基于修正模态的混合遗传算法结构损伤识别

采用模态参数对结构进行损伤识别时,测试模态参数包含的误差使识别结果受到影响,严重时甚至不能反映结构的实际破损情况.结构损伤检测可以作为一优化问题.为此提出一种基于修正测试模态的损伤识别方法,即将基于测试频率和修正后的测试振型组成的函数作为优化目标,由具有鲁棒性及易于处理非确定性信息能力的遗传算法和局部搜索算法组成的混合遗传算法作为优化工具.基于桁架的分析结果表明,即使在测试数据包含误差的情况下,采用该方法也能获得满意的识别结果.

Structural damage detection using updated modes based on hybrid genetic algorithm

When the dynamic modal parameters are used to detect the damage of structure, the identification always has the discrepancy for the measurement error, sometimes the practical damage of structure can not be identified. So the detection of structural damage is formulated as an optimization problem. An identification method based on updated test mode is presented. The optimization function uses the test frequency and the updated test vibration mode as optimized target, a hybrid genetic algorithm constructed with genetic algorithm and a compatible local search operator is also used for its robustness in coping with uncertainty and insufficient information. A truss is presented to demonstrate the effectiveness of the present method with good accuracy and high efficiency, even if the measurement is contaminated by measurement error.