A Frequency Domain Approach to Pretest Analysis and Model Updating in the Presence of High Modal Density
Future high-performance space vehicles pose new challenges to the model validation process that will require advances in structural modeling and analysis, subsystem testing, and model calibrating. Currently there exist vibration analysis techniques that are well suited for the low and high-frequency ranges; however these techniques break down in the mid-frequency range where there can be large variation in the modal density. This research proposes to extend model calibration beyond the low-frequency range by making direct use of the frequency response functions, as opposed to modal analysis. Modal based techniques breakdown outside the low-frequency range where errors are introduced from model uncertainty, test noise, mode coupling, and mode extraction. In lieu of mode shapes, a principal component decomposition of the analytical frequency response is used to form an optimal set of energy based modes that describe the system response. These energy based modes are used as a frequency domain analog to the normal modes, but with the advantage being that they automatically account for input location, damping, and out-of-band modes. New pretest analysis techniques for sensor placement and analytical model reduction are introduced, as well as new methods for test-analysis correlation and model updating. The following research results in a set of tools that are suitable for use in the FEM validation process in and beyond the low-frequency range.
- Pub Date:
- Engineering, Aerospace