传感器性能退化与结构损伤响应异常区分研究

针对现行结构损伤检测方法仅考虑了环境和荷载对损伤的影响，未考虑传感器因时间和环境变化导致的性能退化从而严重影响损伤诊断率的问题，考虑到传感器性能退化的局部性及结构损伤的全局性，利用最小均方差估计与广义极似然比（GLRT）理论检测结构响应异常，结合统计质量控制图的控制指标超出界限的波动样式判断异常来源，并通过简支梁数值算例验证了本文提出的方法.研究结果表明:利用广义极似然比可以有效检测响应异常，利用控制图可以辨别响应异常来源；传感器性能退化导致控制指标超出界限的波动为非平稳波动，而损伤导致控制指标在一定范围内的波动为平稳波动. 

Distinguishing between Sensor Deterioration and Structural Damage

Current detection methods for structural damage only consider environmental or operational effects on damage detection, but neglect the problem that sensor performance degradation can be caused by environmental changes, and thus seriously affects the damage detection accuracy. In this paper, a method for distinguishing between sensor performance degradation and structural damage is proposed, which is based on the fact that the sensor performance degradation is local, while structural damage is global. In this method, the minimum mean square error and the generalized likelihood ratio test (GLRT) are used to detect and locate the abnormal response in a structural system. Besides, the modes of the control index fluctuation out of the limit of the statistical quality control chart are used to judge the source of abnormal responses. A numerical study is performed to validate the proposed method via a simple-supported beam. The results show that the GLRT can effectively detect the abnormal responses, and the statistical quality control chart is capable of distinguishing between the degradation sensor and damage accurately. If the control index fluctuation out of the limit of the statistic control chart shows a stable mode, the abnormal response comes from structural damage; otherwise, it comes from sensor performance degradation.