Structural System Identification by Sequential Quadratic Programming

• Main Authors: Wu, Guo Wei, 伍國維
• Other Authors: Yeong-Kang Chang
• Format: Others
• Language: zh-TW
• Published: 1997
• Online Access: http://ndltd.ncl.edu.tw/handle/34201510423579724266

碩士 === 淡江大學 === 機械工程學系 === 85 === The purpose of this study is to propose a general method to correlate thefinite element analysis data and modal test data systematically. Mathematically the structural system identification problem is identical to the optimum design problem in which the difference between analysis data and test data are used as objective function. When all constraints are satisfied and the objective function becomes minimum, we can obtain the new finite element analysis model which is similar to the test model. The Sequential Quadratic Programming（SQP）is adapted to solve this problem. The difference of analysis/test data in natural frequency, mode shape and design parameter will be considered in objective function. The natural frequency, mode shape, design parameter and structural mass also will be controlled in constraints to limit their allowances. With the helpof SQP and the improve move limit, the design analyst can obtain the more accurate finite element model easily and the saving of computer time is significant. Sensitivity is adopted to establish reanalysis modal and determine search direction. The better search direction avail to solve the optimum problem. A few numerical examples will be solved to demonstrate the capability of the above method.