Evaluating the Stiffness of Conic Interfacing Parts : a Practical Method for Finite Element Model Updating Based on Experimental Modal Testing

摘要

The current work describes an iterative process to optimize the stiffness properties in a finite element model with respect to the actual stiffness of a complex joint. When solving a finite element model in structural dynamics, simplifications come from defining certain boundary conditions, material properties or model geometry as known parameters. These parameters must be properly evaluated in order to obtain a realistic representation of the structure's reaction to dynamic loading. Vibro-acoustic analysis of a solar panel substrate contains frontiers which need to be described in finite element modelling, to make sure that the experimental responses of the panel can be properly correlated with its theoretical representation. An iterative procedure is performed to reduce stiffness errors based on the orthogonality of the finite element modes with the actual physical modes. Using Siemens NX model updating module, a genetic algorithm is used to optimize the correlation. The practical method described here was found to be effective to reduce stiffness based correlation errors. The presented practice made possible to increase mode shape correlation by 11% of the modal assurance criterion (MAC) and decrease the average frequency error by 43%.