Kevin Dillion

Framework for wind noise studies

Morgan and Raspet [J. Acoust. Soc. Am. 92, 1180–1183 (1992)] performed simultaneous wind velocity and wind noise measurements and determined that the wind noise spectrum is highly correlated with the wind velocity spectrum. In this paper, two methods are developed for predicting the upper limits of wind noise pressure spectra from fluctuating velocity spectra in the inertial range. Lower limits on wind noise are estimated from two theories of the pressure fluctuations that occur in turbulence when no wind screen or microphone is present. Empirical results for the self-noise of spherical and cylindrical windscreens in substantially nonturbulent flows are also presented. Measurements of the wind velocity spectra and wind noise spectra from a variety of windscreens are described and compared to the theoretical predictions. The wind noise data taken at the height of the anemometer lies between the upper and lower limits and the predicted self-noise is negligible. The theoretical framework allows windscreen re...

Frequency-wavenumber processing for infrasound distributed arrays

The work described herein discusses the application of a frequency-wavenumber signal processing technique to signals from rectangular infrasound arrays for detection and estimation of the direction of travel of infrasound. Arrays of 100 sensors were arranged in square configurations with sensor spacing of 2 m. Wind noise data were collected at one site. Synthetic infrasound signals were superposed on top of the wind noise to determine the accuracy and sensitivity of the technique with respect to signal-to-noise ratio. The technique was then applied to an impulsive event recorded at a different site. Preliminary results demonstrated the feasibility of this approach.

Acoustical properties of porous hose wind noise filters

Observation of infrasound signals in the atmosphere is often masked by wind noise. A common means of filtering out the wind noise is to connect a commercial porous (or soaker) hose to the manifold of a micro barometer. The filtering effect is attributed to the ability of the porous hose to average the pressure variations over its length. The pressure variations due to the turbulent wind field are incoherent on a scale equal to the hose length and, therefore, are reduced in the averaging process. Infrasound signals, with wavelengths much longer than the hose, are reduced little in the averaging process. There remains the question, ‘‘How does the porous hose respond to infrasound signals that have wavelengths comparable to the hose length?’’ To answer this question, measurements have been made of infrasound signals radiated from a jet engine during takeoff using an infrasound sensor with a porous hose connected to its port, and an array of infrasound sensors distributed along the length of the porous hose. ...

Advances in distributed arrays for detection of infrasonic events

Infrasound arrays normally consist of four to eight microbarometers spaced kilometers apart. Each of these microbarometers is connected to a pipe or porous hose array to reduce wind noise. This presentation describes a 100 sensor all weather distributed array and its use in continuously logging infrasound signals over an extended period of time. The array has been deployed next to a conventional porous hose array connected to a Chaparral 2.5 microbarometer. The effectiveness of the two arrays in canceling wind noise and detecting infrasound signals is compared. The hardware and software used by the distributed array to log the data and to process it so as to cancel wind noise, identify and separate the infrasound signals, and locate their sources will be discussed.

An investigation of the correlation of infrasound signals as a function of sensor separation and wind velocity

Wind noise is a common obstacle to the detection and analysis of infrasonic signals. The masking caused by the wind noise is often reduced by spatial averaging via pipe or porous hose arrays. An alternate method is to average the signals from multiple sensors in distributed arrays. In order to reduce the wind noise, the sensors in the distributed array must be far enough apart so that their wind noise signals are incoherent, but not so far apart that the sound loses its coherence. Few measurements have been reported of the infrasound correlation distance in the presence of wind. This paper reports results of measurements of this infrasound correlation when the infrasound sensors are separated by distances up to hundreds of meters. Techniques for combining infrasound signals from widely separated sensors and measuring their correlation in the presence of wind noise are discussed.

Application of a blind source separation algorithm for the detection and tracking of tornado-generated infrasound emissions during the severe weather outbreak of 27 April 2011

April 25-28, 2011 has been identified by many as the most significant and severe single-system outbreak of tornadoes in recorded history. One day in particular, the 27th of April, has been classified by the National Oceanic and Atmospheric Administration (NOAA) as the fourth deadliest tornado outbreak in US history. Severe tornadic activity on this day levied catastrophic damage to life and property across areas of Mississippi, Alabama, Georgia and Tennessee. During this outbreak, multiple Ducommun Miltec-developed infrasound sensors collecting continuous, high resolution data were deployed in two-dimensional array configurations in Northern Alabama. Prior research on the collection and analysis of infrasonic emissions from severe weather phenomenon has provided much insight on the nature of tornado-generated infrasound. Our effort focuses on the application of novel bearing estimation algorithms using closely spaced (4-6 m) array elements. Direction of Arrival (DOA) estimates, derived from Blind Source Separation (BSS) techniques, will be presented for at least two significant tornadoes: the long-track EF5 that impacted Hackleburg and Phil Campbell, AL and the large multi-vortex EF4 that struck Cullman, AL. Correlation of infrasound detection and bearing estimate initiation and termination with NOAA Storm Prediction Center (SPC) Storm Reports will also be reviewed.

Locating infrasound events in wind with dense distributed arrays

Applications for infrasound suffer due to the presence of wind noise in the sensor data. Several techniques have been and are being used to mitigate the influence of wind noise in locating infrasonic signals of interest. One of these methods is the averaging of data from multiple sensors in a distributed array. A distributed array was employed to study infrasonic signals from airborne sources. Wind noise during testing masked the location of the infrasonic signals in the time domain. Postprocessing techniques using statistical measures for dense arrays were employed to recover the onset of the infrasonic events. Data from these tests and an explanation of the postprocessing techniques will be discussed.

Time‐reversal maximum‐length sequence pairs for simultaneous acoustical dual source measurements

A binary sequence generated by a linear shift‐register (maximum‐length sequence, M‐sequence) and its reversed‐order sequence form a time‐reversal (reciprocal) M‐sequence pair. Their correlation property includes a two‐valued pulse‐like autocorrelation function and a relatively smaller‐valued crosscorrelation function. This unique property, along with other number‐theory properties, makes the time‐reversal MLS pairs suitable for simultaneous dual source crosscorrelation measurements. In the simultaneous measurement of a dual source system, each of the time‐reversal MLS pair simultaneously excites one of two separate sources, one or several receiver signals cross‐correlate in turn with each of the MLS pairs resulting in impulse responses associated with two separate sources. The proposed method is particularly valuable for system identification tasks with multiple sound/vibration sources and receivers that have to be accomplished in a limited time period. A fast algorithm: fast MLS transform pair is developed for the crosscorrelation. Its feasibility and potential applications in the acoustical measurements are demonstrated using recent field experimental results.A binary sequence generated by a linear shift‐register (maximum‐length sequence, M‐sequence) and its reversed‐order sequence form a time‐reversal (reciprocal) M‐sequence pair. Their correlation property includes a two‐valued pulse‐like autocorrelation function and a relatively smaller‐valued crosscorrelation function. This unique property, along with other number‐theory properties, makes the time‐reversal MLS pairs suitable for simultaneous dual source crosscorrelation measurements. In the simultaneous measurement of a dual source system, each of the time‐reversal MLS pair simultaneously excites one of two separate sources, one or several receiver signals cross‐correlate in turn with each of the MLS pairs resulting in impulse responses associated with two separate sources. The proposed method is particularly valuable for system identification tasks with multiple sound/vibration sources and receivers that have to be accomplished in a limited time period. A fast algorithm: fast MLS transform pair is develop...

Array processing for direction of arrival of infrasound.

Pipe arrays have conventionally been used to detect infrasound. The advantage of this approach is that the arrays have a large aperture over which wind noise can be averaged. In recent years, distributed arrays have been used for the same effect. In addition to reducing wind noise via spatial averaging, these arrays have the ability to process the individual signals from each sensor to estimate direction of arrival (DOA) of acoustic signals and, potentially, the direction of the wind flow. This is especially true for infrasound and low frequency acoustic sources of tactical interest, in the 1–100 Hz range. A frequency‐wavenumber (F‐K) processing technique has been applied to signals from rectangular infrasound arrays to estimate DOA. This approach will be discussed as well as the dependence of accuracy on signal‐to‐noise ratio.

Detecting blast-induced infrasound in wind noise

Current efforts seek to monitor and investigate such naturally occurring events as volcanic eruptions, hurricanes, bolides entering the atmosphere, earthquakes, and tsunamis by the infrasound they generate. Often, detection of the infrasound signal is limited by the masking effect of wind noise. This paper describes the use of a distributed array to detect infrasound signals from four atmospheric detonations at White Sands Missile Range in New Mexico, USA in 2006. Three of the blasts occurred during times of low wind noise and were easily observed with array processing techniques. One blast was obscured by high wind conditions. The results of signal processing are presented that allowed localization of the blast-induced signals in the presence of wind noise in the array response.