This study evaluates the influence of different (audio-)visual parameters in the perceived noise annoyance reported in VR experiments. For this purpose, an open-source application developed in Unity was employed to simulate 16 different VR scenarios based on real-life locations. These scenarios were characterized by different binary visual aspects (e.g. rural vs. urban, sunny vs. cloudy, or artificial vs. natural). In each scene, the same binaural aircraft flyover recording was employed to focus on the effect of the different environmental conditions. However, the background noise differed per soundscape, providing different signal-to-noise ratio (SNR) values. The influence of the aircraft visibility (not rendered in some cloudy scenarios) was also evaluated. Preliminary results show that, in general, cloudy, rural, and natural environments were perceived as more annoying. Moreover, a significant and strong correlation was observed between the annoyance ratings and the SNR, showing that background noise can partly mask the presence of aircraft.

With the continuous growth of the aviation sector, concerns regarding the effects of aircraft noise on the health and well-being of communities living in the vicinity of airports have been increasing. Aircraft noise annoyance is inherently subjective and its accurate prediction and quantification represent challenging tasks. There is a lack of consensus in the scientific community regarding which metrics are the best predictors for this type of annoyance. Additionally, many of the metrics employed in the field of (psycho)acoustics are typically computationally expensive. This study aims at developing a methodology which leverages machine learning techniques for instant predictions of various sound metrics and for annoyance rating predictions from input aircraft flyover recordings. The two-step framework involves a Convolutional Neural Network (CNN) for the former, followed by artificial intelligence (AI) models which use the CNN predictions as input for the latter, such as the Support Vector Machine and Random Forest. A listening experiment was conducted in order to gather labeled annoyance data from 60 aircraft flyover recordings of both landings and take-offs and a correlation analysis was subsequently made considering a large pool of sound metrics. The results show that, in general, metrics derived from Psychoacoustic Annoyance models, especially those of Zwicker and Di et al., present better performance as predictors compared to conventional metrics and most Sound Quality Metrics taken individually. Moreover, the AI framework achieves very promising results for both the annoyance ratings and metric predictions (overall, mean absolute errors of the annoyance ratings of approximately 0.4 and below, and R2 values above 0.85), highlighting the potential for bypassing the typically long overhead associated with computing SQMs (which involve expensive and complex algorithms). The drawn conclusions are dependent on the limited amount of data that was available. Hence, the results require further validation using more recordings of various types of aircraft.

Turbofan engines are still one of the loudest noise sources of modern commercial aircraft. Noise radiation from the engine is highly directional and it is, therefore, important to obtain this directivity pattern from microphone-array data from static engine noise tests. Recently, Sijtsma has shown that the deconvolution method CLEAN-SC is able to break down engine noise sources and that it is capable of determining the directivity of these sources in the far field using measurements of a DGEN380 static engine test. This study further investigates this capability of CLEAN-SC by performing acoustic experiments under controlled conditions in the anechoic chamber at the faculty of applied sciences at Delft University of Technology. The experiments consist of a variety of tests with 2 different speakers placed inside an aluminum cylindrical pipe simulating an engine. In these experiments, the location and directivity of the individual noise sources can be measured separately, so that the CLEAN-SC results can be compared to the ground-truth reference. Moreover, the results are also compared with other acoustic imaging methods, such as DAMAS and conventional beamforming.

Predicting aircraft Take-Off Weight (TOW) has been a long-sought task by aviation stakeholders, especially for operational and regulatory bodies involved in flight planning. Unfortunately, TOW being a sensitive parameter to operational trends and cost indices, aircraft operators tend to keep it confidential. In recent years, Machine Learning (ML) algorithms have achieved increased prediction accuracy and capabilities in the field, provided the availability of TOW data. This paper studies the implementation of gradient boosting algorithms as well as Random Forests to better understand which algorithm is best-suited for aircraft TOW prediction (prior to take-off) solely based on Flight PLan (FPL) and Terminal Aerodrome Forecast (TAF) parameters. The study focused on flights at Amsterdam Airport Schiphol (AMS) for training the algorithms, using an 80-20% train-test split. Between Gradient Boosting Decision Trees (GBDTs), LightGBM, XGBoost, and Random Forests, GBDTs achieved the smallest Mean Absolute Percentage Error (MAPE) with 1.71 and 2.17% on the training and testing datasets, respectively. The most influencing feature proved to be the requested cruise speed, followed by great circle distance between airports, and aircraft type. The model was validated on Paris - Charles de Gaulle Airport (CDG) and Brussels South Charleroi Airport (CRL), proving its independence from airport type. However, the distribution of flights in the training dataset, especially that of aircraft and airline types, proved to be an influencing factor for the model's applicability to other airports. Future work includes expanding the training dataset to all flights in the European network, and introducing trajectory-based features such as aircraft speed intent. With a larger training dataset, neural network algorithms could also be explored. Finally, regarding the improvement of trajectory predictions, it was found that better accuracy of TOW predictions does not suffice and that other operational parameters' effect should be investigated, especially speed profiles.

Aeroacoustic testing in wind tunnels is crucial for understanding and mitigating the noise generation mechanisms in several devices while maintaining satisfactory aerodynamic performance in the conceptual design stage. However, current measurements in closed-section wind tunnels face challenges in terms of installation, due to the effect of the boundary layer of the wind tunnel walls, and accuracy. To address these issues, the proposed methodology integrates advanced signal processing techniques and cost-effective and limited alterations in a closed-section wind tunnel. Different configurations, such as a perforated panel, a perforated panel with melamine foam rings, and the addition of melamine foam panels behind the array and inside the wind tunnel combined with the use of a microphone array consisting of 88 microphones, recessed behind an acoustically transparent stainless steel mesh, has led to significant improvements in signal-to-noise ratio and measurement accuracy compared to the baseline aeroacoustic testing. In general, this setup enables the identification of noise sources with a signal-to-noise ratio of at least -10 dB. Additionally, the utilisation of advanced beamforming techniques (CLEAN-SC and DAMAS) in post-processing yields clearer outcomes. Finally, the effectiveness of the set-up was evaluated, resulting in an approximate 15 dB improvement in peak prominence of the flow-induced noise source due to the higher number of microphones and beamforming.

Wind turbine noise is one of the grand challenges in the public acceptance of onshore wind farm projects. The field of psychoacoustics identifies the auralisation of wind turbine noise as a link between technical design and annoyance estimation. There is currently limited work on the auralisation of wind turbine noise, and none targets an application in psychoacoustic research.

This work investigates the auralisation of the aeroacoustics output of DTU's HAWC2 for use in annoyance estimation. A Gaussian beam tracing approach propagates the frequency domain output to observer locations. The resulting spectrograms are converted into sound signals by applying random phase and the inverse short-time Fourier transform. This work includes a binaural rendering module to enable future VR applications. The methodology's implementation results in the Wind Turbine Auralisation tool, WinTAur.

The noise signal output of WinTAur is validated using the HAWC2 model of a stall-controlled NTK 500/41 wind turbine and corresponding acoustic field measurements. Psychoacoustic sound quality metrics show significant differences between the auralised and measured noise. In the overall psychoacoustic annoyance metric, these differences mainly depend on the observer's position around the turbine. All metrics show this directionality dependence, while the loudness, sharpness and tonality metrics also indicate a dependence on wind speed. Differences in fluctuation strength show a minor dependence on the simulation case but are difficult to relate to a specific simulation parameter.
Spectral analysis of the simulation output samples reflects the limitations of HAWC2, demonstrating that it is the primary source of discrepancy. The analysis especially highlights the inaccurate prediction of the directionality and stall noise of the HAWC2 code. The choice of ground type is another probable source of discrepancy, as it does not accurately represent the measurement setup.
A subjective listening experiment demonstrates the significance of these discrepancies in human perception with generally high difference ratings between the simulated and recorded noise. The results illustrate a dependence on wind speed and the position around the turbine. These dependencies match well with the findings from the numerical validation.\\

Future work should focus on a sensitivity analysis of WinTAur since the case-independent parameters may be additional sources of discrepancy. Another recommendation is to investigate the unveiled errors in the underlying methodology. Lastly, better propagation modelling concerning the wind turbine wake and turbulence should be part of future wind turbine noise modelling.

Overall, using modelled wind turbine noise for the auralisation in psychoacoustic research has shown promising results. Validation with sound quality metrics provides good insights into the discrepancies found in subjective listening experiments. Eliminating the existing discrepancies through modelling improvements will allow this work to be applied in a fully modelled approach to estimate wind turbine noise annoyance.