Developing a data based method to quantify the effects of flight track, aircraft weight and engine setting on the received aircraft noise levels

Airports in The Netherlands are subjected to tangent environmental laws to restrain pollution and noise nuisance. Amsterdam Airport Schiphol (AAS) is one airport dealing with this regulatory framework but nevertheless they are resolute to continue growth with respect to aircraft movements. To cope with the law related to aircraft noise, the department Stakeholder Strategy and Development (SSD) of AAS is responsible for the implementation of Noise Abatement Measures (NAMs). NAMs are used to minimize aircraft noise as to be able to maximize the number of aircraft movements within the environmental law as set by the Dutch government. SSD demands to be able to visualize the effect of a NAM by measuring aircraft noise with its Noise Monitoring System (NOMOS). However, in practice it appears that the effect of a NAM to the exposed noise level cannot easily be determined since the total set of measurements show a high degree of scattering. This is caused by the fact that many other parameters are contributing to the exposed noise level as, for example, engine setting and aircraft configuration. Therefore, AAS encounters difficulties evaluating the effectiveness of implemented noise reducing measures using the noise levels as measured by NOMOS. Hence, the research question becomes: How can the distinctiveness between noise measurements effectively be improved as to evaluate the direct effect of a Noise Abatement Measure to the measured noise level? As a first step towards answering this question, aircraft mass m and aircraft engine setting N1 were identified which were expected to mask the effect of a NAM to the measured noise level. Then, the Peak Find Method (PFM) is developed to determine N1 from the associated acoustic time series as retrieved from NOMOS. Thirdly, aircraft mass m was found to be very difficult to determine from aircraft performance theories. Therefore, the lift-off speed at take-off Vlof2 is taken as an aircraft mass representative. With the two predictors N1 and Vlof2 available and the measured maximum loudness levels Lmax retrieved from NOMOS, a Multivariate Linear Regression Analysis (MLRA) is carried out to assess the effect of the two predictors to variations in Lmax. Last, the identified MLRA model is used to subtract the contribution of N1 and Vlof2 from the received noise levels, hence leaving the direct effect of a NAM to the measured noise level. Initial correlation analysis showed no correlation between N1 and Lmax and neither between Vlof2 and Lmax. While the MLRA model is based upon the identified values of the predictors, it was therefore not expected that the high variations in Lmax would decrease when using these predictors, bearing in mind the results of both correlation analysis. And ultimately, by using the MLRA model only 7% of the total variation in Lmax could be explained, which turned out to be too less to evaluate the direct effect of a NAM to the measured noise level.Air Traffic Performance and the EnvironmentControl & OperationsAerospace Engineerin