Uncertainty Quantification of Numerical Transmission Loss Calculations of an Aircraft Fuselage Section
* Presenting author
The prediction of aircraft cabin noise is important in preliminary design stages as potential acoustic problems can be detected in these stages to consider damping measures as early as possible. In these early design stages a final design is not yet given and design parameters are uncertain. Especially for the cabin noise assessment concerning new technologies, the predicted behavior is crucially sensitive to the chosen deterministic design parameters. Hence, the combination of an uncertainty quantification with a deterministic acoustic model is a promising approach. In this contribution, a finite element model of an aircraft fuselage section is used to predict the transmission loss. The finite element model is treated as black box with certain input parameters on the basis of which a statistical analysis and a global sensitivity analysis is conducted. A non-intrusive surrogate modeling technique is applied in frequency domain to reduce the computational costs significantly. The results show a reliable fit of the transmission loss curve by the polynomial surrogate model. However, a limit of the polynomial surrogate model is indicated by a convergence study (polynomial order and numbers of samples) due to the occurrence of problems in the strongly non-monotonic resonances.