Structural Model Tuning via Vector Optimization.

This report examines tuning a finite element model using vector optimization techniques. Structural models using finite element theory often need to be adjusted so they can accurately simulate the real structure. The goal is to tune the model such that it will reproduce data derived from the structure. The tuning procedure is complicated by many factors. First of all, the model may be required to reproduce several sets of different performance data which may have conflicting effects on the model. The number and kinds of parameters to modify may be large or uncertain. Also, there may be many constraints on the model. With all these considerations and several performances indices to extremize, there may be many compromise solutions to examine in order to select the best one. Therefore, the problem has been broken into two parts. First, the performance indices are extremized using multiple objective optimization theory, producing a set of possible solutions. Next, the solutions are ranked ordered according to a decision makers preferences. The solution ranked number one is, then, the best answer. The tuning process was applied to a T-38 horizontal stabilator. Four static deformation and two natural frequency experimentally determined data sets were used as the objective functions for the three parameter model.