The standard method for computing noise contours around civil airports is SAE-AIR-1845 (FAA Integrated Noise Model, European ECAC-CEAC Doc. 29). It is subject to the inaccuracies implicit in the model as well as those caused by erroneous or imprecise input data. Regarding the latter, the existing errors and/or uncertainties, may be amplified in the output results, to a greater or lower extent, in some cases offering unreliable predictions.
In order to study this phenomenon, the Institute of Sound and Vibration Research (ISVR – Southampton) carried out a theoretical sensitivity analysis based on the segmentation technique, regarding the input parameters of the SAE-AIR-1845, and obtaining the input variables of the model, the variation of which implied greater changes in the output variables. The results were validated by using the FAA Integrated Noise Model Version 6.0 software, based on the aforementioned document. It has been revealed that the model has a greater sensitivity to factors that modify the flight path, and a lower sensitivity to the other parameters. Thus, an error greater than 10 % in the variable “gross weight” offers an additional error of between 3 and 7 dB. However, parameters such as the ID of the flaps hardly modify the results obtained for the least favourable case by 1 dB.
As a result of this research, the sensitivity of the model was quantified for each of the input parameters (taken alone and also in interaction with other parameters), and criteria for the minimisation of global error, resulting from uncertainties in the input parameters, were stated.
