【作者】 刘娟；

【导师】 黄维平；

【作者基本信息】 中国海洋大学 ， 港口、海岸及近海工程， 2003， 硕士

【摘要】 本文论述了结构损伤诊断中的传感器优化配置及诊断方法的研究进展；讨论了遗传算法的发展现状、基本思想和特点；重点研究了海洋平台的传感器优化配置和损伤诊断。 从各阶模态向量的相关性和变形能出发，讨论了模态置信度和变形能两种传感器优化配置的准则；在原有基于QR分解的逐步累积法基础上，提出了修正的逐步累积法，直接将所有QR分解后的剩余自由度按照对目标函数贡献最大既累积的原则进行优化，得到了较好的优化效果；基于逐一削去能使目标函数减小最快的自由度的原则，提出了逐步削减法，并指出传感器优化配置必须结合经济和优化效果两方面考虑；将遗传算法应用于传感器优化配置，针对遗传操作中出现的约束条件的满足问题，提出二重结构编码遗传算法，并将局部优化方法(累积法、削减法)和全局优化方法(遗传算法)所得结果作了比较，证明了遗传算法的优越性；给出各种情况下海洋平台模态试验中的传感器配置方案。 以实测的固有频率和振型为诊断依据，将损伤诊断问题归结为优化问题，采用遗传算法加以研究，以10单元固端梁的损伤诊断为基础，讨论了诊断中出现的损伤指标，损伤对称性和变量精度等问题；将遗传算法的损伤诊断应用于海洋平台，验证了遗传算法用于大型结构损伤诊断的可能性；并用MATLAB将遗传算法应用于损伤诊断的全过程编制成损伤诊断系统。更多还原

【Abstract】 This paper discusses the development and the methods of sensor placement and damage detection, and the basic theory and the characteristic of genetic algorithm. It focuses on the sensor placement and the damage detection of offshore platforms.Two optimal principles of sensor placement, Modal Assurance Criterion (MAC) and Displacement Energy, have been discussed based on the relativity of modes and the displacement energy. Updated cumulative method, which optimize the objective by cumulating directly all the freedoms left by QR decompose according to their contribution to the objective, has been proposed. Reduced Method has been proposed based on the principle of reducing the freedom affecting the objective the most. It is suggested that both optimal result and economy must be considered in a sensor placement; Applies genetic algorithm to the sensor placement and presents dualistic coding genetic algorithm aiming at restriction problem in genetic operation; Comparing the results getting from partial and global optimal methods proves superiority of genetic algorithm and presents each sensor placement schemes of offshore platforms.Taking natural frequency and mode as damage principle, damage detection is studied as optimal problem based on genetic algorithm, and damage indices, symmetry and precision in the damage detection is discussed. Numerical simulation of the damage detection of offshore platforms using genetic algorithm proves the feasibility of genetic algorithm in damage detection of large structure; Program all the process of geneticalgorithm in damage detection into damage detection system.更多还原