Evaluation of Loss Factor Estimation Techniques for Free Hanging Flat Panels Excited Mechanically

Abstract

To establish the "best" technique to estimate a damping loss factor for mechanically-excited panels, three loss factor estimation techniques--PIM, IRDM, and RDT--are compared. In experimental and computational analyses, panels with two damping levels and three sizes were tested. The loss factor estimates from each of the three techniques are then evaluated in four distinct frequency bands centered at one-third octave frequencies of 500 Hz, 1000 Hz, 2000 Hz and 4000 Hz (for computational analysis only). Unlike IRDM and RDT, the quality of PIM-based loss factor estimates have presented a strong correlation between the region of response measurement and it is distance from the excitation location. PIM-based loss factors were significantly underestimated when responses are measured inside the direct field. PIM-based loss factors are relatively accurate only if the measurements are made from wide-spread response locations. For a lightly damped panel, loss factor estimates using PIM, IRDM and RDT with direct averaging agree within reasonable accuracy. For intermediately to highly damped panels, IRDM and RDT with direct averaging under-predicted the loss factor; RDT with an autocorrelation function averaging approach slightly over-predicted the loss factor. Both RDT approaches might be used to set a bound on panel loss factor. Even when significantly fewer response locations are considered, it is evident that loss factor estimates from RDT are as reliable as IRDM and more reliable than PIM especially for highly damped panels. For the analysis of freely hanging plates, excitation "close to an edge", especially for PIM, is not recommended. When analyzing the panel loss factor, arbitrary or central excitation is acceptable.