A STATISTICAL MODEL UPDATING METHOD OF BEAM STRUCTURE USING STATIC MEASUREMENT DATA WITH NON-GAUSSIAN ERRORS

作者：Wu Zhifeng^*; Huang Bin

来源：14th International Symposium on Structural Engineering (ISSE-14), 2016-10-12 to 2016-10-15.

• model updating
• random parameter
• static condensation
• multivariate power series
• measurement error

In the operation stage of engineering structure, there is more or less difference between the measurement values, e.g. displacements, of real structure and the calculation results based on the initial finite element model taken in design. It is very essential to reduce the errors of the initial finite element model for the evaluation of structural safety. The objective of this paper is to propose a new statistical model updating method to modify the finite element model using static measurement data with non Gaussian errors. In this paper, a static model updating equation with random modification factors is established. The modification factor of each element is defined as a variation of the element stiffness, and is expressed as a random multivariate power series expansion. A high order perturbation method is presented to solve the random model updating equation after it is condensed. Then the coefficients of random multivariate power series expansion can be determined, which yields the statistical characteristics of element modification factor. In the numerical examples, the direct Monte Carlo simulation method is taken to verify the accuracy of the proposed method. Numerical results show that this proposed statistical model updating method can ensure the accuracy and efficiency of calculation.

• 出版日期2016
• 单位武汉理工大学