结构损伤检测中首先运用自由界面模态综合法对结构进行模态分析获得整 体结构及各子结构相应的模态信息并获得神经网络的训练样本集在节省大量计算工作量同 上海交通大学博士学位论文 摘 要 II 时通过组合神经网络将损伤子结构与
5、整体结构的模态频率变化相结合进行损伤检测一定 程度上可有效提高损伤检测结果的准确性 8. 损伤子结构频率变化特征在较复杂损伤模式识别中的应用第八章针对结构内较为复杂的 多构件损伤模式在运用神经网络进行结构损伤检测时首先用单层感知器识别损伤子结构 然后用最小误差匹配法近似得到损伤子结构模态频率变化再与整体结构频率变化相结合进 行损伤检测 关键词失效模式损伤检测载荷增量法灵敏度分析水平位移变化增量模糊推理BP 神经网络模态综合法 上海交通大学博士学位论文 摘 要 III Abstract This paper seeks to deal with the issues concerning the
6、 determination of main failure elements and the detection of structural damages. The results thus obtained in this dissertation are valuable for both theoretical research and engineering application. This dissertation contains eight chapters. Chapter One: This chapter briefs the current situations i
7、n research and the key issues in structural reliability analysis and damage detection. It also gives an introduction to the originality and the main contents of this dissertation. Chapter Two: Modification to the Incremental Loading Method. After the original loading model in the incremental loading
8、 method is theoretically analyzed in detail, the possibilities and reasons that make the failure elements move into the other structural members through the original loading model are explored. Furthermore, reasonable modifications to the loading model are made in the incremental loading method. Cha
9、pter Three: Sensitivity Analysis of Natural Frequencies in Structural Damage Detection Based on Matrix Perturbation Theory. On the basis of matrix perturbation theory, the definite condition for the existence of the sensitivity parameters of natural frequencies and model shapes is made clear. Then,
10、the sensitivity parameters of natural frequencies are analyzed and deduced in detail according to the matrix perturbation theory; and the damage detection method is proposed for the single-damage and multi-damage in structures. Chapter Four: Application of the Increment of Lateral Displacement Chang
11、e (IOLDC) to Damage Detection of Frame Structure. A simplified structure model is proposed to formulate the flexibility of frame structures. IOLDC, which is found to be very sensitive to the local damage in structures, is obtained. At the same time, it is found that the damage in different types of
12、members in structures generates distinctive influence on the lateral displacement of the frame. Based on these characteristics, a damage detection method, which can make effective identification of the damage in frame structures, is proposed. Chapter Five: Damage Detection Method Based on the Fuzzy
13、Inference System (FIS): Taking IOLDC as one damage identification parameter, this chapter proposes a new damage detection method based on the fuzzy inference theory (FIT), which aims at the fuzzy, indeterminate and complicated relationship between the damage symptom and damage reason in structure. T
14、he effectiveness in applying FIT in structural damage detection is verified through numerical examples. Chapter Six: Application of IOLDC to Damage Detection of Space Frame Structures: With regard to the space structures, distinctive rules by which the damage in different types generates characteris
15、tic influence on the lateral displacement are analyzed in this chapter. After IOLDC at the storey level of the frame along two normal lateral dimensions are selected as the input parameters of the neural network, a step-by-step damage detection method based on the BP neural network technique is prop
16、osed. The identification through the BP neural network can be further modified and improved according to the rules of the damage in different types of members corresponding to distinctive 上海交通大学博士学位论文 摘 要 IV influence models on the lateral displacement of frame structures. Chapter Seven: Application of Frequency Changes in Substructures and the Combined BP N
