【摘要】 针对传统的广义卡尔曼滤波算法不能有效地追踪结构刚度的变化情况,本文以传统卡尔曼滤波理论为基础,得到了基于衰减记忆的广义卡尔曼滤波算法公式,利用该算法对所得到的地震响应信号进行分析,提取结构的特性,辨识结构的参数,并从中判断结构损伤发生的时刻、位置及程度,改善了广义卡尔曼滤波的效果。衰减记忆的广义卡尔曼滤波算法只能够判断出结构参数变化的时间并且容易出现振荡,因此采用了一种新的自适应追踪技术,用一个自适应因子矩阵代替了原有的遗忘因子,这种技术可以有效追踪结构参数变化的时间、位置和大小,从而能够在线识别出结构的损伤。更多还原

【Abstract】 Considering that the traditional generalized Kalman filter algorithm can not effectively track the changes in stiffness,based on the traditional Kalman filter,this paper works out the formulas of the fading Kalman filter.This paper uses the algorithm to analyze the responses of the earthquake signals to extract the characteristics of the structure,to identify the structural parameters,and determine the time,location and extent of the structural damage.The algorithm improves the generalized Kalman filtering effect.But fading Kalman filter algorithm can only determine the time when the structural parameters change,and may be prone to oscillation.This paper uses a new adaptive tracking technology,with an adaptive factor matrix instead of the original forgetting factor.The technology can effectively track the changes in structural parameters of time,location and extent,and thus applies to on-line identification.更多还原