摘要
基于静态结构识别方法,提出了服役结构损伤探测及状态评估的概率分析方法。首先通过有限元法把结构离散成用基本参数及核心矩阵表示的分析模型,根据两种分析模型和实际结构间的误差定义,用 Gauss-Newton法推导了在不完全测量情况下两种参数识别方法。在设计的测量情况下,进一步采用 Monte Carlo 法模拟测量数据,详细分析和比较这两种算法在有测量误差时的有效性和稳定性问题。第一种算法中的敏感矩阵不受测量误差的影响而具有比较好的性能,在确定了其识别结果的概率特性后,引入假设试验,探测和评估损伤的位置及程度。数字模拟显示了这一方法的有效性和准确性。
A method is proposed to detect damages and assess the status of an existing structure from static test data. The real structure is represented by finite element model that is described by the member constitutive: parameters and kernel matrix. Two algorithms by Gauss-Newton method for identifying those parameters based on two kinds of error definition under incomplete measurement are developed. By using Monte Carlo method, the effectiveness and stability of two algorithms are investigated when measurement errors exist. The first algorithm is distinguished in which the sensitivity matrix is not affected by the measurement error of displacements. The identified constitutive parameters are at the element level, and a statistical approach by a hypothesis test is introduced to locate and assess the damages of elements. Numerical examples are presented to demonstrate the validity and accuracy of the damage assessment algorithm.
出处
《工程力学》
EI
CSCD
北大核心
2004年第6期76-83,共8页
Engineering Mechanics
关键词
基本参数
敏感矩阵
系统识别
假设试验
状态评估
Computer simulation
Measurement errors
Monte Carlo methods
Structural loads
