【摘要】 为探讨通过实测模态数据识别混凝土梁损伤的可行性,采用异常值分析方法,提出了通过判断实测低阶模态参数是否偏离正常状态来确定结构是否发生损伤的单参考点方法。针对结构状态连续监测的情况,提出了更具鲁棒性的逐次参考点异常值分析方法。预制了一根16m长的钢筋混凝土梁,并对健康状态和多种已知人工损伤状态下的梁结构进行了实验模态分析。实验数据的异常值分析结果表明:基于模态数据异常值分析的损伤识别方法,不仅能识别出钢筋混凝土梁的损伤,而且对损伤还很灵敏,甚至能检测出小于1%的微小损伤;逐次参考点分析方法能追踪钢筋混凝土梁累积损伤的过程。更多还原

【Abstract】 A single reference outlier analysis method was developed for damage detection in reinforced concrete beams with measured modal data.The measured lower order modal parameters were used so that the method need only identify the signal deviations from the normal condition from the statistical discipline of the outlier analysis.Then,a more robust successive references outlier analysis method was developed for damage detection for continuous monitoring of the structural state.The model was compared against data for a 16 m reinforced concrete beam with the modal parameters measured for the healthy state and 17 known artificial damage scenarios.The outlier analysis of the experimental data shows that these methods are not only able to detect damage in reinforced concrete beams,but are also sensitive to minute damage of less than 1%.The successive references outlier analysis method is able to track the cumulative damage in the reinforced concrete beam.更多还原