DOI: 10.5176/2301-394X_ACE18.141
Authors: Fu-Shang Wei and Clifford Anderson
Abstract:
Complex structural dynamic models require an analysis that includes frequency responses in the design considerations. In the structural dynamic model improvement analyses, the theoretical mass and stiffness matrices obtained from finite element analysis are often modified based on an incomplete set of modal test data. This paper uses the generalized inverse technique to investigate structural dynamic model improvement and expands the theoretical bases used to derive the governing equations. Two different approaches are used to find the desired matrices. The first is to minimize the orthogonality equation to satisfy both the dynamic equation and the symmetry matrix relationship. The second is to minimize the dynamic equation to satisfy both the orthogonality constraint and symmetry matrix relationship. A simple numerical example is presented to demonstrate the analysis. Three different numbers of modes are used to simulate the incomplete modal test data. Four different design configurations of analytical model diagonal elements have been varied from +3% to +10% of expected values to simulate the real system design uncertainty between the true and analytical models. Results indicate that the percentage root mean square changes for the improved model as compared to the true model are relatively small even with up to +10% errors in the design. In addition, increasing the number of modes used in the analysis, improved model performance as compared to the true value. Further research is strongly recommended.
Keywords: Structural dynamic model modification; Generalized inverse method; Incomplete model anlysis; Structural System Identification.
Updating...