Doctor of Philosophy / Department of Mechanical and Nuclear Engineering / Zhongquan Zheng / Windscreens are widely used in outdoor microphone measurement for acoustic sensing systems. In many cases of outdoor microphone applications, wind noise interferes with the signals. The performance of measurement microphones thus heavily depends on correct designs of windscreens that are used to maximize the signal to noise ratio of the sensing system. The purpose of the study is to investigate the wind noise reduction between the unscreened microphone and the screened microphone under different frequencies of incoming wind turbulence. In this study, a modified immersed boundary method using the distributed forcing term has been applied to simulate the flow/acoustic interaction between air and the porous medium. Because of the high accuracy requirement in the vicinity of the interface between air and the porous medium, spatial derivatives of flux need to be discretized using high order schemes. In this study, several different schemes have been tested in the vicinity of the interface including a second-order upwind scheme, a third-order upwind scheme, and a fifth-order Weighted Essentially Non-Oscillatory (WENO) scheme. Based on the test results, the fifth-order WENO scheme is selected for most of the simulation cases. The model equations for flow outside the windscreen are the Navier-Stokes equations; flow inside the windscreen (porous medium) uses the modified Zwikker-Kosten equation. The wind turbulence in this study is generated by two different ways. The first is to place different sizes of solid cylinders and spheres in the upstream of the microphone under two-dimensional and three-dimensional conditions. The second is to use a Quasi-Wavelet method to generate the background atmospheric turbulence to simulate the real physical phenomena. Both two-dimensional and three-dimensional simulations for the flow over the unscreened and the screened microphone are presented and discussed under both low Reynolds number and high Reynolds number flow conditions. The results show that the windscreen effect is significant and the wind noise reduction level between the unscreened and the screened microphone can reach around 20dB either for low Reynolds number cases or for high Reynolds number cases. For low Reynolds number cases, Low flow resistivity windscreens are more effective for low frequency turbulence; high flow resistivity windscreens are more effective for high frequency turbulence. For high Reynolds number cases, the medium flow resistivity windscreens perform better compared to low flow resistivity windscreens and high flow resistivity windscreens.

Predictors of people’s responses to noise inside cars are used by car companies to identify and address potential noise problems from tests. Because significant advances have been made in the reduction of engine, powertrain, and tire/road noise, it is now important to pursue reductions in wind or aerodynamic noise. While models of loudness are commonly used to predict people’s responses to stationary wind noise, some wind noises are less acceptable than is predicted by the loudness metric. Additional sound characteristics may account for this. The research described in this dissertation was conducted in two main stages. The focus of the first stage was on improving acceptability predictions for stationary noise, by using additional sound quality metrics along with predictions of loudness. Three listening studies were designed and conducted, including one study with aspiration noise. Test sounds were a combination of recordings made on cars in a wind tunnel and modified recordings. Methods to modify individual sound characteristics were developed to de-correlate metrics across the set of test sounds, and to examine trends of acceptability with particular sound characteristics. Models of acceptability for stationary wind noise are significantly improved when a metric that predicts the sharpness of a sound is included in the model with the loudness metric. The focus of the second stage of the research was on improving acceptability predictions for non-stationary noise, particularly noise with the kind of variations that are expected from wind gusts. A simulation method was developed to generate sounds with controlled gusting features by modifying stationary noise recordings. Two listening studies were conducted containing simulated gusting sounds, and a gusting unacceptability metric was designed to predict subjects’ responses based on the strength, modulation rate, and duration of the gusts. The inclusion of this gusting metric significantly improved the goodness of fit of linear and logistic models of non-stationary noise acceptability containing Loudness and Sharpness.

No description available.

When creating data-based models it is important to include the underlying physical characteristics and constraints of the data. If physical characteristics are not properly included in the model, results may be infeasible or physically impossible. Acoustic environments are better characterized by ensuring that models include the fundamental spatial, spectral, and temporal characteristics of noise sources, or how they change based on location, frequency, and time. When model data are limited, in availability or in reliability, additional care must be taken to ensure models predict feasible results. This dissertation focuses on physics-guided modeling of acoustic environments using limited data, taking into consideration spatial, spectral, and temporal characteristics of noise sources, specifically focused on wind noise and traffic noise. Wind noise contamination in spectral data can be significant, even when using a windscreen. By modeling spectral characteristics of temporally varying wind noise contamination, a method for automatically detecting and reducing wind noise was developed. Reducing non-acoustic wind noise contamination allows for better characterization of outdoor acoustic environments and is useful for accurately measuring other noise sources. Traffic noise varies spatially, spectrally, and temporally, and depends on traffic volume (the number of vehicles per unit time) and traffic class mix (e.g., the relative number of small vehicles compared to large trucks). Using the temporal variation found in reported traffic volume at thousands of locations, a model was developed to represent and predict the spatio-temporal variability of traffic volume nationwide. Further models were created to include dynamic changes in traffic class mix and to predict spectral source traffic noise. The resulting model for predicting source traffic noise is known as VROOM, the Vehicular Reduced-Order Observation-based Model. The physics-guided modeling techniques presented in this dissertation are intended for characterizing acoustic environments, which has applications for such diverse areas as human health and wellness, bioacoustics, wildlife conservation, urban and roadway planning, land development and conservation, noise ordinance legislation, homebuying, and more.

Over the last years, due to the development of quieter engines and drivetrains, the importance of addressing the vehicle wind noise problem has significantly increased.In this thesis work, several existing Transfer Path Analysis methods have been applied on an experimental database acquired during a wind tunnel test on a passenger car with the objective of analyzing the distribution of the wind noise sources and their contribution to the target microphones located inside the vehicle. A major challenge for the Transfer Path Analysis (TPA) consists of the high complexity of the aerodynamic sources exciting the structure. Moreover, the existence of multiple incoherent source phenomena, and the presence of distributed coherent source regions of different correlation scales make the analysis very complex.The thesis work provides a solid and comprehensive analysis of the results obtained by different methods. The outcomes can be potentially useful for optimizing the vehicle NVH performance in future practical cases. / Under de senaste åren har vikten av att arbeta med vägfordons problem med aerodynamisk ljudgenerering ökat avsevärt på grund av utvecklingen av tystare motorer och drivlinor. I det här projektet har flera existerande metoder för Transfer Path Analysis (TPA) tillämpats på en databas med experimentella data som samlats in vid vindtunneltest på en personbil, med målet att analysera fördelningen av källorna orsakade av vindbruset och deras påverkan på ljudnivån vid de uppsatta målmikrofonerna inuti fordonet. En stor utmaning för TPA är den höga komplexiteten hos de aerodynamiska källorna som exciterar strukturen. Vidare gör förekomsten av flera okorrelerade källor, och närvaron av distribuerade koherenta källregioner med olika korrelationsskalor, analysen mycket komplex. Arbetet presenterar en solid och omfattande analys av resultat som erhållits med olika metoder. Resultaten är potentiellt användbara för att optimera fordonets NVH-prestanda i praktiktiken i framtiden.

<p dir="ltr">As speeds and directions of the vehicle and wind change, the unsteady flow creates variations in wind noise, which can be referred to as non-stationary wind noise. To investigate people's perceptions of non-stationary wind noise, a method to simulate the non-stationary wind noise is needed. Previously, a method was developed that used stationary recordings taken at several wind speeds and directions to form functions that relate the 1/3 octave sound pressure level with wind speed and direction. These functions are used to create time-varying filters based on provided time histories of wind speed and direction. A reference wind noise measurement is then filtered to produce the sounds. To reduce the time cost of taking many stationary measurements, an improved method was investigated. At each yaw angle, one speed sweep wind tunnel measurement was used to estimate the relationship between sound pressure level and wind speed. Two partially correlated white noise signals were then filtered to simulate binaural sounds that had a similar coherence structure between the left and right ear sounds to that observed in binaural measurements in the vehicle. The accuracy of the simulations was validated by comparing wind noise simulations with wind tunnel and on-road vehicle interior noise measurements. For the on-road measurements, a noise decomposition method based on noise source measurements was used to estimate the road/tire noise, which was then added to the simulated wind noise to make it comparable with the measured on-road noise. The time-varying loudness, and power spectral densities of the simulated and measured sounds were found to be well consistent. Besides, a method to simulate the turbulent wind speed time histories, which can be used as inputs to the wind noise simulation method, was developed. The von Karman model predicts the spectra of wind turbulence by assuming it to be a stationary random process. White noise signals can then be filtered to simulate the stable variations of wind speeds. The discrete gusts, which are the transient events in wind speed time histories, were also simulated by using an 8-parameter piecewise function. Eventually, the non-stationary wind noise and the turbulent wind speed simulation method can be a powerful tool when investigating sound perceptions of vehicle interior wind noise.</p>