ABSTRACT Statistical Energy Analysis (SEA) has been used widely by industry and academia for more than 20 years to predict the mid-to-high frequency range behavior of complex acoustic systems. At Gulfstream Aerospace Corporation (GAC), SEA models have been developed to predict the interior cabin noise levels of completed Gulfstream aircraft. These models are also used for acoustic evaluations of design changes prior to implementation as well as a diagnostic tool for investigating noise and vibration issues. Throughout the development of the SEA models, extensive experimental testing in GAC's Acoustic Test Facility (ATF) was conducted on numerous aircraft components represented in the models. This paper demonstrates the importance of using experimental data to improve the accuracy of the SEA predictions by accurately adjusting the material properties and acoustic parameters of the SEA model to better match the ATF experimental data. This is particularly important for complicated SEA models with thousands of subsystems and junctions. INTRODUCTION SEA has been used widely in the industry and academic field for more than 20 years to predict the mid-to-high frequency behavior of acoustic components. By combining statistical modeling and energy approaches to quantify the subsystem response, SEA is used to study the vibroacoustic response of large complex systems very efficiently. At GAC, SEA models have been developed to model the entire completed aircraft [ 1, 2] with the SEAM3D software package [ 3]. The largest model developed so far in GAC consists of 2450 subsystems and 8578 junctions. Figure 1 shows a graphical representation of a Gulfstream aircraft SEA model.SEA is a powerful tool, but the user needs to have accurate parameters to input into the model in order to get reasonable results out of the model. Typical parameters include the following: 1. Geometric dimensions 2. Subsystem material properties 3. Junction coupling factors 4. Input loading The general modeling steps covering these input parameters are discussed in Refs [ 1, 2]. This paper specifically targets the subsystem and junction property characterizations. Through the development process of the GAC Aircraft SEA models, hundreds of parametric studies have been conducted to help define the material and coupling inputs to correlate the SEA predictions with the test data, measured both on the aircraft and in GAC's ATF. The experimental data was found to have an important role in improving the accuracy of SEA predictions, not only in terms of absolute subsystem response, but more importantly, for improving the transfer path analysis (TPA). For SEA models with thousands of subsystems and junctions and numerous transfer paths, an accurate TPA is vitally important for making recommendation on weight- effective design improvements [ 4]. This paper illustrates the importance of test data by first introducing the acoustic testing conducted in the ATF, and then showing the necessities of using experimental data to improve the SEA predictions. Discussions will also be made regarding the TPA portion of the SEA predictions to illustrate the importance of modeling each path as accurately as possible. Finally, the predictions of the aircraft SEA model with correlated material and junction properties are compared with the test data. Improving SEA Predictions with Experimental Data2011-01-1701 Published 05/17/2011 Tongan Wang and John Maxon Gulfstream Aerospace Corporation Copyright © 2011 SAE International doi:10.4271/2011-01-1701Downloaded from SAE International by University of Auckland, Saturday, August 04, 2018Figure 1. Graphical representation of a Gulfstream aircraft SEA Model GAC ACOUSTIC TEST FACILITY The GAC ATF features a 252 m3 reverberation chamber, and a 215 m3 hemi-anechoic chamber coupled by a transmission loss opening of up to 2.4m × 2.4m between the two chambers, as shown in Figure 2. A separate control room is used to collect and analyze the data acqu