Compressive Sampling as an Enabling Solution for Energy-Efficient and RapidWideband RF Spectrum Sensing in Emerging Cognitive Radio Systemsкод для вставкиСкачать
Design and Characterization of the UTIAS Anechoic Wind Tunnel by Derrick H.F. Chow A thesis submitted in conformity with the requirements for the degree of Master of Applied Science Graduate Department of Institute for Aerospace Studies University of Toronto © Copyright 2016 by Derrick H.F. Chow ProQuest Number: 10043015 All rights reserved INFORMATION TO ALL USERS The quality of this reproduction is dependent upon the quality of the copy submitted. In the unlikely event that the author did not send a complete manuscript and there are missing pages, these will be noted. Also, if material had to be removed, a note will indicate the deletion. ProQuest 10043015 Published by ProQuest LLC (2016). Copyright of the Dissertation is held by the Author. All rights reserved. This work is protected against unauthorized copying under Title 17, United States Code Microform Edition © ProQuest LLC. ProQuest LLC. 789 East Eisenhower Parkway P.O. Box 1346 Ann Arbor, MI 48106 - 1346 Abstract Design and Characterization of the UTIAS Anechoic Wind Tunnel Derrick H.F. Chow Master of Applied Science Graduate Department of Institute for Aerospace Studies University of Toronto 2016 The anechoic open-jet wind tunnel facility at the University of Toronto Institute for Aerospace Studies was updated and characterized to meet the needs of current and future aeroacoustic experiments. The wind tunnel inlet was resurfaced and flow-conditioning screens were redesigned to improve the freestream turbulence intensity to below 0.4% in the test section. The circular nozzle was replaced with a square secondary contraction that increased the maximum test section velocity to 75 m/s and improved flow uniformity to over 99% across a usable cross-sectional area of 500 mm × 500 mm. Acoustic baffles were installed in front of the wind tunnel inlet and foam wedges were installed in the anechoic chamber. The overall background sound pressure levels in the chamber were improved by 8-18 db over the range of operational freestream velocities. The anechoic chamber cut-off frequency is 170 Hz and the reverberation time for a 60 dB sound power decay is 0.032 s. ii Acknowledgements I would like to thank Professor Philippe Lavoie for all the knowledge and wisdom he imparted on me throughout my time in his lab. His patience and support were key to completing the work in this thesis. I also appreciate all the professors on my research assessment committees for lending their time and expertise to ensure my work met high standards of quality. I am deeply grateful to Philip McCarthy, a.k.a. “British Phil”, for the incredible dedication and effort he displayed in helping me throughout the wind tunnel retrofit process. I am glad that he was there with me during this journey and I have learnt a lot from him. Finally, I’d like to express my sincerest thanks to all my colleagues in the FCET and EFD lab groups, as well as several others in the UTIAS community, who gave me the critical support I needed to complete my work and ensure that I had a good time at UTIAS. iii Contents 1 Introduction 1 1.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.2 Airframe Noise Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 1.3 Outline of Thesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 2 Background 2.1 7 Anechoic Wind Tunnel Design . . . . . . . . . . . . . . . . . . . . . . . . . 7 2.1.1 Aerodynamic Specifications . . . . . . . . . . . . . . . . . . . . . . 8 2.1.2 Acoustic Specifications . . . . . . . . . . . . . . . . . . . . . . . . . 10 2.2 Survey of Existing Facilities . . . . . . . . . . . . . . . . . . . . . . . . . . 13 2.3 Existing UTIAS facility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 3 Facility Updates 3.1 23 Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 3.1.1 Aerodynamic objectives . . . . . . . . . . . . . . . . . . . . . . . . 23 3.1.2 Acoustic objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 3.2 Flow Conditioning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 3.3 Secondary Contraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 3.4 Acoustic Wedges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 3.5 Front-end Baffles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 4 Facility Characterization 40 4.1 Flow Speed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 4.2 Flow Uniformity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 4.3 Freestream Turbulence Intensity . . . . . . . . . . . . . . . . . . . . . . . . 45 4.4 Background Noise Levels . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47 4.5 Anechoic Cut-off Frequency . . . . . . . . . . . . . . . . . . . . . . . . . . 54 4.6 Reverberation Time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 iv 5 Conclusions 62 5.1 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 5.2 Recommendations for Future Work . . . . . . . . . . . . . . . . . . . . . . 63 A Anechoic Foam Material Data Sheet v 65 List of Tables 1.1 1.2 ICAO subsonic aircraft noise reduction standards. . . . . . . . . . . . . . . ACARE and NASA aircraft noise reduction targets. . . . . . . . . . . . . . 2.1 Specifications of select aeroacoustic research facilities around the world. . . 15 3.1 Summary of aerodynamic improvement objectives for the UTIAS AWT facility. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 Summary of acoustic improvement objectives for the UTIAS AWT facility. 24 Comparison of new and old wire mesh screen specifications. . . . . . . . . . 29 3.2 3.3 4.1 4.2 4.3 5.1 2 2 Summary of sound transmission distances for baffle noise attenuation test. 52 A-weighted OASPL white noise attenuation by front-end baffles and wind tunnel components. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53 SPL error bounds for free-field assumption. . . . . . . . . . . . . . . . . . . 55 Summary of tunnel performance objectives and results from the facility characterization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 vi List of Figures 1.1 Major sources of airframe noise. . . . . . . . . . . . . . . . . . . . . . . . . 4 2.1 Examples of anechoic wind tunnel facilities around the world. . . . . . . . 8 2.2 A-weighting SPL curve. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 2.3 Comparison of background OASPL of select anechoic wind tunnels around the world. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 2.4 Views of the existing UTIAS AWT facility. . . . . . . . . . . . . . . . . . . 18 2.5 Freestream turbulence intensity plots for old AWT. . . . . . . . . . . . . . 19 2.6 Test section velocity profiles for old AWT. . . . . . . . . . . . . . . . . . . 20 2.7 Views of the fibreglass wedges in the old anechoic chamber. . . . . . . . . . 21 2.8 Background SPL spectra for old AWT. . . . . . . . . . . . . . . . . . . . . 22 2.9 Background SPL spectra for ISVR anechoic wind tunnel. . . . . . . . . . . 22 3.1 Comparison of predicted and measured SPL spectra for NACA0012 airfoil in ISVR anechoic wind tunnel. . . . . . . . . . . . . . . . . . . . . . . . . . 25 3.2 Schematic plan view of updates to UTIAS AWT facility. . . . . . . . . . . 26 3.3 Photos of the old bell-mouth under construction. . . . . . . . . . . . . . . 27 3.4 Photo of the tunnel front-end with the resurfaced bell-mouth. . . . . . . . 27 3.5 Close-up scale comparison of the old and new honeycomb mesh screens. . . 28 3.6 Comparison of old and new AWT test section inlets with approximate shapes of the open jets. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 3.7 Plot of the new secondary contraction wall profile. . . . . . . . . . . . . . . 31 3.8 Photos of the secondary contraction construction process. . . . . . . . . . . 32 3.9 Views of the new polyurethane wedges in the anechoic chamber. . . . . . . 33 3.10 Acoustic absorptivity curve of fibreglass and open-cell polyurethane foam for wedges. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 3.11 Acoustic absorptivity curve of eggcrate foam sheets. . . . . . . . . . . . . . 34 3.12 Chamber construction photo: Beginning to frame the bare concrete chamber. 35 3.13 Chamber construction photo: Putting up the last studs. vii . . . . . . . . . . 36 3.14 Chamber construction photo: Sheeting the ceiling and walls. . . . . . . . 3.15 Chamber construction photo: Installing wedges and building frame for false door. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.16 Photo from inside completed anechoic chamber. . . . . . . . . . . . . . . 3.17 Schematic plan view of the front-end acoustic baffles. . . . . . . . . . . . 3.18 Photos of the new front-end acoustic baffles. . . . . . . . . . . . . . . . . 4.1 4.2 4.3 4.4 4.5 4.6 4.7 4.8 4.9 4.10 4.11 4.12 4.13 4.14 4.15 4.16 4.17 . 36 . . . . 37 37 38 39 Freestream velocity versus VFD output frequency. . . . . . . . . . . . . . . Variation in static pressure coefficient along the open-jet centreline. . . . . Freestream velocity profiles of new AWT. . . . . . . . . . . . . . . . . . . . Test section flow uniformity versus streamwise location. . . . . . . . . . . . Power spectral density of streamwise velocity fluctuations. . . . . . . . . . Comparison of centreline freestream turbulence intensity for old and new AWT. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Plan view of free-field microphone measurement locations. . . . . . . . . . Background SPL spectra for new anechoic chamber. . . . . . . . . . . . . . Comparison of background OASPL of improved UTIAS AWT with select anechoic wind tunnels around the world. . . . . . . . . . . . . . . . . . . . A-weighted OASPL versus fan motor rotational speed. . . . . . . . . . . . Comparison of A-weighted third-octave SPL spectra before and after baffle attenuation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Schematic of anechoic chamber for free-field characterization. . . . . . . . . Chamber free-field characterization traversing 45◦ relative to the centreline. Chamber free-field characterization traversing 90◦ to the centreline. . . . . Chamber free-field characterization traversing 135◦ to the centreline. . . . . Pressure response signal of impulse noise used in reverberation time test. . Comparison of reverberation times for acoustically-wedged anechoic chamber and bare concrete chamber. . . . . . . . . . . . . . . . . . . . . . . . . 41 42 43 44 46 46 47 49 50 51 53 55 56 57 58 60 61 A.1 Material data sheet for polyurethane foam used for the anechoic wedges in the improved UTIAS AWT. . . . . . . . . . . . . . . . . . . . . . . . . . . 66 viii Chapter 1 Introduction This thesis describes the improvements to the Anechoic Wind Tunnel facility (AWT) at the University of Toronto Institute for Aerospace Studies (UTIAS) for the benefit of future researchers interested in experimentally studying aerodynamic noise phenomena. This first chapter opens with a brief description of the motivation behind this work, then provides some examples of research projects that could be performed in the improved AWT, followed by an overview of the presentation of this thesis. 1.1 Motivation Over the past several decades, environmentally sustainable aviation has become a major topic of discussion among the public, government bodies, the aviation industry, and academia. In particular, regulations and research have focused mainly on three environmental pollutants: carbon dioxide (CO2 ), nitrogen oxides (NOx ), and aviation noise. Of these, aviation noise around airports was the first to be regulated by the International Civil Aviation Organization (ICAO) in 1972 and remains a significant area of consideration by all stakeholders involved in aviation today. Table 1.1 shows a summary of ICAO’s increasingly stringent noise standards over the last five decades. This owes in part to the public’s direct perceptibility of and adverse reaction to aviation noise, as indicated by the historically high proportion of community complaints that are related to airport noise. Indeed, aircraft noise has various behavioural and physiological impacts including sleep and work disturbance, resulting in increased stress, lower productivity, and lower property values in communities surrounding airports [1–3]. With global commercial air travel projected to grow on average by 5% per year over the next 20 years [5, 6], new airports will have to be built and existing ones expanded. As a 1 2 Chapter 1. Introduction Table 1.1: Subsonic jet-powered aircraft noise reduction standards set by ICAO . Year 1972 1977 2001 2014 Version Chapter 2 Chapter 3 Chapter 4 Chapter 14 Standard Reference level 10 dB cumulative below Ch.2 10 dB cumulative below Ch.3 7 dB cumulative below Ch.4 consequence, it is estimated that the population affected by aircraft-related noise around airports will rise from 13.7 million in 2005 to 23.0 million people in 2035 [2,3]. To address this growing problem, the National Aeronautics and Space Administration (NASA) and the Advisory Council for Aerospace Research in Europe (ACARE) have individually set targets for noise reduction for the next two to three generations of aircraft. Table 1.2 presents a summary of these targets, along with converted values for comparison to the ICAO standards. Table 1.2: Noise emission reduction targets set by ACARE and NASA, and their equivalent values relative to ICAO standards . Year Target 2020 ACARE Vision 2020: 50% reduction in perceived noise relative to year-2000 aircraft (revised to FlightPath 2050) 2025 2035 2050 ICAO Equivalent 40 dB cumulative below Ch.4 NASA N+2: 42 dB cumulative decrease relative to ICAO Ch.4 (equivalent) NASA N+3: 71 dB cumulative decrease relative to ICAO Ch.4 (equivalent) ACARE FlightPath 2050: 65% reduction in perceived noise relative to year-2000 aircraft 55 dB cumulative below Ch.4 Historically, the majority of aircraft noise originated from turbojet engine-related sources such as the compressor and exhaust jet, and hence these formed the main focus of aircraft Chapter 1. Introduction 3 noise reduction research. With the introduction of bypass turbofan engines in the 1970s and the subsequent increases in fan bypass ratios and improvements in engine lining and serrated nozzle technology, aircraft manufacturers have largely been able to achieve the increasingly stringent ICAO noise regulations without sacrificing significant performance in other areas. Today, the engine noise levels of an aircraft on approach and landing is comparable to the noise levels generated by the turbulent flow over airframe components . However, the current engine noise reduction techniques are yielding diminishing returns as high bypass ratio engine sizes have reached a ceiling with aerodynamic drag and ground clearance considerations, among others. With the ambitious noise reduction goals set by ACARE and NASA for the next three generations of aircraft, airframe noise sources are now seen as the barrier to further significant improvements in aviation noise, especially from aircraft flying over airport communities on approach for landing. For this reason, an increasing number of people in industry and academia alike have turned their attention to airframe noise reduction research. This thesis seeks to make a contribution to those efforts by providing a high quality aeroacoustics test facility that is unique among Canadian academic institutions. 1.2 Airframe Noise Research The basis for airframe noise research was first formed in the theoretical aeroacoustics work by Lighthill , Curle , and Ffowcs Williams . The first efforts to quantify airframe noise came in the form of full-scale flyover measurements in the early 1970’s [12–16], which were able to isolate the landing gear and high lift devices as the most significant sources of airframe noise for an aircraft on approach to landing (see Fig. 1.1). In the four decades since then, researchers have performed numerous empirical investigations of these airframe components in aeroacoustic wind tunnels to relate the local flow physics to the observed far-field sound characteristics. For landing gears, these research efforts include studies on the broadband noise generated by turbulent boundary layer separation and interaction off major gear components (e.g. main struts, wheels), and the tonal/narrowband noise arising from laminar separation and coherent vortex shedding off smaller gear components (e.g., smaller support struts, hydraulic lines, etc. with Reynolds number below 500,000 based on diameter) [17–20]. For high-lift devices, examples include broadband noise studies on the unsteady flow inside slat coves [21,22] and over slat tracks and flap side-edges . As well, experiments have been conducted to understand the noise generated by the interaction between the landing gear wake and extended flaps . Researchers have also identified many other sources of airframe noise, such as tonal reso- Chapter 1. Introduction 4 Figure 1.1: Some of the major sources of airframe noise, including landing gear, slats, flaps, and wing-tips. Image adapted from . nance from pin-holes and vent cavities , and broadband noise from wing tip vortices and turbulent boundary layers . The majority of these studies were performed in aeroacoustic open-jet wind tunnel facilities, where detailed flow-field characterizations and far-field noise measurements could occur in tandem. Some of the common measurement techniques include Particle Image Velocimetry (PIV) for non-intrusive flow-field quantification, hot-wire anemometry for local velocity spectra, microphone phased arrays to pinpoint physical noise sources, and free-field microphones to determine model noise spectra and directivity. Using high fidelity test models, the wind tunnel far-field noise measurements could be corrected for model characteristic dimensions and freestream Mach number to give accurate full scale noise spectra and directivities, an ability that makes aeroacoustic facilities unique and critical to airframe noise research. Several semi-empirical noise prediction models do exist for landing gears [27–31], but these are only useful for determining the aerodynamic noise of major components on traditional gear architectures and depend on educated users to apply “dressing/complexity” factors to compensate for noise from smaller details. As well, research is well underway on computational aeroacoustic methods [32–36], but these computationally expensive noise prediction tools are currently not self-sufficient and still require extensive empirical validation. The importance of aeroacoustic wind tunnels in investigating aerodynamic noise generation mechanisms and predicting far-field noise levels has been especially apparent in the Chapter 1. Introduction 5 recent surge in efforts to develop novel airframe noise reduction methods. With the ambitious noise emission targets described previously, researchers have been experimenting with a wide range of noise control strategies on a variety of airframe components. Passive control studies on landing gear and high-lift devices have mainly focused on using add-on treatments. Landing gear add-on fairings have yielded up to 10 dB in noise reduction [37, 38] and splitter plates between major landing gear components have resulted in up to 4 dB reduction [39, 40]. Flap side-edge fences and brush extensions have shown 5-10 dB improvements [41–43], while slat hook extensions and slat cove fillers, covers, and strips have individually provided 4-5 dB reductions [44–46]. Some researchers have also investigated active flow control strategies for noise reduction, including dielectric barrier discharge plasma actuators and blowing/suction systems. The use of plasma actuators on landing gear struts have suggested the possibility of up to 13 dB in noise reduction [47–49], while blowing systems on struts have shown 5-10 dB reductions [50, 51]. Applications of flap side-edge blowing systems have yielded 3-4 dB reductions , while slat cove suction/blowing systems have provided 2-4 dB reductions [53, 54]. Researchers have also looked at cavity tone suppression using plasma actuators [55, 56]. With the varying levels of success from the airframe noise reduction research to date, it is clear that there is still a long way to achieving the NASA and ACARE noise reduction targets. In the future, high quality aeroacoustic wind tunnels will continue to play the critical role of facilitating interdisciplinary research between airframe aerodynamics and aeroacoustics, and be in demand for the validation of novel quiet landing gear architectures and high lift device configurations. 1.3 Outline of Thesis The UTIAS AWT facility was constructed in 1975 and underwent improvements in 1985 (replacement of honeycomb and wire mesh screens) and 2011 (repair of motor windings, installation of a Variable Frequency Drive (VFD), and addition of a silencer aft of the fan/motor). This facility has been used for various studies, including graduate-level research on the acoustic excitation of Tollmien-Schlichting Waves  and the aeroacoustics of landing gears in the wake of propellers , as well as many industry-led initiatives. The purpose of this thesis is to address some key facility deficiencies and improve the aerodynamic and acoustic performance to keep up with the demands of high quality aeroacoustics research. Potential future study subjects include landing gear struts and wheels, slats, flaps, propellers, and turbojet compressor fans, among others. Chapter 1. Introduction 6 Chapter 2 provides some background information for this thesis, beginning with a survey of existing anechoic wind tunnels at other institutions, followed by a description of the existing UTIAS facility. Chapter 3 discusses the specific objectives of this thesis project and details the design rationale, construction, and installation of the various facility improvements. Chapter 4 describes the experimental procedure and presents the results of the overall facility characterization. Finally, Chapter 5 summarizes the work of this thesis and provides a list of recommendations for future consideration. Chapter 2 Background The research objectives of aeroacoustic experiments are often to characterize, understand, and predict the aerodynamic generation of noise on a model of interest. These pose many requirements on the capabilities of the research facility, the chief of which is the ability to conduct high quality aerodynamic and acoustic investigations simultaneously. This chapter discusses the rationale behind the design of anechoic wind tunnels and provides a brief survey of some of the aeroacoustic research facilities around the world, including the existing one at UTIAS at the time this thesis project began. 2.1 Anechoic Wind Tunnel Design Aeroacoustic test facilities can come in many shapes and sizes, ranging from small anechoic rooms to hemi-anechoic tunnels to hangar-like facilities (see Fig. 2.1). Their general purpose is to provide a high quality flow region to simulate operating conditions on test models, within a test environment that is also suitable for measuring the acoustics generated by the interaction between the airflow and the models. This is usually accomplished with a wind tunnel that has an open-jet test section enclosed within a larger anechoic chamber. The test models are positioned inside the test section jet flow region along with the necessary aerodynamics sensors such as pressure transducers and hot-wires. The acoustic measurement sensors (i.e., microphones) are placed outside the test section flow region at various specific locations in the surrounding anechoic chamber to collect data on the far-field directivity, intensity, and frequency spectra of the aerodynamic noise generated by the test model. Building or improving a dedicated aeroacoustics testing facility involves the consideration of many aerodynamic- and acoustic-related parameters, most of which derive from the 7 8 Chapter 2. Background requirements of the future experiments to be conducted in the facility. Within the test section, aerodynamic specifications such as usable flow area, velocity, uniformity, and turbulence intensity must be known and appropriate for the experiments at hand. In the surrounding anechoic chamber, acoustic specifications such as the anechoic cut-off limit, background noise levels, and useable anechoic volume are key. (a) Florida State University facility; 1.12 m × 0.74 m test section. Image from . (b) ISVR facility; 0.45 m × 0.15 m test section. Image from . (c) Daimler Chrysler AAWT hemi-anechoic facility; 6.9 m × 4.0 m test section. Image from . (d) German-Dutch DNW-LLF facility; 8.0 m × 6.0 m test section. Image from . Figure 2.1: Examples of anechoic wind tunnel facilities around the world. 2.1.1 Aerodynamic Specifications For airframe noise tests in a typical university laboratory setting, it is often difficult to investigate full size components at real operating conditions due to space and wind tunnel fan power limitations. As with aerodynamic experiments testing scaled models, aeroacoustic investigations must ensure that the flow regimes in the wind tunnel setting match those of real operating conditions. These flow regimes can be represented by several 9 Chapter 2. Background important similarity parameters, including the non-dimensional Reynolds, Strouhal, and Mach numbers, Re, St, and M a, respectively. The Reynolds number, the ratio of inertial to viscous forces in the working fluid, is given by: U∞ D Re = , (2.1) ν where U∞ is the freestream velocity, D is the characteristic model dimension (e.g. airfoil chord length or cylinder diameter), and ν is the kinematic viscosity of the working fluid. This similarity parameter is crucial for characterizing the aerodynamic regime around the test model because it is used to describe the boundary layer, shear layer, and wake dynamics. This information is obviously important in aeroacoustic tests as researchers aim to understand noise generation mechanisms associated with specific aerodynamic phenomena. The Strouhal number is a non-dimensional measure of the frequency of oscillatory flow phenomena, such as bluff body vortex shedding or free jet shear layer instabilities. It is given by: fD , (2.2) St = U∞ where f is the corresponding physical frequency as measured by experimentation microphones, and D is the separating distance between opposing shear layers (i.e. the diameter of a cylinder/bluff body or of a free jet). The Strouhal number characterizes shedding regimes, and is used to scale the frequencies generated by shedding phenomena. The Mach number is the ratio of inertial to elastic forces in the fluid, and is defined as: Ma = U∞ , a (2.3) where a is the speed of sound. The importance of the Mach number lies in the dependence of aerodynamic sound generation intensity on this similarity parameter; the intensity of aerodynamic noise scales with Mach number to the eighth power for mechanisms related to flow turbulence (e.g. open-jet shear layers), to the sixth power for fixed surface pressure fluctuations (e.g. periodic vortex shedding off a surface), and to the fourth power for oscillating surface fluctuations (e.g. cavity resonance) . From the above similarity parameters, D and U∞ are typically the only variables that can be controlled by the researcher, but both of their upper ends are limited by the test 10 Chapter 2. Background facility. Usually, new anechoic wind tunnel designers looking to accommodate a variety of aeroacoustic tests will attempt to maximize the test section area and freestream velocity. Of course, installation and operating costs, as well as background noise levels in the anechoic chamber, rise with test section area and velocity so consideration must be given to all of these criteria. For modifications to existing facilities where the tunnel drive fan and motor will not be changed, as is the case in this thesis, the maximum mass flowrate through the test section is fixed and hence test section area and maximum velocity are inversely proportional with one another. Therefore a balance must be struck between the two, taking into account the types of aeroacoustics research planned for the facility. With respect to the other major aerodynamics-related parameters, the goal is to create laminar, uniform and steady test section flow conditions. The spatial uniformity of the freestream velocity should be maximized across as large portion of the test section area as possible. A statistical measure of uniformity based on the root-mean-square of local time-averaged velocity deviations is given by: qP n i (Ui γ= − U )2 U , (2.4) where Ui is the local time-averaged velocity and U is the mean test section velocity. The freestream turbulence intensity (TI) should be minimized across the test section to simulate the laminar flow experienced in most real flight conditions. It is given by: TI = u0rms , U∞ (2.5) where u0rms is the root-mean-square of time-resolved velocity fluctuations at a point in space. 2.1.2 Acoustic Specifications While it is theoretically possible to conduct aeroacoustic tests on a model inside typical hard-walled wind tunnels, there are several key factors that make high quality acoustic measurements difficult to obtain within these facilities . Specifically, the closed test section creates a reverberant environment in which individual microphones have difficulty distinguishing the test model incident noise from reflections off the walls. Placing the microphones within the closed test section also exposes them to airflow and causes high levels of microphone self-noise from turbulent flow over the microphone grid and casing, which in cases of high flow velocities can decrease the signal-to-noise ratio to unusable 11 Chapter 2. Background levels. As well, there is usually limited space within closed test sections and microphones will be within the acoustic near-field of the test model, which is unacceptable in the case that measurements are needed for the study and prediction of far-field noise levels. Nevertheless, some limited forms of aeroacoustics research can be performed inside closed test section wind tunnels using microphone phased arrays and coherence/correlation analysis techniques to determine noise source locations and frequency spectra, even in the high background noise environments . However, noise intensity and directivity, which are used to predict full-scale far-field noise emission levels, are best evaluated in anechoic open-jet wind tunnels. Acoustic researchers often express noise intensity in terms of Sound Pressure Level (SPL) in the logarithmic unit decibels (dB) via the equation: SPL = 20 × log10 p0 pref , (2.6) where p0 is the measured fluctuating pressure signal, and pref is a reference pressure level, which is usually chosen to be 2×10−5 Pa (approximately the lower limit of sensitivity of the human ear). In spectral analysis, these SPLs correspond to sound intensities over particular narrow bands of frequency, and may be integrated over a large range of frequencies to yield an overall measure of sound intensity, called Overall Sound Pressure Level (OASPL). The frequency limits of integration depend on the particular study at hand; a commonly used range is 20 Hz to 20 kHz, which forms the approximate limits of human hearing. Prior to integration, researchers may also apply a correction to the SPL spectra known as A-weighting, which adjusts SPL values for the frequency-dependent loudness perception levels of the human ear (the ear is most sensitive to noise in the 1 kHz to 6 kHz frequency range). The A-weighting adjustment levels for frequencies from 20 Hz to 20 kHz are shown in Fig. 2.2. The purpose of having an anechoic chamber surrounding the open-jet test section is to provide a non-reverberant, quiet, still-air environment for noise measurement. The anechoic property is crucial for simulating an acoustic free-field environment, and is typically provided by wedge-shaped sound absorbent material covering all chamber interior surfaces (minus the floor for hemi-anechoic chambers). Materials for wedge construction include polyurethane foam, fibreglass, rock-fibre, and proprietary acoustically-absorbent materials, among others. The shape of these wedges are designed such that most incident acoustic waves do not reflect back directly into the test section measurement region; any sound waves with one-quarter wavelength smaller than the depth of the wedges are 12 Chapter 2. Background 10 SPL attenuation [dB] 0 −10 −20 −30 −40 −50 10 2 10 3 10 4 Frequency [Hz] Figure 2.2: A plot of the A-weighting SPL across the frequency range 20 Hz to 20 kHz. Data from . absorbed. That is, the theoretical frequency cutoff limit above which the chamber is considered ‘anechoic’, is determined from the wedge depth, Lw , via: fc = a . 4 × Lw (2.7) Aeroacoustic researchers are often interested in surveying noise frequency spectra for interesting sound intensity peaks or other phenomena, in search for connections to aerodynamic phenomena. There may also be motivation to determine how the noise under study may affect humans if it is within their sensible frequency range. The anechoic cutoff frequency is important in determining the bandwidth of frequencies permitted into this spectral analysis, and should be made as low as possible. However, increasing wedge depth increases installation cost and, in the case of retrofitting an existing facility, decreases the usable anechoic volume as described later in this section. Even with acoustic reverberations eliminated, high background noise levels in the anechoic chamber will significantly affect the ability of microphones to discern noise generated by the test model from that of other sources. The signal-to-noise ratio of measurement microphone systems should be at least 10 dB above the background noise to ensure high quality sound intensity measurements . Some typical background noise sources include the wind tunnel drive fan and other components, model and apparatus support stands in the flow, the open-jet shear layer, vibrating surfaces, and general environmental noise. Chapter 2. Background 13 It is crucial in the design and construction of anechoic wind tunnels to minimize these extraneous noise sources so as to maximize the signal-to-noise ratio of the microphone measurements. Some strategies employed by past researchers to minimize background noise include baffles and silencers to stop the transmission of fan noise into the test section and heavily insulated jet collectors to decrease the noise generated from turbulent jet shear layer impingement [59,68,69]. Unwanted vibrations must be eliminated by securing or stiffening parts, and environmental noise should be sealed out using thick concrete wall construction for the chamber. As well, model and apparatus support struts that extend into the test section flow should have cross-sections in the form of the M-H airfoil, which was specially designed to exhibit minimal shedding (thus low self-noise) while providing stiffness through a 30% thickness-to-chord ratio . In order to accommodate far-field noise measurements in the anechoic chamber, the physical dimensions of the chamber must be large enough to accommodate an arc of far-field microphones positioned around the test model. These microphones must be in the acoustic far-field of the model, which is defined as the region that is at least one acoustic wavelength and two noise source dimensions away from the test model . These microphones must be within the usable anechoic volume of the facility, away from the acoustically absorbent wedges on the walls to avoid interference with the noise measurements. The angular position of the measurement arc would depend on the test at hand, but typical arcs may extend from 20◦ to 135◦ relative to the jet centreline downstream of the test model. One more acoustic specification that some, but not all, researchers quote is RT60 , defined as the time it takes for the SPL of an impulse noise and subsequent reverberations to decrease by 60 dB from the peak of the impulse. This is a measure used mostly in regular rooms and concert halls, and is related to the size and shape of the room and the quality of the anechoic treatment on surfaces. Lowering reverberation time means improving the anechoic property of the test facility. 2.2 Survey of Existing Facilities Anechoic wind tunnel facilities are used in various settings around the world, including at aircraft and auto manufacturer test facilities, national research centres, and universities. These facilities can be categorized into open- and closed-circuit types. Open-circuit tunnels draw in and expel out air to the surrounding environment, while closed-circuit tunnels continuously recirculate air with limited exchange with the surrounding envi- Chapter 2. Background 14 ronment. Open-circuit tunnels require less space and incur substantially lower costs to construct. Closed-circuit tunnels have more parts in the form of turning vanes and air return sections, but are able to offer higher flow quality and achieve test section velocities equivalent to open-circuit tunnels using less power and generating less background noise. Tests involving combustion or smoke visualization should be performed in open-circuit tunnels. The type of tunnel driver also has a profound impact on tunnel operation, and existing facilities can be categorized into blowdown, blower, or suction driver types. Blowdown tunnels supply air from a pressurized reservoir, while the blowing and suction type tunnels drive air continuously using a fan upstream and downstream of the test section, respectively. The blowdown tunnels are limited in test runtime, but usually offer superior background noise levels to the continuously driven tunnels because there is no fan and motor noise. Blower type open-jet test sections have inferior flow quality in the test section and may exhibit low-frequency flow instabilities from collector spillage. Suction type open-jet test sections operate below atmospheric pressure, but exhibit less jet spreading and offer higher flow quality than the blower type tunnels do . For the tunnels with variable pitch fans and variable rotational speed motors, electrical power consumption and background noise may be improved via an adjustment of these settings for the lowest motor RPM . In the past two decades, a handful of new facilities have been built and existing ones improved [67,68]. The auto industry has taken the lead in constructing larger and quieter facilities, due in large part to their ability to pay for the enormous costs of installation and operation. For example, the Daimler Chrysler Aeroacoustic Wind Tunnel (AAWT) in Detroit, U.S.A., which has one of the largest test section areas among anechoic wind tunnels at 28 m2 and is capable of velocities up to 71 m/s, cost US$37.5 million to construct . Universities have also pushed the envelope with new aeroacoustic test facilities that have impressive specifications, albeit on a smaller scale. An example is the tunnel at the Institute for Sound and Vibration Research in Southampton, U.K., which offers a maximum test section velocity of 100 m/s within an anechoic chamber that has one of the lowest background noise levels in the world . Table 2.1 compares key facility dimensions and specifications for the aerodynamic and acoustic capabilities of leading facilities around the world. Tunnel type Open-jet, closed-circuit Open-jet, closed-circuit Open-jet , closed-circuit Open-jet, closed-circuit Hessert Center, Open-jet, Notre Dame open-circuit, suction University of Open-jet, Florida open-circuit, suction ISVR Open-jet, Southampton blowdown UTIAS Open-jet, (old facility) open-circuit, suction Florida State Open-jet, University open-circuit, suction Virginia Tech Kevlar wall, closed-circuit Brandenburg Univ. Open-jet of Tech. at Cottbus open-circuit, blower Universitat Open-jet, Siegen open-circuit, blower NASA Langley Quiet Flow Facility Ford DTF WT8 Daimler Chrysler AAWT Audi Facility name Maximum velocity [m/s] 58 54 71 83 28 76 100 60 75 57 72 24 Test section dimensions [m×m] 0.91 × 0.61 5.56 × 3.34 6.9 × 4.0 3.94 × 2.8 0.61 × 0.61 1.12 × 0.74 0.45 × 0.15 0.7 diameter 1.22 × 0.91 1.83 × 1.83 0.15 × 0.12 0.13 × 0.13 0.40 0.20 0.03 0.12 0.5 0.1 0.07 0.04 0.30 0.16 0.34 4.5 × 3.2 × 2.9 5.0 × 4.5 × 3.9 7.0 × 4.2 × 2.4 4.8 × 4.5 × 2.7 7.33 × 7.33 × 5.50 4.9 × 4.9 × 2.4 5.5 × 5.0 × 2.3 9.1 × 7.3 × 3.7 - 21 × 15 × 10 - Turbulence Chamber intensity dimensions [%] [m×m×m] 9.1 × 6.1 × 7.6 125 500 190 250 140 80 100 125 - - - Anecoic freq. cut-off [Hz] - Chapter 2. Background 15 Table 2.1: Survey of select aeroacoustic research facilities around the world. Data from [59, 60, 73–76]. 16 Chapter 2. Background In order for background noise level data from all the different facilities to be comparable, they must be converted to a common basis with respect to microphone measurement distance from the test section centreline as well as the open-jet cross-sectional area. The conversion is given by : OASPLconv = OASPLfac + 20 × log10 rfac rref − 10 × log10 Afac Aref , (2.8) where OASPLfac is the reported background sound pressure level, rfac is the original microphone measurement distance from the open-jet centreline, rref is the reference microphone distance from the open-jet centreline (set at 2 m in this case), Afac is the open-jet crosssectional area of the facility, and Aref is the reference cross-sectional area (set at the new test section inlet area of 0.36 m2 , for ease of comparison with results later). Figure 2.3 summarizes the background noise levels over a range of test section velocities for the aeroacoustic facilities of interest. Background A−weighted OASPL [dB] 80 70 60 Audi DaimlerChrysler Ford DTF Florida State ISVR Southampton Univ. Cottbus Univ. Florida Univ. Siegen Univ. Notre Dame Old UTIAS 50 40 30 20 0 10 20 30 40 50 60 Test section velocity [m/s] 70 80 90 100 Figure 2.3: Comparison of overall background sound pressure levels of select anechoic wind tunnels around the world, over their range of capable test section velocities. Data adapted from [59, 60, 72–74]. Chapter 2. Background 2.3 17 Existing UTIAS facility The existing aeroacoustic research facility at UTIAS consists of an open-circuit, suctiontype wind tunnel with an open-jet test section enclosed inside an anechoic chamber. The tunnel is driven by a nine-blade axial fan powered by a 150 h.p. A/C induction motor that is capable of rotational speeds up to 1800 RPM. The test section inlet diameter is 0.70 m, with a maximum velocity of 60 m/s and freestream turbulence intensity of less than 0.5%. The anechoic chamber is lined with fibreglass wedges with tip-to-base depth of 0.61 m, yielding a theoretical anechoic cut-off frequency of 140 Hz. The chamber inner dimensions as measured from wedge-tip to wedge-tip are 4.9 m × 4.9 m × 2.4 m (L × W × H). The wind tunnel is situated outdoors, with its front-end attached to the side of a geodesic dome that is 55 m in diameter. Fig. 2.4 provides views of the existing facility from outside and inside the geodesic dome, as well as a schematic plan view. The geodesic dome is used by the UTIAS robotics research groups as a Martian surface simulator, and also doubles as a sort of supply plenum for the wind tunnel. Air from inside the dome is guided into the tunnel by a square bell-mouth section measuring 2.74 m by 2.74 m and then sent through a series of flow conditioning screens. Unfortunately, activity in the dome creates large amounts of airborne dust, which over time clogs up the flow conditioning screens and renders them ineffective. The fine dust particles that are not caught in the screens enter the test section, where they may accumulate on and damage sensitive instrumentation. The first of the flow conditioning screens is an aluminum honeycomb mesh with hexagonal cells of 1.59 mm width and 50 mm depth used to straighten the flow. This honeycomb screen is bound by two 120-mesh (120 openings per inch length) stainless steel wire screens to keep it in place. The air then passes through a 200 mm settling passage before entering a series of 4 stainless steel, 54-mesh, 0.10 mm-diameter wire screens, with 50 mm streamwise separation between each of them. The purpose of these wire screens is to reduce freestream turbulence intensity and promote spatial flow uniformity. The issue with these existing honeycomb and wire mesh screens is that their physical specifications are too small by the standards of generally accepted wind tunnel design rules. As a result, they are less effective at reducing freestream turbulence intensity and incur excess static pressure drops, which negatively impact the maximum achievable test section velocity. After the conditioning screens, the flow accelerates through a square contraction section with a contraction ratio of 9.3, followed by a square-to-circle conversion section. After 18 Chapter 2. Background (a) Outdoors view of the faciity. (b) Inside the geodesic dome looking at the existing tunnel bellmouth inlet. (c) Schematic plan view. Figure 2.4: Views of the existing UTIAS AWT facility. 19 Chapter 2. Background a circular tube section, the flow accelerates through a cone-shaped nozzle that sharply converges the flow from 0.9 m to 0.7 m in diameter just before entering the test section inside the anechoic chamber. The open-jet section from nozzle exit to collector entrance is 2.78 m long. After entering the 0.9 m diameter collector, the flow passes through diffuser and silencer sections, followed by the driving fan section and finally out into the atmosphere. An aerodynamic characterization of the existing test section open-jet was performed by Chekiri . Figure 2.6 shows time-averaged freestream velocity surveys across the diameter of the open-jet, for test section streamwise locations from 0 mm to 1190 mm (i.e., from 0 to 1.70 nozzle diameters). The wake surveys of the first several streamwise locations clearly show accelerated regions near the shear layers at the edges of the jet, and this is due to the contracting nozzle just before the entrance of the test section. As a result, the spatially uniform flow velocity area is reduced, and the usable test section diameter is restricted to 0.5 m. Figure 2.5a shows the freestream turbulence intensity along the jet centreline for freestream Mach numbers 0.10 and 0.15. Figure 2.5b shows the variation of freestream turbulence intensity with radial position in the openjet, for various streamwise positions in the test section at freestream Mach of 0.10. It is clear that as the flow moves downstream from the nozzle exit, the converging open-jet shear layer causes a significant increase in freestream turbulence intensity in the test section. Improving the test section area, flow velocity, and flow quality are crucial to future experimental research in the facility and thus is one of the main areas of focus in this thesis. (a) Centreline TI versus streamwise location at freestream Mach 0.15 and 0.10. (b) TI versus radial location at various streamwise locations at freestream Mach 0.10. Figure 2.5: Freestream turbulence intensity plots for old AWT. Figures from . Chapter 2. Background 20 Figure 2.6: Test section time-averaged velocity surveys over a range of streamwise locations, at freestream velocities of (a) Mach 0.10 and (b) Mach 0.15 for old AWT. Figure from . 21 Chapter 2. Background The anechoic chamber housing the test section was constructed from 300 mm thick concrete walls, and all hard interior surfaces were covered in wedge-like fibreglass blocks (see Fig. 2.7). These wedge blocks were constructed from individual 51 mm thick yellow fibreglass batts friction-fit into “step-function” wedge-like shapes, instead of the more commonly installed triangular prism wedge shape. The fibreglass wedges were not secured by any form of metal mesh or cloth material and over time many of them lost their form or completely dislodged from the ceiling and walls due to gravity and perturbations from recirculating chamber flow during tunnel operation. As well, the bare fibreglass wedges released many loose fibres into the air, which posed a serious health risk to researchers who worked inside the chamber. (a) Inside the anechoic chamber looking at the test section. (b) Drawing of two pairs of fibreglass wedges. Dimensions in centimetres. Figure from . Figure 2.7: Views of the fibreglass wedges in the old anechoic chamber. Some background SPL characterization was performed on the old anechoic chamber and is summarized in the SPL spectra in Fig. 2.8. For comparison, the background SPL spectra of the ISVR anechoic wind tunnel is given in Fig. 2.9. At the maximum freestream velocity shown for the existing UTIAS facility, 61.2 m/s, the background noise levels in the UTIAS facility are between 15-25 dB higher than those in the ISVR facility across the range of frequencies from 100 Hz to 10,000 Hz. This relatively poor acoustic performance in the existing tunnel limits the signal-to-noise ratio of microphone measurements and hampers the ability of researchers to collect high quality acoustic data. For this reason, chamber acoustics is another main area of focus in the current facility improvement efforts. 22 Chapter 2. Background 80 19.0 m/s 29.6 m/s 37.7 m/s 44.9 m/s 48.3 m/s 54.4 m/s 61.2 m/s 60 SPL [dB ref 2x10 −5 Pa] 70 50 40 30 20 10 2 10 10 3 10 4 Frequency [Hz] Figure 2.8: Background SPL narrowband frequency spectra of old AWT, for select freestream velocities from 19.0 m/s to 61.2 m/s. Data adapted from . Figure 2.9: Background SPL narrowband frequency spectra of ISVR anechoic wind tunnel, for select freestream velocities from 33.1 m/s to 99.6 m/s. Figure from . Chapter 3 Facility Updates 3.1 Objectives The specific thesis objectives and associated performance metrics are discussed below. One of the overarching objectives was to keep the total cost below CA$40,000. 3.1.1 Aerodynamic objectives Table 3.1 lists the aerodynamics-related targets of the facility update efforts. The test section freestream velocity and dimension targets were driven by minimum Reynolds number requirements and maximum model flow blockage limits. The freestream turbulence intensity and spatial uniformity targets were motivated by the desire to provide as high quality flow in the test section as possible. Table 3.1: Summary of aerodynamic improvement objectives for the UTIAS AWT facility. Objective Target Test section maximum freestream velocity Increase to at least 70 m/s from current 60 m/s Test section dimensions The usable cross-sectional area should be at least 0.5 m in diameter and the streamwise length be at least 1.2 m Test section freestream Below current value of 0.5% within first 1.5 test turbulence intensity section widths downstream of inlet Test section spatial flow uniformity At least 99% 23 24 Chapter 3. Facility Updates 3.1.2 Acoustic objectives Table 3.2 lists the acoustics-related targets of the facility update efforts. Table 3.2: Summary of acoustic improvement objectives for the UTIAS AWT facility. Objective Target Background noise level 15 dB SPL reduction from current levels across frequencies 200 Hz - 20 kHz Anechoic limit cut-off frequency Below 200 Hz Usable anechoic area around test section At least enough space to place an arc of microphones 2 m from the jet centreline Reverberation time Below 0.2 s to 60 dB SPL decay The expected experimental microphone measurements across the frequency bandwidths of interest must achieve a signal-to-noise ratio of at least 10 dB above the tunnel background noise level. The NACA0012 airfoil self-noise is commonly used as a benchmark to approximate experimental noise levels, and a semi-empirical noise prediction method for this airfoil was developed by Brooks et al. . Figure 3.1 shows both predicted and experimental (from ISVR) noise spectra for a NACA0012 airfoil with chord length 0.15 m at velocities 33.3 m/s and 80 m/s, values which approximate the freestream velocity lower and upper bounds for future UTIAS AWT experiments. Comparisons between these airfoil noise spectra and the UTIAS AWT background SPL spectra in Fig. 2.8 reveals that the UTIAS AWT requires a large background noise level improvement of around 15 dB across all frequencies in order to achieve the 10 dB target for signal-to-noise ratio. The target for the anechoic cut-off frequency is controlled by the depth of the acoustic wedges, which was determined with primary consideration given to the space available inside the existing anechoic chamber and the cost of acoustic wedges (they were by far the most expensive item on the budget). The upper limit of 200 Hz is within the range of other facilities around the world, and provides adequate bandwidth for future scaledmodel aeroacoustic tests in the facility. The anechoic space requirement was set to ensure that an arc of free-field microphones can be situated in the acoustic far-field of any aerodynamic noise with frequencies above 200 Hz generated from sources within the test section. Finally, the reverberation time should be as low as possible, with 0.2 s considered Chapter 3. Facility Updates 25 Figure 3.1: Comparison of predicted and measured narrowband SPL spectra for NACA0012 airfoil with chord length 0.15 m at freestream velocities of 33.3 m/s and 80 m/s in the ISVR anechoic wind tunnel. Also shown is the ISVR background narrowband SPL at 33.3 m/s and 80 m/s. Figure from . a reasonable upper limit for anechoic chambers of this size. In comparison, the bare concrete room housing the anechoic chamber has a reverberation time of approximately 1.5 s. In order to achieve these thesis objectives, several key areas in the UTIAS AWT facility were targeted for improvement (see Fig. 3.2). First, the bell-mouth intake surfaces were refinished and the flow conditioning screens were replaced. Second, the square-to-circle flow section and nozzle were replaced with a square secondary contraction. Third, the fibreglass wedges in the test chamber were replaced with polyurethane foam wedges, and a false door was installed inside the existing chamber entrance. Finally, baffles were added to the front-end of the facility. The following sections describe in detail the rationale, design, and implementation of these facility updates. Chapter 3. Facility Updates 26 Figure 3.2: Schematic plan view of updates to UTIAS AWT facility. Front-end baffles were added, flow conditioning screens were replaced, existing nozzle was replaced with a secondary contraction, and new anechoic wedges and false door were installed. 3.2 Flow Conditioning The wind tunnel front-end, which consists of a bell-mouth, a honeycomb screen, and three wire mesh screens, performs important flow intake and conditioning functions on the air heading into the test section. The bell-mouth can be separated into four corner parts and is made of painted steel sheets that are about 2 mm thick (see Fig. 3.3). The surfaces of the old bell-mouth pieces were very rough as a result of paint stripping and severe rusting. As a first step towards improving the flow conditioning, these surfaces were ground down to the bare steel and then repainted with three coats of Tremclad exterior rust-proof paint. Figure. 3.4 shows an image of the front-end with the resurfaced bell-mouth. The honeycomb screen primarily serves to straighten the flow going into the test section. This is achieved as long as there are at least 150 cell openings across the width of the flow section . As well, the honeycomb screen can help reduce freestream turbulence intensity if the cell depths are approximately 6-8 cell diameters. In order to minimize the pressure drop, the cross-sectional shape for each cell should be hexagonal . Following these guidelines, the new honeycomb design has aluminum hexagonal cells with 9.53 mm cell width and 76.2 mm cell depth. The honeycomb screen was installed with wire mesh screens (discussed later) attached on both sides to keep it in place. Figure 3.5 provides a scale comparison of the old honeycomb cells with the new ones. Chapter 3. Facility Updates Figure 3.3: Photos of the old bell-mouth under construction. Figure 3.4: Photo of the tunnel front-end with the resurfaced bell-mouth. 27 28 Chapter 3. Facility Updates (a) Old honeycomb screen with 1.59 mm cell width and 50.8 mm cell depth. (b) New honeycomb screen with 9.53 mm cell width and 76.2 mm cell depth. Photo taken with 18-mesh, 0.229 mm diameter wire screen in front of honeycomb. Figure 3.5: Close-up scale comparison of the old and new honeycomb mesh screens. The purpose of the wire mesh screens is to cause a static pressure drop that makes the flow velocity profile more spatially uniform and reduces boundary layer growth . The freestream turbulence intensity can also be reduced, given a minimum settling spacing of 500 wire diameters downstream of each screen to allow the screen-generated turbulence to dissipate . Mesh screens are characterized by their open-area ratio, β = (1 − M d)2 , (3.1) where M is the number of mesh openings per unit length and d is the wire diameter, and by their pressure drop coefficient as formulated by Wieghardt’s empirical formula , − 1 ∆P 1 − β Ud 3 K= = 6.5 , q β βν (3.2) where ∆P is the static pressure drop, q is dynamic pressure, U is freestream velocity entering the screen, and ν is the kinematic viscosity. Wind tunnel design guidelines [78, 82] suggest β > 57% to avoid instabilities associated with jets forming from the individual screen openings. As well, a series of 3-5 screens with K = 1.6 − 2.0 for each is suggested to achieve flow uniformity while maintaining an acceptable overall pressure drop. The maximum streamwise velocity of air entering the bell-mouth inlet was measured to be approximately 3.7 m/s. 29 Chapter 3. Facility Updates The old set of screens consisted of M = 120-mesh preliminary screens on both sides of the honeycomb grid, followed 200 mm downstream by four 54-mesh main screens spaced 50 mm apart from each other. This old set-up needed to be replaced because the low open-area ratio of the 120-mesh screens compounded with the heavy accumulation of dirt on all of the screens caused severe flow blockage. Following the design guidelines, a new set of five screens were made. Two 18-mesh, 0.229 mm diameter wire preliminary screens were installed on both sides of the honeycomb, followed 200 mm downstream by three 34-mesh, 0.165 mm diameter wire main screens spaced 76 mm apart from each other. Table 3.3 lists the specifications of the old and new screens. Table 3.3: Comparison of new and old wire mesh screen specifications. Screen type 3.3 M (per inch) d [mm] β [%] K New preliminary 18 0.229 70.2 0.92 Old preliminary 120 0.067 47.0 4.83 New main 34 0.165 60.7 1.73 Old main 54 0.086 66.8 1.55 Secondary Contraction The facility has a square primary contraction situated in the concrete wall separating the anechoic chamber from the interior of the geodesic dome. In the old set-up, the primary contraction connected to a square-to-circle conversion section, followed by a straight circular section, and capped off by a conic nozzle at the test section inlet. This nozzle produced an undesirable region of accelerated flow near the shear layers, with a contracted uniform velocity test section area about 500 mm in diameter. The new design maintains a square cross-section throughout; the primary contraction is followed by a straight square section, which is in turn followed by a secondary contraction that ends in a tangential exit to get the desired “top-hat” velocity profile in the test section. Figure 3.6 shows images of the old nozzle and new secondary contraction sections. There are several important factors to consider in the design of a wind tunnel contraction, including flow separation, crossflow, and flow uniformity [75, 82, 83]. Flow separation at the entrance of the contraction can be caused by an abrupt increase in the wall pressure Chapter 3. Facility Updates 30 (a) Old upstream flow section with square-to-circle conversion, then circular straight section, capped off with conic nozzle at test section inlet. (b) New upstream flow section with straight square section followed by secondary contraction with tangential exit lip at test section inlet. Figure 3.6: On the left, comparison of new and old flow sections leading to the test section. On the right, renders of the shear layer and test section velocity profiles expected from the two different test section inlets. coefficient in the convex (fore) section of the contraction. This can be avoided by using a longer convex section and a lower order polynomial to shape the walls in that section. Crossflow occurs due to having a non-axisymmetric cross-section, but this can be reduced via a longer contraction length and having a square cross-section. Since the contraction is the last tunnel component that the flow passes through before entering the test section, it is critical that flow uniformity at the contraction exit be maximized. This can be done using a higher order polynomial to shape the walls in the concave (aft) section of the contraction, and ensuring the exit lip is tangential to the desired direction of the test section freestream flow. 31 Chapter 3. Facility Updates The wall profile chosen for the secondary contraction consists of two matched-polynomial curves, with a 3rd order polynomial for the fore section and a 8th order polynomial for the aft section (see Fig. 3.7). The convex- and concave-section match point, where the first- and second-derivatives of the polynomial curves were set equal to each other, was set at 70% of the contraction length. The contraction ratio for this new section alone is 1.91, while the overall ratio for the primary and secondary contractions combined is 20.9. A side benefit of this relatively high contraction ratio is a reduction in the longitudinal freestream turbulence intensity going into the test section. 500 Wind tunnel height/width [mm] 400 300 200 100 0 -100 -200 -300 -400 -500 0 200 400 600 800 1000 Streamwise location [mm] Figure 3.7: Plot of the secondary contraction wall profile. Each curve is constructed from 3rd and 7th order polynomials with the matching point at the 70% streamwise location. The secondary contraction section was manufactured in-house via a fibreglass hand-layup method on a positive mould. This mould consisted of a wooden frame supporting Owens Corning extruded polystyrene rigid foam insulation sheets, which were formed into the contraction wall profile using a computer-guided hot-wire foam cutter. The formed foam sheets were sealed with three coats of a thick epoxy mixture, consisting of West Systems 105 Resin, 205 Fast Hardener, and 406 Colloidal Silica, to yield a hard and smooth moulding surface. To help with mould release, two layers of TR104 carnauba blend wax were applied to the mould surface using a soft cloth, followed by a light spray of Polytek Pol-Ease 2300 multipurpose silicone release agent. The actual contraction section walls were made from a sandwich composite of 5.8 oz. woven fibreglass ‘E’ cloths and 1.59 mm 32 Chapter 3. Facility Updates thick balsa sheets. These were formed on the mould by first curing two layers of thick epoxy mixture, followed by a fibreglass cloth and epoxy, then a balsa sheet and epoxy, and then another fibreglass cloth and epoxy, and then sealed with two layers of epoxy. Two 100 mm wide, 12.7 mm thick plywood ribs in the shape of the contraction profile were epoxied onto the outer surface of each wall for structural stiffening. As well, a 50 mm wide, 12.7 mm thick plywood flange was epoxied onto the downstream opening of the contraction to reinforce the square shape and provide a smooth tangential exit into the test section. Finally, a 100 mm wide, 3.18 mm thick steel flange was epoxied onto the upstream opening of the contraction so that it could be bolted onto the square flow section extending from the primary contraction. Figure 3.8 shows some images of the secondary contraction construction process. (a) Surface polishing for half the mould. (b) Full fibreglass and balsa contraction on mould with metal flange on back. Reinforcing joints with fibreglass strips. Figure 3.8: Photos of the secondary contraction construction process. 3.4 Acoustic Wedges Covering the walls with anechoic wedges serves to minimize acoustic reverberation in the chamber. This is achieved via two mechanisms: one is the pseudo-random wall geometry that minimizes the reflection of sound waves back into the chamber, and the other is the large surface area of anechoic material that absorbs the sound waves. Fig. 3.9a shows the new anechoic wedges fully installed in the chamber. 33 Chapter 3. Facility Updates (a) Photo of inside the anechoic chamber looking at the test section inlet. (b) Drawing of two pairs of polyurethane wedges. Dimensions in milimetres. Figure 3.9: Views of the new polyurethane wedges in the anechoic chamber. The shape and dimension of the wedges depend on several design factors. Ideally, the shape would be conical so that there are no parallel surfaces to produce acoustic reflections. However, due to wedge stiffness, cost, manufacturing, and installation considerations, a triangular prism shape was chosen (see Fig. 3.9b). The wedge dimensions were chosen to minimize the anechoic cut-off frequency, while satisfying the anechoic volume and material cost constraints. With the chosen wedge dimensions shown in Fig. 3.9b, the anechoic chamber dimensions ended up being 4.88 m × 4.88 m × 1.83 m as measured from wedgetip to wedge-tip, and the theoretical anechoic cut-off frequency is approximately 150 Hz. The acoustic absorption properties of the wedge material can be described by its sound absorption coefficient, which is defined as the ratio of absorbed sound energy to incident sound energy on the material surface . Anechoic wedges are commonly made of porous materials due to their ability to absorb incident sound waves by incurring viscous and momentum losses on the oscillating air molecules within the numerous irregular pores in the material . A polyurethane open-cell foam was sourced from Foamite, a local foam mattress manufacturer, for the new wedges (see Appendix A for a material data sheet for this foam). This foam material poses a much lower health risk because it does not shed minuscule fibres into the air. An acoustic absorptivity test conducted at Bombardier Inc. by Raymond Wong using the Impedence Tube Method (ASTM C384) found that the foam material has a superior acoustic absorption coefficient compared to the old fibreglass material (see Fig. 3.10). Polyurethane open-cell foam sheets with a thickness of 38 mm were also purchased to line the chamber wall edges, wind tunnel flow sections, and other Chapter 3. Facility Updates 34 hard surfaces in the chamber that could not be covered with wedges. These foam sheets had ‘egg-crate’-shaped surfaces to help reduce acoustic reverberation (see Fig. 3.11 for the absorption coefficient spectrum). Figure 3.10: Acoustic absorptivity spectra for fibreglass and open-cell polyurethane foam. Data acquired by Bombardier using Impedance Tube Method (ASTM C384). Figure 3.11: Acoustic absorptivity spectrum for open-cell polyurethane foam eggcrate sheet, 38 mm thickness. Data from  using Reverberation Room Method (ASTM C423). Chapter 3. Facility Updates 35 Before installing the anechoic wedges, the bare anechoic chamber was first framed with beams across the ceiling and studs along the walls (see Figs. 3.12-3.13). All contact points between wood and concrete were protected by plastic vapour barrier sheets. The wall studs were secured at their top- and base-plates, with an approximately 15 mm air gap between the studs and concrete walls to prevent vibration noise. The ceiling was then sheeted with 6.35 mm thick oriented strand board and the walls with 6.35 mm thick plywood strips (see Fig. 3.14). Finally, the anechoic wedges were attached to the sheeted ceiling and walls, and the eggcrate foam sheets were attached to the remaining exposed surfaces, both using 3M Super-77 spray adhesive (see Fig. 3.15). A false door covered with wedges was also installed just inside of the existing metal door entranceway. In comparison to the old design that only had one layer of fibreglass covering the metal door, the false door provides additional acoustic attenuation of outside noise sources as well as decreased acoustic reverberation from inside noise sources. This false door swings inward and has wedges positioned on it such that they interlock with wedges on the wall when the door is fully opened (see Fig. 3.16). Figure 3.12: Chamber construction photo: Beginning to frame the bare concrete chamber. Chapter 3. Facility Updates Figure 3.13: Putting up the last studs. Figure 3.14: Sheeting the ceiling and walls. 36 Chapter 3. Facility Updates Figure 3.15: Installing wedges and building frame for false door. Figure 3.16: View of the completed chamber and false door. 37 Chapter 3. Facility Updates 3.5 38 Front-end Baffles The final major addition to the aeroacoustic facility was the set of acoustic baffles installed upstream of the wind tunnel bellmouth inside the geodesic dome. These baffles serve as silencers by blocking direct line of sight from the interior of the geodesic dome into the anechoic chamber. Seven baffles were installed in a staggered manner (see Fig. 3.17) to maximize the total airflow area into the tunnel bellmouth. Each baffle is 100 mm thick and extends from the gravel floor to the ceiling of the geodesic dome. They consist of wooden frames filled with 100 mm thick fibreglass batts that were recycled from the old anechoic chamber, and covered in black broadcloth to keep the fibreglass in place. Figure 3.18 shows the completed set of baffles. Figure 3.17: Schematic plan view of the front-end acoustic baffles. Dimensions are in metres. 39 Chapter 3. Facility Updates (a) Left-hand side view. (b) Head-on view. (c) Right-hand side view. Figure 3.18: Photos of the new front-end acoustic baffles as viewed from inside the geodesic dome looking towards the wind tunnel inlet. Chapter 4 Facility Characterization 4.1 Flow Speed The wind tunnel test section freestream velocity was calibrated against the frequency output of the tunnel motor VFD. The VFD was controlled via an analog voltage output from a National Instruments (NI) BNC-2120 terminal block connected to a National Instruments PCIe-6361 data acquisition card (DAQ). This voltage signal was in the range of 0-10 V, corresponding linearly to the limits of the VFD frequency output range of 0-60 Hz. The frequency output is also controllable via a keypad terminal on the VFD unit. The freestream flow velocity was measured using a 3 mm diameter pitot-static tube situated at the centre of the test section inlet plane. This pitot tube was connected via Tygon tubing to a MKS Baratron 120AD differential pressure transducer, which has a maximum measurement range of 100 torr with an accuracy of 0.25% of reading. Freestream air temperature was acquired using a T-type thermocouple. All signals were acquired by the NI PCIe-6361 DAQ via a BNC-2120 terminal block. The wind tunnel axial fan has a hub diameter of 1.38 m and a tip diameter of 2.10 m. There are a total of 9 blades with hub-to-tip twist angles of approximately 20◦ . Each blade can be individually set to equally-spaced pitch angle settings labelled number “-3” to “6”, corresponding roughly to pitch angles of -12◦ to 42◦ at three-quarters blade radius relative to the fan plane. The freestream velocities were measured at the range of blade angle settings between -3 and 5. Setting number 6 could not be physically obtained because the trailing edges of the blades come into contact with the motor housing behind the fan hub. The differences in freestream velocity between the different blade angle settings were negligible, but the motor rotational rate did decrease by 2% at the highest blade angle compared to the lowest angle. This would result in slightly lower power consumption when 40 41 Chapter 4. Facility Characterization Mean freestream velocity [m/s] operating the wind tunnel, as well as slightly lower fan and motor noise levels transmitted into the anechoic chamber (discussed later). Thus, the fan blades were fixed at setting number 5 for the remainder of the tests in this thesis. Fig. 4.1 shows the calibration points and linearly fitted curve for the test section freestream velocity. 80 60 40 20 0 0 10 20 30 40 50 60 VFD setting [Hz] Figure 4.1: Freestream velocity versus VFD output frequency for fan blades at angle setting number ‘5’. Equation for line of best fit is: U∞ = (1.27 m s−1 Hz−1 )VFD−0.14 m s−1 , with R2 of the linear fit at 0.9998. 4.2 Flow Uniformity The spatial uniformity of the test section freestream flow was characterized using the flow velocity measurement system described above. The pitot-static tube in this case was mounted on top of a two-axis linear traverse system capable of placing the probe tip anywhere within an area from 0 mm to 2200 mm downstream of the test section inlet and up to 400 mm cross-stream distance from the test section centreline. The two traverses were driven by stepper motors powered by a Motion Group MMC 4S-SC Motor Controller. The spatial resolution of the streamwise traverse was approximately 0.032 mm/full motor step, while that of the cross-stream traverse was approximately 0.0064 mm/full motor step. As a first step, the static pressure along the test section centreline was measured to check if there were any large pressure gradients in the test section (e.g., caused by the jet collector). The pitot-static tube was placed on a linear streamwise traverse with the 42 Chapter 4. Facility Characterization probe tip at the jet centreline pointing upstream. One end of the differential pressure transducer was connected to the static-pressure port of the traversing probe, while the other end was connected to a static pressure port at the test section inlet. The probe was traversed the 2200 mm length of the open jet from inlet to collector, in 100 mm increments. Pressure data were acquired at a freestream velocity of 55 m/s for 60 s at a sampling rate of 4000 Hz at each measurement location, and are presented in Fig. 4.2 in terms of coefficient of pressure, given by: Cp = Ps − P∞ , q∞ (4.1) where Ps is the static pressure reading at the traversing probe, P∞ is the freestream static pressure at the inlet, and q∞ is the freestream dynamic pressure at the inlet. As expected, there is a slight adverse pressure gradient moving downstream from the inlet due to the streamlines expanding towards the collector. The pressure coefficient within 2 test section widths downstream of the inlet remains below an acceptable value of 1%. 0.05 0.04 0.03 Cp 0.02 0.01 0 -0.01 -0.02 0 0.5 1 1.5 2 2.5 3 3.5 4 Streamwise location x/D Figure 4.2: Variation in static pressure coefficient along the open-jet centreline from inlet to collector. The freestream velocity surveys were taken in the mid-height plane of the test section in the form of 6 cross-stream profiles, every 300 mm from 0 mm to 1500 mm downstream of the inlet. Each of the 6 profiles consisted of measurement points along a 800 mm line centred on the test section centreline, with 10 mm cross-stream spacing between each point from 0 mm to 200 mm, 40 mm spacing from 200 mm to 600 mm, and then 10 mm spacing from 600 mm to 800 mm. These profiles for freestream velocities of 75 m/s, 55 m/s, and 35 m/s are presented in Fig. 4.3. The profiles at each streamwise location are generally consistent across all three freestream velocities, with the exception of the profiles at the 43 Chapter 4. Facility Characterization Cross−stream location normalized by test section width 0.8 0.6 0.4 0.2 0 −0.2 x = 0D 0.5D 1.0D 1.5D 2.0D 2.5D −0.4 −0.6 −0.8 −0.2 0 0.2 0.4 0.6 0.8 1 1.2 0.8 1 1.2 0.8 1 1.2 Freestream velocity normalized by mean inlet velocity (a) U∞ = 75 m/s. Cross−stream location normalized by test section width 0.8 0.6 0.4 0.2 0 −0.2 x = 0D 0.5D 1.0D 1.5D 2.0D 2.5D −0.4 −0.6 −0.8 −0.2 0 0.2 0.4 0.6 Freestream velocity normalized by mean inlet velocity (b) U∞ = 55 m/s. Cross−stream location normalized by test section width 0.8 0.6 0.4 0.2 0 −0.2 x = 0D 0.5D 1.0D 1.5D 2.0D 2.5D −0.4 −0.6 −0.8 −0.2 0 0.2 0.4 0.6 Freestream velocity normalized by mean inlet velocity (c) U∞ = 35 m/s. Figure 4.3: Freestream velocity profiles at 75, 55, and 35 m/s measured at a range of streamwise locations between x = 0 - 2.5 test section widths downstream of the inlet. 44 Chapter 4. Facility Characterization inlet plane. At U∞ = 35 m/s, there is a 3% undershoot in freestream velocity near the centre of the jet, within 0.3 test section widths of the centreline. At U∞ = 75 m/s, there is a 5% overshoot near the edges of the jet, around 0.4-0.5 test section widths from the centreline. These deviations from the mean freestream velocity are not present in the profiles at the other streamwise locations and hence should not pose a problem for future experiments. A test section flow uniformity of at least 99% (i.e., non-uniformity = 1−uniformity ≤ 1%) was chosen as an acceptable threshold. For each of the three measured test section velocities, the useable test section area was estimated by finding the maximum test section width that satisfied this uniformity threshold at each of the six streamwise measurement locations. The variation of this usable test section width with streamwise location is presented in Fig. 4.4; there is no substantial difference for the three measured velocities. Future experiments will mainly be conducted in between 0.5 - 2.0 test section widths downstream of the inlet and will have a minimum usable cross-sectional area of 500 mm × 500 mm. 580 Usable test section width [mm] 570 560 550 540 530 520 510 500 490 480 0 0.5 1 1.5 2 2.5 x/D, Streamwise location normalized by test section inlet diameter Figure 4.4: Minimum test section width that satisfies the 99% uniformity threshold, as a function of streamwise location, for test section velocities of 75, 55, and 35 m/s. Chapter 4. Facility Characterization 4.3 45 Freestream Turbulence Intensity The freestream turbulence intensity along the centreline of the test section was measured via hot-wire anemometry. A 5 µm diameter tungsten wire was soldered onto a 2 mm wide single-prong hot-wire probe and coated with copper to yield a 0.5 mm sensing length on the wire. The probe was connected to a Dantec 56C17 Constant Temperature Anemometer and calibrated against a pitot-static tube in the test section from 0 m/s to 75 m/s. For the turbulence intensity measurements, the hot-wire probe was positioned in the centre of the test section via an aluminum sting that was fastened to a linear streamwise traverse. The hot-wire voltage signal was passed through a 10.5 kHz low-pass filter and acquired by a NI PCIe-6361 DAQ for 20 s at a sampling rate of 21 kHz. The power spectra of the streamwise velocity fluctuations at various streamwise locations for a freestream velocity of 55 m/s are given in Fig. 4.5. The spectral frequency is expressed in terms of Strouhal number based on the test section width, Dts : St = f Dts . U∞ (4.2) The velocity fluctuations observed within one test section width downstream of the inlet are mainly due to vorticity passed into the test section from the upstream tunnel components (i.e., dome environment, screens and contraction). Moving farther downstream from the test section inlet, the velocity fluctuations increase due to shear layer-induced pressure fields on the free-jet, as indicated by the broad hump in the spectra around St = 0.5 for x/D = 1.0 to 2.5 . The narrow peaks on the humps are generated by resonant pressure fluctuations between the test section inlet and the jet collector. The centreline freestream turbulence intensity value at each of the streamwise measurement locations was calculated via integration of the velocity fluctuation power spectral densities over a frequency range from a certain lower cut-off value up to where the noise floor begins. The lower cut-off frequency, fc,low , was found based on the wavelength, λ, of the largest longitudinal turbulent flow structures that could exist in the test section: U∞ , λ (4.3) λ = 2 × Lts , (4.4) fc,low = where and Lts is the length of the test section. At U∞ = 55 m/s and Lts = 2.2 m, this translated 46 Chapter 4. Facility Characterization Power Spectral Density of Velocity Fluctuations [m 2 s −2Hz −1] 10 10 10 10 10 10 −2 x/D = 0.5 x/D = 0.75 x/D = 1.0 x/D = 1.5 x/D = 2.0 x/D = 2.5 −3 −4 −5 −6 −7 10 −1 10 0 10 1 St Figure 4.5: Power spectral density of streamwise velocity fluctuations along test section centreline at freestream velocity of 55 m/s, for streamwise locations between 0.5 and 2.5 test section widths downstream of inlet. Frequency is given in terms of Strouhal number based on the test section width. All spectra have reached their noise floors by St = 10. to a lower cut-off frequency of 12.5 Hz. All the velocity fluctuation spectra had reached their noise floors by 1 kHz, so this value was chosen as the upper cut-off frequency. Figure 4.6 shows the centreline freestream turbulence intensity values for both the old and new AWT test sections. The comparison shows that the new tunnel flow conditioning screens, as well as the increased contraction ratio, helped improve the turbulence intensity values in the test section. −2 TI 10 Old AWT Improved AWT 10 −3 0 0.5 1 1.5 2 2.5 x/D, streamwise location normalized by test section inlet width Figure 4.6: Comparison of centreline freestream turbulence intensity values for old and new AWT test sections at freestream velocity of 55 m/s over a range of streamwise locations from 0.5 to 2.5 test section widths downstream of the inlet. Chapter 4. Facility Characterization 4.4 47 Background Noise Levels The background noise levels in the anechoic chamber were measured using a 0.5” diameter G.R.A.S. 40AC 200 V externally polarized free-field microphone connected to a G.R.A.S. 26AS preamplifier via a 0.5 in -to- 0.25 in adapter. The microphone and preamplifier were powered by a G.R.A.S. 12AA power module that also provided 20 Hz high-pass filtering and optional 20 dB and 40 dB amplification to the microphone signals. The voltage signal from the 12AA module was acquired by the NI PCIe-6361 DAQ, at a sampling rate of 42 kHz for 60 s for each data point. This measurement system had a SPL dynamic range of -5 dB to 134 dB when using the 40 dB amplification setting. The pressure response of the microphone and preamplifier is designed to be flat over the frequency range from 3.15 Hz to 40 kHz, with an SPL uncertainty of ±2 dB. Using a G.R.A.S. 42AB sound calibrator that emits a 1 kHz tone at 114 dB, the sensitivity of the microphone and preamplifier pairing was determined to be 13.55 mV/Pa. Background noise measurements were taken over the range of tunnel freestream velocities from 0 to 75 m/s. The microphone was placed at two different locations in the acoustic far-field region of the test section jet (see Fig. 4.7). Both positions were at the height of the test section centreline, with the microphone diaphragm pointed towards an imaginary point on the jet centreline 0.6 m downstream of the test section inlet. The diaphragm was oriented 90◦ to the centreline in the first position, and 60◦ to the centreline facing upstream in the second position. Figure 4.7: Plan view of free-field microphone locations for tunnel background noise measurements. Figure not to scale. Chapter 4. Facility Characterization 48 The narrowband SPL spectra measured at these two microphone locations for a range of test section velocities from 15 - 75 m/s are shown in Fig. 4.8. The noise spectra for velocities below 15 m/s were omitted because the microphone measurement system had reached its noise floor. This is not a significant issue as the large majority of future aeroacoustics investigations in this tunnel will be conducted at velocities above 15 m/s. As expected from an open-jet tunnel, the noise spectra are representative of broadband noise. Through careful sealing of the anechoic chamber and wind tunnel components, all noticeable stray whistling noises and cavity resonance tones that would have caused peaks in the spectra have been removed. Compared to the background SPL spectra of the old UTIAS tunnel, the new spectra present reductions in background noise levels in the test chamber across all frequencies of interest. For example, at a freestream velocity of 60 m/s, the reduction is approximately 10 dB at the lowest frequencies between 100-400 Hz, and as much as 15 dB above 500 Hz. This substantial reduction means that microphones in the improved chamber can achieve at least 10 dB signal-to-noise ratio for many airframe noise studies, including landing gear and airfoil trailing edge noise. If an array of microphones and beamforming signal processing algorithms are used, the signal-to-noise ratio may be further increased by at least 5-6 dB, maybe up to 30 dB depending on the algorithm [88–90]. The background noise in the anechoic chamber mainly comes from the fan and motor, open-jet shear layer impingement on the collector, freestream turbulence, and noise transmitted into the open-circuit tunnel from the surrounding up- and downstream environments. The motor contributes high frequency tonal electrical noise (some noticeable peaks in the spectra at freestream velocities of 15-30 m/s) and low frequency mechanical vibrations, the latter of which could be reduced through motor and fan blade balancing, as well as bearing greasing. Shear layer impingement on the collector would produce low frequency broadband noise, and previous researchers (e.g. [59,75,91]) have had some success reducing this noise by expanding the collector or padding it with acoustically absorbent material like foam or fibreglass. Turbulence in the shear layer emits broadband noise. Environmental sources are unpredictable and could contribute both broadband and tonal noise to the spectra, depending on the source. The main method of attenuating this type of noise is through silencer/baffle noise blockage. 49 Chapter 4. Facility Characterization 80 15 m/s 20 m/s 30 m/s 40 m/s 50 m/s 60 m/s 70 m/s 75 m/s Background SPL [dB ref 2x10 Pa] 70 −5 60 50 40 30 20 10 0 −10 2 10 3 10 Frequency [Hz] 10 4 (a) Microphone diaphragm oriented 90◦ to tunnel centreline. 80 15 m/s 20 m/s 30 m/s 40 m/s 50 m/s 60 m/s 70 m/s 75 m/s Background SPL [dB ref 2x10 −5 Pa] 70 60 50 40 30 20 10 0 −10 2 10 3 10 Frequency [Hz] 10 4 (b) Microphone diaphragm oriented 60◦ to tunnel centreline, facing upstream. Figure 4.8: Narrowband background SPL spectra in anechoic chamber, measured with free-field microphone positioned at height of test section centreline, with diaphragm oriented at two different angles towards a point on the centreline that is 0.6 m downstream of inlet. 50 Chapter 4. Facility Characterization As mentioned in Section 4.1, the angle of attack of the tunnel fan blades only had a slight effect on the motor rotational rate and hence the motor noise. The chosen blade angle setting number ‘5’ was the quietest, with the other blade angle settings about 1-2 dB louder (within measurement error of microphone). Figure 4.9 shows the overall sound pressure level in the anechoic chamber measured at the two microphone locations, over a range of test section velocities from 15-75 m/s with the fan blade angle at setting number ‘5’. The curves for the old UTIAS tunnel and several of the quietest tunnels around the world are also provided for comparison. There is a clear improvement in the background noise levels in the UTIAS tunnel, with the microphone measurements at 90◦ to the centreline achieving OASPL reductions between 8-18 dB over the whole range of operating freestream velocities. As well, the noise levels do not change substantially between the two microphone measurement angles, which is important for future far-field directivity measurements. Background A−weighted OASPL [dB] 80 70 Audi (52 deg) DaimlerChrysler (32 deg) Ford DTF (49 deg) Florida State (90 deg) ISVR Southampton (90 deg) Univ. Cottbus (45 deg) Univ. Florida (65 deg) Univ. Siegen (45 deg) Univ. Notre Dame (60 deg) Old UTIAS (90 deg) Improved UTIAS (90deg) Improved UTIAS (60deg) 60 50 40 30 20 0 10 20 30 40 50 60 Test section velocity [m/s] 70 80 90 100 Figure 4.9: Comparison of overall background sound pressure levels between new and old UTIAS AWT and select anechoic wind tunnels around the world, over their range of capable test section velocities. Microphone orientation angle relative to tunnel centreline is shown beside the facility name in the legend. Comparison data adapted from [59, 60, 72–74]. 51 Chapter 4. Facility Characterization The OASPL versus freestream velocity data points can be fitted with a logarithmic curve to give some indication of the types of contributing noise sources in the tunnel. Past empirical studies have shown that fan noise OASPL in dB versus fan rotational speed can be described by the following relation [92, 93]: ∆ OASP L = 10 × log10 R2 R1 n , (4.5) where R2 and R1 are motor rotational speeds, and n is some empirically-determined number. For dipole-type noise sources related to pressure fluctuations on axial fan blades, n is typically between 5 and 6. For quadrupole noise sources related to freestream turbulence, n is approximately 7-8. The motor rotational rate of the UTIAS tunnel is approximately directly proportional to the freestream velocity. The A-weighted OASPL measured by the microphone at the 90◦ orientation is plotted versus motor rotations per minute (RPM) in Fig. 4.10 along with a logarithmic curve fitted through the data points. The curve of best fit has an exponent n = 6.17, indicating that most of the background noise comes from dipole-type sources such as the fan and shear layer interaction with the jet collector, as well as some noise from quadrupole sources such as freestream turbulence. Figure 4.10: A-weighted OASPL measured by far-field microphone in the 90◦ orientation versus fan motor rotational speed. Logarithmic curve of best fit is ∆OASP L = 10 × log10 (R2 /R1 )6.17 . Prior to the addition of front-end baffles, noise disturbances from inside the reverberant geodesic dome entered the test section through the tunnel bellmouth inlet. This noise was from activity in the dome and thermal creaking of the dome walls, and it could not be filtered out of acoustic measurements due to its randomness. On the back-end of the tunnel, there are already baffles built into the diffuser section that help reduce the level 52 Chapter 4. Facility Characterization of fan and motor noise and outdoor disturbances propagating back upstream into the anechoic chamber. After the addition of the front-end baffles, an acoustic characterization was performed to estimate their effectiveness in attenuating sound waves travelling towards the tunnel inlet. This characterization also determined the sound attenuation by the wind tunnel front-end components (i.e., honeycomb, mesh screens, and contractions). The process involved generating white noise using an Altec Lansing ACS22 computer speaker inside the geodesic dome and measuring the sound transmission loss through the left, centre, and right side of the baffles. For each of the three test directions, microphones measured the sound pressure level in front of the baffles, behind the baffles at the centre of the bellmouth inlet, and inside the anechoic chamber at the 90◦ far-field noise measurement location described earlier. The distances between each of these microphone locations and the speaker are summarized in Table 4.1. Table 4.1: Summary of sound transmission distances between the speaker and each of the three microphones at the baffle, the bellmouth inlet, and in the anechoic chamber. Transmission directions are defined facing the baffles when viewing from inside the geodesic dome (same directions as in Fig. 3.18). Transmission Direction Baffle mic [m] Bellmouth mic [m] Chamber mic [m] Leftside of baffles 2.9 7.6 13.0 Centre / head-on 2.9 5.2 10.6 Rightside of baffles 2.9 7.6 13.0 The noise measurements were corrected for spherical transmission spreading loss via the equation: SPLloss = 20 × log10 r2 r1 , (4.6) where r2 and r1 are the distances between the speaker and the two microphone locations under comparison. The atmospheric attenuation of sound at 20 ◦ C and 60% relative humidity across the frequency range of 100 Hz to 10 kHz is between 0.0002 dB/m and 0.1 dB/m, which, given the short transmission distances at hand, was considered negligible. The A-weighted OASPL attenuation by the baffles and wind tunnel front-end components are summarized in Table 4.2. 53 Chapter 4. Facility Characterization Table 4.2: Summary of white noise attenuation by the front-end baffles and tunnel frontend components, in terms of A-weighted OASPL. Transmission directions are defined facing the baffles when viewing from inside the geodesic dome (same directions as in Fig. 3.18). Transmission Direction Baffles [dB] Tunnel Front-end [dB] Total [dB] Leftside of baffles -11 -14 -25 Centre / head-on -16 -14 -30 Rightside of baffles -8 -13 -21 There are considerable gains in front-end noise level attenuation by the addition of the baffles in all three directions. The centre of the baffles offer the strongest attenuation due to the wood sheeting on the front of the baffles and minimal airflow openings. The righthand side of the baffles have the weakest attenuation because of the relatively large opening for air to flow into the bellmouth inlet. While there is no direct line-of-sight into the tunnel inlet from any direction outside of the baffles, it is conjectured that lower frequency sound waves are able to bend around the baffles. Figure 4.11 compares the third-octave A-weighted OASPL spectra measured in front and behind the righthand side of the baffles. It is clear that at frequencies below around 100 Hz, the baffle attenuation is weaker than at frequencies above 100 Hz. 80 A−weighted OASPL [dB] 75 70 65 60 55 50 45 40 Behind baffles at tunnel inlet In front of baffles pointed at speakers 35 30 1 10 2 10 3 10 4 10 5 10 Third−octave frequency bands [Hz] Figure 4.11: Third-octave A-weighted SPL of white noise generated by the speaker in front of the baffles, as measured by microphones in front of and behind the baffles on the right hand side of the wind tunnel inlet. 54 Chapter 4. Facility Characterization 4.5 Anechoic Cut-off Frequency As discussed earlier, the theoretical cut-off frequency based on the depth of the wedges is approximately 150 Hz, but construction details and wind tunnel components protruding into the chamber can increase this value. In order to determine the actual cut-off frequency, a series of microphone traverses were performed using the tonal noise discrete sampling method found in ISO-3745:2012 for ten tonal frequencies from 1 kHz down to 125 Hz. A speaker was set up on the tunnel centreline, 0.6 m downstream of the test section inlet, and was driven by an Agilent 33210A function generator. Three free-field microphone traverses were performed on paths extending from the speaker to the chamber walls at 45◦ , 90◦ , and 135◦ relative to the downstream tunnel centreline (see schematic of set-up in Fig. 4.12), with the microphone diaphragm facing the speaker at all times. The microphone hung under a traverse line made of 0.3 mm diameter fishing line and was pulled by a take-up line attached to a spool on a rotary stage driven by a stepper motor. The whole microphone traversing system had a linear resolution of 0.5 mm per full step of the motor. In order to resolve the spatial characteristics of the tonal noise, the separation distance between measurement locations along a traverse must not exceed 15% of the acoustic wavelength of interest . This translated to 0.05 m for the shortest wavelength that was tested (i.e., at frequency of 1 kHz). Traverses for all the tonal frequencies were performed from 0.3 m to 2.2 m away from the speaker diaphragm, at 0.05 m increments. A stationary reference microphone was placed on the tunnel centreline 0.3 m from the speaker to monitor the stability of the sound source during the traverses. Based on the assumption that the anechoic chamber approximates an acoustic free-field, the SPL readings at distances r1 and r2 from the sound source should adhere to the following relation: SPL(r1 ) = SPL(r2 ) − 20 × log10 r1 r2 . (4.7) This relation is valid for measurement locations that are in the acoustic far-field region of the sound source, which is given by : 95 m/Hz , r ≥ MAX 0.5 m, f (4.8) where f is the noise frequency of interest in Hz. The acceptable bounds for any deviation of the traverse SPL readings from the free-field assumption are defined in ISO-3745:2012 55 Chapter 4. Facility Characterization Figure 4.12: Schematic of speaker and reference microphone locations and three microphone traverse lines for the free-field characterization test. All traverse lines are in the mid-plane of the anechoic chamber coinciding with the tunnel centreline. and summarized in Table 4.3 for the frequencies of interest. The SPL measurements for the three traverse directions, expressed as deviation from the free-field assumption, are given in Figs. 4.13-4.15 for the measurement locations that are in the far-field of the speaker. Table 4.3: Acceptable SPL error bounds for free-field characterization of anechoic chambers. Values defined in ISO-3745:2012. Frequency range SPL error bound ≤ 630 Hz ± 1.5 dB 800 Hz - 5 kHz ± 1.0 dB 56 Chapter 4. Facility Characterization SPL deviation from acoustic freefield assumption [dB] 2 0 -2 2 0 -2 5 0 -5 2 0 -2 2 0 -2 2 0 -2 2 0 -2 2 0 -2 2 0 -2 1 0 -1 f = 125 0 0.5 1 f = 150 1.5 2 2.5 0 0.5 1 f = 160 1.5 2 2.5 0 0.5 1 f = 170 1.5 2 2.5 0 0.5 1 f = 180 1.5 2 2.5 0 0.5 1 f = 190 1.5 2 2.5 0 0.5 1 f = 200 1.5 2 2.5 0 0.5 1 f = 250 1.5 2 2.5 0 0.5 1 f = 500 1.5 2 2.5 0 0.5 1 f = 1000 1.5 2 2.5 0 0.5 1 2 2.5 1.5 Distance from speaker [m] Figure 4.13: Deviation of SPL measurements from free-field assumption along the traverse path at 45◦ relative to the downstream centreline (towards the door), for various speaker frequencies from 125 Hz to 1 kHz. Acceptable SPL error bounds from ISO-3745:2012. 57 Chapter 4. Facility Characterization SPL deviation from acoustic freefield assumption [dB] 2 0 -2 2 0 -2 2 0 -2 2 0 -2 2 0 -2 2 0 -2 2 0 -2 2 0 -2 2 0 -2 1 0 -1 f = 125 0 0.5 1 f = 150 1.5 2 2.5 0 0.5 1 f = 160 1.5 2 2.5 0 0.5 1 f = 170 1.5 2 2.5 0 0.5 1 f = 180 1.5 2 2.5 0 0.5 1 f = 190 1.5 2 2.5 0 0.5 1 f = 200 1.5 2 2.5 0 0.5 1 f = 250 1.5 2 2.5 0 0.5 1 f = 500 1.5 2 2.5 0 0.5 1 f = 1000 1.5 2 2.5 0 0.5 1 2 2.5 1.5 Distance from speaker [m] Figure 4.14: Deviation of SPL measurements from free-field assumption along the traverse path at 90◦ relative to the downstream centreline, for various speaker frequencies from 125 Hz to 1 kHz. Acceptable SPL error bounds from ISO-3745:2012. 58 Chapter 4. Facility Characterization SPL deviation from acoustic freefield assumption [dB] 2 0 -2 2 0 -2 2 0 -2 2 0 -2 2 0 -2 2 0 -2 2 0 -2 2 0 -2 5 0 -5 2 0 -2 f = 125 0 0.5 1 f = 150 1.5 2 2.5 0 0.5 1 f = 160 1.5 2 2.5 0 0.5 1 f = 170 1.5 2 2.5 0 0.5 1 f = 180 1.5 2 2.5 0 0.5 1 f = 190 1.5 2 2.5 0 0.5 1 f = 200 1.5 2 2.5 0 0.5 1 f = 250 1.5 2 2.5 0 0.5 1 f = 500 1.5 2 2.5 0 0.5 1 f = 1000 1.5 2 2.5 0 0.5 1 2 2.5 1.5 Distance from speaker [m] Figure 4.15: Deviation of SPL measurements from free-field assumption along the traverse path at 135◦ relative to the downstream centreline, for various speaker frequencies from 125 Hz to 1 kHz. Acceptable SPL error bounds from ISO-3745:2012. Chapter 4. Facility Characterization 59 For the 45◦ traverse, the anechoic chamber and false door provide a good acoustic freefield environment at frequencies down to 170 Hz and within 2.2 m of the speaker. At 160 Hz and 150 Hz, and measurement locations greater than 1.8 m from the speaker, the SPL values exceed the acceptable error bounds and therefore the chamber cannot be considered anechoic in those conditions. For the 90◦ traverse, the SPL values for all measured frequencies and at measurement points up to 2 m from the speaker are within the allowable error bounds. Beyond 2 m from the speaker, the wedges on the closest wall begin to have an adverse effect on the free-field quality, pushing the SPL deviations at frequencies of 190 Hz and 200 Hz slightly above acceptable limits. Finally, for the 135◦ traverse, the chamber largely adheres to the free-field assumption at frequencies down to 170 Hz and within 2.2 m of the speaker, except for small (≤0.5 dB) deviations beyond allowable limits at frequencies of 500 Hz and 1 kHz. This is most likely due to acoustic reflections from the protruding tunnel contraction and flow sections, which were only covered in one layer of egg-crate foam sheets. A potential improvement to the acoustic treatment on these tunnel sections could be miniature wedges with a depth of 0.15-0.2 m, but this is non-critical and the additional wedges would be expensive. In summary, the anechoic cut-off frequency of the chamber is around 170 Hz, and the free-field assumption is valid for a minimum distance of 2 m from a point on the tunnel centreline that is 0.6 m downstream of the test section inlet, thus meeting the relevant thesis objectives. 4.6 Reverberation Time The integrated impulse response method found in ISO-3382-1:2009 was used to determine the reverberation time of the anechoic chamber. The standard normally requires an impulse noise source that has a peak SPL of at least 75 dB above the noise floor in order to directly measure RT60 . However, this was not possible with the readily available lab equipment, so instead an approximation method within the standard was used. RT20 , the time for an impulse response to decrease by 20 dB, was calculated from an impulse noise with a peak SPL at least 35 dB above the noise floor. RT20 was then linearly extrapolated to get RT60 . A microphone was set-up at the 90◦ far-field measurement location as previously described, and a loud clap in the middle of the anechoic chamber served as the noise source. The pressure response signal from this impulse noise is shown in Fig. 4.16. 60 Chapter 4. Facility Characterization 8 6 Pressure signal [Pa] 4 2 0 −2 −4 −6 −8 0 200 400 600 800 1000 1200 1400 1600 1800 2000 Time [s] Figure 4.16: Pressure response signal of a single clap impulse noise in the centre of the anechoic chamber, as measured by a microphone at the 90◦ far-field measurement location. Using cumulative trapezoidal integration on the impulse response pressure signal, noise decay curves were estimated for octave bands with centre frequencies between 250 Hz and 8 kHz . The reverberation time for each octave band was then approximated using a least-squares fit on the corresponding decay curve between the points -5 dB and -25 dB below the peak of the impulse noise to get RT20 , and then linearly extrapolated to get RT60 . Figure 4.17 compares the reverberation times for the improved anechoic chamber and the bare concrete chamber. The presence of acoustic wedges greatly reduces the reverberation time from an average of 1.5 s for the bare concrete chamber to 0.032 s for the acoustically-treated chamber. In general, the reverberation time of the anechoic chamber increases as frequency decreases towards the anechoic cut-off frequency. Interestingly, there is also a slight increase in the RT60 of the treated anechoic chamber around 4 kHz, which corresponds to the dip in the acoustic absorption coefficient of the foam used for the wedges (see Fig. 3.10). As mentioned in Section 2, the literature does not usually discuss reverberation time for acoustic wind tunnels so comparisons cannot be made. However, the values here all satisfy the thesis objective of RT60 ≤ 0.2 s and meet the expectations of future aeroacoustic experiments in this facility. 61 Chapter 4. Facility Characterization 10 New anechoic chamber Bare concrete chamber 0 RT 60 [s] 101 10-1 10-2 2 10 103 104 Frequency [Hz] Figure 4.17: Comparison between the RT60 of the improved anechoic chamber and of the bare concrete chamber, for octaves with centre frequencies between 250 Hz and 8 kHz. Chapter 5 Conclusions 5.1 Summary This thesis discussed the motivation, design, construction, and characterization of the improvements to the UTIAS Anechoic Wind Tunnel facility. Reduction of aircraft noise forms a key part of the sustainable aviation agenda currently being pushed by international governing bodies. Facilities like the UTIAS AWT are rare and play a critical role in the aeroacoustic research efforts led by both industry and academia. The work performed in this thesis helped to update the UTIAS AWT for current and foreseeable future research needs. At the beginning of this thesis, a set of aerodynamic and acoustic performance objectives were laid out. A full characterization of the facility has shown that these objectives were either met or exceeded, with the exception of the background noise levels at frequencies below 500 Hz. Table 5.1 summarizes the results. Table 5.1: Summary of tunnel performance objectives and results from the facility characterization. Objective Target Result Test section maximum freestream velocity ≥70 m/s 75 m/s Test section dimensions Cross-sectional diameter ≥0.5 m, streamwise length ≥1.2 m 0.5 m × 0.5 m, 1.2 m 62 63 Chapter 5. Conclusions Test section freestream turbulence intensity ≤ 0.5% for x/D ≤ 1.5 ≤ 0.4% Test section spatial flow uniformity ≥ 99% across test section ≥ 99% Background noise level -15 dB SPL, 200 Hz - 20 kHz -10 dB (200-500 Hz), -15 dB (500-20,000 Hz) Anechoic limit cut-off frequency ≤ 200 Hz 170 Hz Usable anechoic area around test section 2 m arc about a point on centreline & 0.6 m downstream of inlet Satisfied, 45◦ -135◦ relative to centreline Reverberation time Below 0.2 s 0.032 s 5.2 Recommendations for Future Work There are several immediate recommendations for future aeroacoustics tests in the UTIAS AWT. First, if the interest is in the low frequency noise range (200-500 Hz) and the current background levels are too high, then a few things could be investigated. The signal-to-noise ratio of far-field microphone measurements at these frequencies may be improved by using windscreens to reduce microphone self-noise generated by the out-of-jet recirculating flows. The jet-collector may require improvement to reduce the noise generated by shear layer impingement, possibly by adding more/better padding material or by increasing the flow capture area of the collector. If the latter is chosen, care must be taken not to increase the area too much and cause the static pressure in the test section to exceed tolerances. Another source of low frequency noise that may require attention is the vibrational noise from the motor and fan. If balancing and greasing does not sufficiently reduce vibrations, a thin circumferential section may need to be cut out of the diffuser and replaced with a rubber membrane to isolate the vibrations. Chapter 5. Conclusions 64 Second, if the test model is long (i.e., extends past 2 test section widths downstream of the inlet), and far-field noise measurements are required at azimuthal angles less than 60◦ to the centreline pointing upstream, there may not be enough length in the chamber to set up these microphones (the false door and/or collector wall would get in the way). In this case, the wind tunnel straight flow section that currently sits between the primary and secondary contractions may be omitted to shift the test section (and hence model and far-field microphone arc) upstream by 0.6 m. A characterization would have to be performed to determine how this extra 0.6 m length of open-jet affects the freestream flow profiles. Finally, for all future aeroacoustics experiments in the AWT, there should be a way of measuring the local atmospheric pressure and humidity inside the chamber to give an accurate calculation of certain aerodynamic and acoustic parameters (e.g., air density, sound absorptivity in air, etc.). Appendix A Anechoic Foam Material Data Sheet 65 Appendix A. Anechoic Foam Material Data Sheet 66 Figure A.1: Material data sheet for polyurethane foam used for the anechoic wedges in the improved UTIAS AWT. Bibliography  Mahashabde, A., Wolfe, P., Ashok, A., Dorbian, C., He, Q., Fan, A., Lukachko, S., Mozdzanowska, A., Wollersheim, C., Barrett, S., Locke, M., and Waitz, I.A., “Assessing the Environmental Impacts of Aircraft Noise and Emissions,” Progress in Aerospace Sciences, vol. 47, pp. 15-52, 2011.  He, Q., “Development of an Income-Based Hedonic Monetization Model for the Assessment of Aviation-Related Noise Impacts,” Master of Science Thesis, Department of Aeronautics and Astronautics, Massachusetts Institute of Technology, Cambridge, MA, USA, 2010.  Kish, C., “An Estimate of the Global Impact of Commercial Aviation Noise,” Master of Science Thesis, Department of Aeronautics and Astronautics, Massachusetts Institute of Technology, Cambridge, MA, USA, 2006.  Dickson, N., “Aircraft Noise Technology and International Noise Standards,” Environment Branch, ICAO Air Transport Bureau, 2014.  “FAA Aerospace Forecast: Fiscal Years 2013-2033,” Federal Aviation Administration, 2013.  “2013 Global Aerospace and Defense Industry Outlook,” Deloitte Touche Tohmatsu Limited, 2012.  Graham, W.R., Hall, C.A., and Vera Morales, M., “The Potential of Future Aircraft Technology for Noise and Pollutant Emissions Reduction,” Journal of Transport Policy, vol. 34, pp. 36-51, 2014.  Dobrzynski, W., “Almost 40 Years of Airframe Noise Research: What Did We Achieve?,” Journal of Aircraft, vol. 47, no. 2, pp, 353-367, 2010.  Lighthill, M.J., “On Sound Generated Aerodynamically. I. General Theory,” Proceedings of the Royal Society of London, Series A: Mathematical and Physical Sciences, vol. 211, no. 1107, pp. 564-587, 1952. 67 Bibliography 68  Curle, N., “The Influence of Solid Boundaries Upon Aerodynamic Sound,” Proceedings of the Royal Society of London, Series A: Mathematical and Physical Sciences, vol. 231, no. 1187, pp. 505-514, 1955.  Ffowcs Williams, J.E., and Hall, L.H., “Aerodynamic Sound Generation by Turbulent Flow in the Vicinity of a Scattering Half-Plane,” Journal of Fluid Mechanics, vol. 40, no. 4, pp. 657-670, 1970.  Gibson, J.S., “The Ultimate Noise Barrier: Far Field Radiated Aerodynamic Noise,” Proceedings of Inter-Noise 72, Washington, D.C., pp. 332-337, 1972.  Healy, G.J., “Measurement and Analysis of Aircraft Far Field Aerodynamic Noise,” NASA CR-2377, 1974.  Gibson, J.S., “Non-Engine Aerodynamic Noise Investigation of a Large Aircraft,” NASA CR-2378, 1974.  Revell, J.D., “The Calculation of Aerodynamic Noise Generated by Large Aircraft at Landing Approach,” Journal of the Acoustical Society of America, vol. 55, pp. 73, 1974.  Fethney, P., “An Experimental Study of Airframe Self-Noise,” Royal Aircraft Establishment Report AERO1623, 1975.  Dobrzynski, W., and Buchholz, H., “Full-Scale Noise Testing on Airbus Landing Gears in the German-Dutch Wind Tunnel,” AIAA/CEAS Paper 1997-1597, 1997.  Stoker, R.W., “Landing Gear Noise Test Report,” NASA report NAS1-97040, 1997.  Ravetta, P.A., Burdisso, R.A., and Ng, W.F., “Wind Tunnel Aeroacoustic Measurements of a 26% Scale 777 Main Landing Gear Model,” AIAA/CEAS Paper 2004-2885, 2004.  Smith, M.G., Fenech, B., Chow, L.C., Molin, N., Dobrzynski, W., and Seror, C., “Control of Noise Sources on Aircraft Landing Gear Bogies,” AIAA/CEAS Paper 2006-2626, 2006.  Choudhari, M.N., and Khorrami, M., “Slat Cove Unsteadiness: Effect of 3D Flow Structures,” AIAA Paper 2006-0211, 2006.  Khorrami, M., Berkman, M.E., and Choudhari, M.N., “Unsteady Flow Computations of a Slat with a Blunt Trailing Edge,” AIAA Journal, vol. 38, no. 11, pp. 2050-2058, 2000. Bibliography 69  Pott-Pollenske, M., Dobrzynski, W., Buchholz, H., Gehlhar, B., and Walle, F., “Validation of a Semi-empirical Airframe Noise Prediction Method through Dedicated A319 Flyover Noise Measurements,” AIAA Paper 2002-2470, 2002.  Oerlemans, S., and Pott-Pollenske, M., “An Experimental Study of Gear Wake/Flap Interaction Noise,” AIAA/CEAS Paper 2004-2886, 2004.  Dobrzynski, W., Chow, L.C., Guion, P., and Shiells, D., “A European Study on Landing Gear Airframe Noise Sources,” AIAA/CEAS Paper 2000-1971, 2000.  Lockett, B., “Boeing 787-83Q N787ZA on approach to land on runway 30 Right”, 2011.  Fink, M.R., “Airframe Noise Prediction Method,” Federal Aviation Administration RD-77-29, 1977.  Smith, M.G., and Chow, L.C., “Validation of a Prediction Model for Aerodynamic Noise from Aircraft Landing Gears,” AIAA/CEAS Paper 2002-2581, 2002.  Guo, Y.P., Yamamoto, K.J., and Stocker R.W., “An Empirical Model for Landing Gear Noise Prediction,” AIAA/CEAS Paper 2004-2888, 2004.  Guo, Y.P., “A Semi-Empirical Model for Aircraft Landing Gear Noise Prediction,” AIAA/CEAS Paper 2006-2627, 2006.  Burley, C.L., Brooks, T.F., Humphreys W.M., Jr., and Rawls, J.W., Jr., “ANOPP Landing Gear Noise Prediction Comparisons to Model-Scale Data,” AIAA/CEAS Paper 2007-3459, 2007.  Khorrami, M.R., and Mineck, R.E., “Towards Full Aircraft Airframe Noise Prediction: Detached Eddy Simulations,” AIAA/CEAS Paper 2014-2280, 2014.  Liu, W., Kim, J.W., Zhang, X., Angland, D., and Caruelle, B., “Landing-gear noise prediction using high-order finite difference schemes,” Journal of Sound and Vibration, vol. 332, pp. 3517-3534, 2013.  Yang, G, Liu, J., and Sun, J., “Computational Aeroacoustic Simulation of Landing Gear,” Applied Mechanics and Materials Journal, vol. 421, pp. 110-115, 2013.  Ewert, R., “Broadband Slat Noise Prediction Based on CAA and Stochastic Sound Sources from a Fast Random Particle-Mesh (RPM) Method,” Computer and Fluids Journal, no. 37, pp. 369-387, 2008. Bibliography 70  Dahl, M., “A Process for Assessing NASA’s Capability in Aircraft Noise Prediction Technology,” AIAA/CEAS Paper 2008-2813, 2008.  Dobrzynski, W., Schoning, B., Chow, L.C., Wood, C., Smith, M., and Seror, C., “Design and Testing of Low Noise Landing Gears,” International Journal of Aeroacoustics, vol. 5, no. 3, pp. 233-262, 2006.  Remillieux, M.C., Camargo, H.E., Ravetta, P.A., Burdisso, R.A., and Ng, W.F., “Noise Reduction of a Model-Scale Landing Gear Measured in the Virginia Tech Aeroacoustic Wind Tunnel,” AIAA paper 2008-2818, 2008.  Pott-Polenske, M., “Splitter Plate and Deceleration Plate Noise Test,” Report OPENAIR-DLR-D4.1.6-D4.1.7-R1.0, 2010.  Spiteri, M., Zhang, X., Molin, N., and Chow, L.C., “The Use of a Fairing and Split Plate for Bluff Body Noise Control,” AIAA Paper 2008-2817, 2008.  Koop L, and Dillmann A., “Reduction of Flap Side Edge Noise: Passive and Active Flow Control,” AIAA Paper 2004-2803, 2004.  Dobrzynski,W., Bieler, H., Chow, L. C., and Remy, H., The Airframe Noise Reduction Challenge: Lessons Learned from the European SILENCER Project, DLR Congress, Braunschweig, Germany, Nov. 2006.  Herr, M., A Noise Reduction Study on Flow-Permeable Trailing Edges, Proceedings of the 8th ONERA-DLR Aerospace Symposium, Gttingen, Germany, Oct. 2007.  Ma, Z., and Zhang, X., “Broadband Slat Noise Attenuation with Acoustic Liner Treatment,” AIAA Paper 2008-2964, 2008.  Chen, P., Zhang, X., and Angland, D., “Slat Noise Reduction Using a Leading Edge Strip,” AIAA Paper 2011-2908, 2011.  Khorrami, M.R., and Lockard, D.P., “Effects of Geometrical Details on Slat Noise Generation and Propagation,” AIAA Paper 2006-2664, 2006.  Kozlov, A.V., and Thomas, F.O., “Active Noise Control of Bluff-body Flows using Dielectric Barrier Discharge Plasma Actuators,” AIAA Paper 2009-3245, 2009.  Li, Y., Zhang, X., and Huang, X., “The Use of Plasma Actuators for Bluff Body Broadband Noise Control,” Experiments in Fluids Journal, vol. 49, no. 2, pp. 367377, 2010. Bibliography 71  Huang, X., Zhang, X., and Li, Y., “Broadband Flow-induced Sound Control using Plasma Actuators,” Journal of Sound and Vibration, vol. 329, no. 13, pp. 24772489, 2010.  Oerlemans, S., and Bruin, A., “Reduction of Landing Gear Noise using an Air Curtain,” AIAA Paper 2009-3156, 2009.  Angland, D., Zhang, X., and Goodyer, M., “The Use of Blowing Flow Control to Reduce Bluff Body Interaction Noise,” AIAA Paper 2010-3786, 2010.  Koop, L., Ehrenfried, K., Dillmann, A., and Michel, U., “Reduction of Flap Side Edge Noise by Active Flow Control,” AIAA Paper 2002-2469, 2002.  Knacke, T., and Thiele, F., “Slat Noise Reduction using Steady Suction,” AIAA Paper 2008-17, 2008.  Huang, H., Li, W., and Wang, F., “Slat Noise Suppression using Upstream Mass Injection, Journal of Shanghai Jiaotong University (Science), vol. 18, no. 5, pp. 620629, 2013.  Chan, S., Zhang, X., and Gabriel, S., “The Attenuation of Cavity Tones using Plasma Actuators,” AIAA Journal, vol. 45, no. 7, pp. 15251538, 2007.  Huang, X., Chan, S., Zhang, X., and Gabriel, S., “Variable Structure Model for Flow-induced Tonal Noise Control with Plasma Actuators,” AIAA Journal, vol. 46, no. 1, pp. 241250, 2008.  Ball, N.G., “Acoustic Excitation of Tollmien-Schlichting Waves in a Laminar Boundary Layer,” Doctor of Philosophy Thesis, Institute for Aerospace Studies, University of Toronto, Toronto, ON, Canada, 1987.  Chekiri, R., “Experimental Aeroacoustic Study of a Landing Gear in the Unsteady Flow Induced by a Propeller,” Master of Applied Science Thesis, Institute for Aerospace Studies, University of Toronto, Toronto, ON, Canada, 2014.  Pascioni, K., Reger, R., Edstrand, A., and Cattafesta, L., “Characterization of an Aeroacoustic Wind Tunnel Facility,” Proceedings of Inter-Noise 2014, Melbourne, Australia, 2014.  Chong, T.P., Joseph, P.F., Davies, P.O.A.L., “Design and Performance of an Open Jet Wind Tunnel for Aero-Acoustic Measurement,” Journal of Applied Acoustics, vol. 70, pp. 605-614, 2009. Bibliography 72  Daimler AG, “New aeroacoustics wind tunnel in Sindelfingen,” 2013.  Hermans, C., “Acoustic testing at wind tunnel facilities of German-Dutch Wind Tunnels (DNW),” German-Dutch Wind Tunnels, 2014.  Barlow, J.B., Rae, Jr., and W.H., Pope, A., “Introduction, Aeroacoustics,” in LowSpeed Wind Tunnel Testing, 3rd ed. New York: Wiley, pp. 16-17, 1999.  Soderman, P.T., and Allen, C.S., “Microphone Measurements In and Out of Airstream,” in Aeroacoustic Measurements, Mueller, T.J., Ed. Berlin: SpringerVerlag, pp. 1-61, 2002.  Underbrink, J., “Aeroacoustic Phased Array Testing in Low Speed Wind Tunnels,” in Aeroacoustic Measurements, Mueller, T.J., Ed. Berlin: Springer-Verlag, pp. 98-217, 2002.  International Electrotechnical Commission, “Electroacoustics - Sound Level Meters - Part 1: Specifications,” IEC 61672-1:2013.  Duell, E., Walter, J., Arnette, S., and Yen, J., “Recent Advances in Large-Scale Aeroacoustic Wind Tunnels,” AIAA/CEAS Paper 2002-2503, 2002.  Santana, L.D., Carmo, M., and Catalano, F.M., “The Update of an Aerodynamic Wind-Tunnel for Aeroacoustics Testing,” Journal of Aerospace Technology Management, vol. 6, no. 2, pp. 111-118, 2014.  Mathew, J., Bahr, C., Carroll, B., Sheplak, M., and Cattafesta, L., “Design, Fabrication, and Characterization of an Anechoic Wind Tunnel Facility,” AIAA/CEAS Paper 2005-3052, 2005.  Barlow, J.B., Rae, Jr., and W.H., Pope, A., “Wind Tunnels,” in Low-Speed Wind Tunnel Testing, 3rd ed. New York: Wiley, pp. 16-17, 1999.  Walter, J., Duell, E., Martindale, B., Arnette, S., and Geierman, R., “The DaimlerChrysler Full-Scale Aeroacoustic Wind Tunnel,” SAE Paper 2003-01-0426, 2003.  Mueller, T.J., Scharpf, D.F., Batill, S.M., Strebinger, R.B., Sullivan, C.J., and Subramanian, S., “The Design of a Subsonic Low-Noise Low-Turbulence Wind Tunnel for Acoustic Measurements,” AIAA Paper 92-3883, 1992.  Winkler, J., Temel, F.Z., and Carolus, T., “Concept, Design and Characterization of a Small Aeroacoustic Wind Tunnel Facility with Application to Fan Blade Measurements,” Proceedings of Fan Noise 2007, Lyon, France, 2007. Bibliography 73  Sarradj, E., Fritzsche, C., Geyer, T., Giesler, J., “Acoustic and aerodynamic design and characterization of a small-scale aeroacoustic wind tunnel,” Journal of Applied Acoustics, vol. 70, pp. 1073-1080, 2009.  Mathew, J., “Design, Fabrication, and Characterization of an Anechoic Wind Tunnel Facility,” Doctor of Philosophy Thesis, Department of Mechanical and Aerospace Engineering, University of Florida, Florida, Gainesville, FL, USA, 2006.  Remillieux, M.C., Crede, E.D., Camargo, H.E., Burdisso, R.A., Devenport, W.J., Rasnick, M., Van Seeters, P., and Chou, A., “Calibration and Demonstration of the New Virginia Tech Anechoic Wind Tunnel,” AIAA/CEAS Paper 2008-2911, 2008.  Brooks, T.F., Pope, D.S., and Marcolini, M.A., “Airfoil Self-Noise and Prediction,” Technical Report NASA Reference Publication 1218, 1989.  Mehta, R.D., and Bradshaw, P., “Design Rules for Small Low Speed Wind Tunnels,” Aeronautical Journal, 1979.  Pope, A., and Harper, J.J., “Low Speed Wind Tunnel Testing,” 1st ed. New York: Wiley, 1966.  Laws, E.M. and Livesey, J.L., “Flow Through Screens,” Annual Review of Fluid Mechanics, vol. 10, pp. 247-266, 1978.  Wieghardt, K.E.G., “On the Resistance of Screens,” Aeronautical Quarterly, vol. 4, pp. 186-192, 1953.  Lindgren, B., and Johansson, A.V., “Design and Evaluation of a Low-Speed WindTunnel with Expanding Corners,” Technical report TRITA-MEK 2002:14, Department of Mechanics, KTH, 2002.  Su, Y.X., “Flow Analysis and Design of Three-Dimensional Wind Tunnel Contractions,” AIAA Journal, vol. 29, no. 11, pp. 1912-1920, 1991.  Harris, C.M., and Molloy, C.T., “The Theory of Sound Absorptive Materials,” Journal of the Acoustical Society of America, vol. 24, no. 1, pp. 1-7, 1952.  Ayub, M., Xander, A.C., Howard, C.Q., and Cazzolato, B.S., “A Review of Acoustic Absorption Mechanisms of Nanoscopic Fibres,” Proceedings of Acoustics 2011, Paper 77, Gold Coast, Australia, 2011.  Victor, D., and Moyer, D.L., “Foam Factory 1.5” Acoustic Eggcrate Foam,” Test Report RAL-A11-148, Riverbank Acoustical Laboratories, Illinois, U.S.A., 2011. Bibliography 74  Michel, U., Froebel, E., “Lower Limit for the Velocity Fluctuation Level in Wind Tunnels,” Experiments in Fluids, vol. 6, pp. 49-54, 1988.  Shin, H.-C., Graham, W.R., Sijtsma, P., Andreou, C., and Faszer, A.C., “Implementation of a Phased Microphone Array in a Closed-Section Wind Tunnel,” AIAA Journal, vol. 45, no. 12, pp. 2897-2909, 2007.  Dougherty, R.P., “Functional Beamforming,” Proceedings of 5th Berlin Beamforming Conference, Berlin, Germany, 2014.  Dougherty, R.P., “Beamforming in Acoustic Testing,” in Aeroacoustic Measurements, Mueller, T.J., Ed. Berlin: Springer-Verlag, pp. 62-97, 2002.  Kudo, T., Komatsu, Y., Maeda, K., and Nishimura, M., “Techniques for Reducing Low-Frequency Fluctuations in Aeroacoustic Wind Tunnels,” Journal of Environment and Engineering, vol. 4, no. 2, pp. 289-301, 2009.  Soderman, P.T., “Sources and Levels of Background Noise in the NASA Ames 40-by 80-Foot Wind Tunnel - A Status Report,” NASA TM 100077, 1988.  Jaeger, S.M., Allen, C.S., and Soderman, P.T., “Reduction of Background Noise in the NASA Ames 40- by 80- Foot Wind Tunnel,” Proceedings of CEAS/AIAA Aeroacoustics Conference, Munich, Germany, 1995.  Cunefare, K.A., Biesel, V.B., Tran, J., Rye, R., Graf, A., Holdhusen, M., and Albanese, A.-M., “Anechoic Chamber Qualification: Traverse Method, Inverse Square Law Analysis Method, and Nature of Test Signal,” Journal of Acoustical Society of America, vol. 113, no. 2, 2003.