Proper Projection Vectors for the Expansion of Measured Experimental Modal Data

Abstract

Expansion of experimental measured mode shapes using the system equivalent reduction and expansion process (SEREP), or any expansion process that uses mode shapes for that matter, can be hampered due to lack of a well-correlated finite-element model. The lack of adequate correlation between the finite-element model and the experimental data has been blamed for the distortion of the expanded mode shapes. However, the root cause of the expansion has never been clearly identified but rather just stated as unsuccessful by several researchers. Recent work using the variability improvement of key inaccurate node groups (VIKING) in order to remove the variance and errors on measured data has revealed some of the root causes for the success and lack of success with the expansion process. In this paper, the basic reduction and expansion process is reviewed along with some basic system modeling and structural dynamic modification concepts to clearly identify the root cause for the success or lack of success in the expansion process. An example is presented to show the proper manner to select modes for the expansion process to circumvent the distortion of the expanded measured mode shapes.