This research project deals with the development and refinement of analysis techniques necessary to extract valid Statistical Energy Analysis (SEA) parameters from measured system data. SEA is a methodology that permits prediction of system structural dynamic and structural acoustic response in frequency regimes where the finite element method is not feasible. The use of SEA in engineering design could be significantly increased if the necessary SEA parameters can be obtained efficiently and accurately. The goal of this research is to develop system identification techniques and data acquisition methods that will allow for the determination of these required parameters, such as coupling loss factors. This goal will be accomplished through extensive testing using realistic engineering systems. Eigensystem identification methods, applied both in the time domain and in the frequency domain, will be used to estimate SEA parameters as a function of spatial averaging and spatial extent. Both direct inversion of the steady-state energy coupling matrix and a state space realization of the transient response energy will be examined. The results of this research will define the data acquisition and analysis techniques required for obtaining necessary SEA parameters. These results will permit the technique to be implemented more easily and with more confidence by designers and analysts, allowing for a significant expansion of the method in engineering design practice.
