Uncertainty Quantification for Aircraft Noise Emission Simulation: Methods and Limitations

Abstract

Implementing measures for the reduction of aircraft noise impact, such as optimization of flight paths and aircraft, requires sophisticated simulation capabilities. These tools have to incorporate simulation of aircraft noise generation at the source (i.e., emission) and account for prevailing sound propagation effects to ultimately predict noise levels as received on the ground. Obviously, understanding the associated uncertainties is crucial when aiming at a reliable and meaningful assessment. It also becomes essential when comparing different technologies, mixing experimental and numerical data, or using simulation methods of different fidelity (e.g., semi-empirical and first order methods). This research focuses on quantifying uncertainties in the first step in aircraft noise simulation (i.e., the prediction of the emission situation). The quantified uncertainties reflect imperfections of models for different aircraft noise sources and variability of model input parameters at different operating conditions. A general framework is presented, which also accounts for limited knowledge of the underlying data distributions, and quantitative comparisons of different uncertainty methods are provided. In particular, the first order second-moment analysis is compared to higher-order polynomial chaos methods, and the advantages and disadvantages of the different methods are discussed.