William K. Bonness

Low‐wavenumber turbulent boundary layer wall‐pressure measurements from vibration data on smooth and rough cylinders in pipe flow.

The vibration response of a thin cylindrical shell excited by a fully‐developed turbulent boundary layer is measured and used to extract the fluctuating pressure levels generated by the boundary layer. Parameters used to extract the turbulent boundary layer pressure levels are determined via experimental modal analysis of the water‐filled pipe and measured vibration levels from flow through the pipe at 6 m/s. Hydrostatic head from a large reservoir provides the low‐noise source of steady flow for measuring the low‐wavenumber fluctuating pressure levels. Measurements are reported for smooth, transitionally rough, and fully rough conditions and are compared to the turbulent boundary layer pressure models of Chase, Smol’yakov, and Howe.

Low Wavenumber Models for Turbulent Boundary Layer Excitation of Structures

When the spatial correlation length of the turbulent boundary layer (TBL) pressure fluctuations is small compared to the structural wavelengths, the vibration response can be determined by forming an equivalent point drive from the effective correlation area. This approach is equivalent to using the zero wavenumber component of the TBL pressure spectrum, so it only works for TBL models that are wavenumber white at low wavenumbers. In this work, a similar simplification is developed for TBL models with a wavenumber-squared dependence, that works for structural modes with a low-pass cutoff wavenumber. This introduces a boundary layer thickness dependence that results in significantly different predictions for structures excited by a developing boundary layer. Based on the analysis, an experimental setup is proposed that may help resolve some of the controversy surrounding the low wavenumber TBL spectrum.

Removing Unwanted Noise from Operational Modal Analysis Data

Operational modal analysis data includes the measurement of dynamic signals such as structural vibration data and the corresponding excitation force or pressure. In addition to the desired information, measured structural vibration data can include unwanted electrical noise and vibration energy from adjacent structures. Measured dynamic pressures can contain unwanted signals such as acoustic and vibration induced pressures. In this paper, a noise removal technique is presented in which an unlimited number of unwanted correlated signals can be removed from a set of measured data. In its simplest form, this technique is related to coherent output power (COP). However, unlike COP noise removal, multiple signals can be removed from measured data while retaining the magnitude and phase of the original data required for modal analysis processing. This technique is demonstrated using vibration data and dynamic wall pressure measurements from a thin-walled aluminum cylinder filled with water flowing at 20 ft/s.

Acoustic cavitation localization in reverberant environments

Cavitation detection and localization techniques generally require visual access to the fluid field, multiple high-speed cameras, and appropriate illumination to locate cavitation. This can be costly and is not always suitable for all test environments, particularly when the bubble diameter is small or duration is short. Acoustic detection and localization of cavitation can be more robust and more easily implemented, without requiring visual access to the site in question. This research utilizes the distinct acoustic signature of cavitation events to both detect and localize cavitation during experimental water tunnel testing. Using 22 hydrophones and the processing techniques plane-wave beamforming and Matched-Field Processing (MFP), cavitation is accurately and quickly localized during testing in a 12” diameter water tunnel. Cavitation is induced using a Nd:YAG laser for precise control of bubble location and repeatability. Accounting for and overcoming the effects of reflections on acoustic localizatio...

Determination of pipe interior pressures using external accelerometers.

A non‐invasive method utilizing a ring of accelerometers to measure pipe interior pressures is presented. The internal acoustic pressure of the fluid inside a cylindrical pipe is directly related to the vibrations at the surface due to the coupling of the fluid and structure. The n=0 breathing mode provides the basis for this relation and is extracted from operational data using a circumferential modal decomposition routine. The measurement of pipe interior pressures can be useful in determining the integrity of a piping system by detecting high pressures, which may lead to the fatigue or failure of the system. Within the field, a non‐invasive method for measuring these pressures is more practical in comparison to the intrusive but more direct method of using hydrophones. Measurements using the ring of accelerometers are compared to hydrophone measurements to determine accuracy.

Low wavenumber turbulent boundary layer fluctuating wall shear stress measurements from vibration data on a cylinder in pipe flow.

Fluctuating wall shear stress under turbulent boundary layer (TBL) excitation is studied in this experimental investigation. A cylindrical shell with a smooth internal surface is subjected to TBL excitation from water in fully developed pipe flow at 6.1 m/s. The vibration response of the cylinder is used to inversely determine low‐wavenumber TBL shear stress levels. Both the cross‐flow and streamwise directions are examined using directionally uncoupled low‐order cylinder modes in the circumferential and axial directions. These data address a critical gap in available literature regarding experimental low‐wavenumber shear stress data. The low‐wavenumber shear stress levels in both the cross‐flow and streamwise directions are determined to be roughly 10 dB higher than those of normal pressure. As is the case for various models of TBL pressure, these measurements suggest that a nearly constant value for normalized shear stress at low wavenumber is valid over a broad range of frequencies. A simple wavenumber...

Turbulent boundary layer shear stress transmitted through a viscoelastic layer.

Transfer functions are developed for the transmission of unsteady shear stress, generated by a turbulent boundary layer in water, through a viscoelastic layer backed by a rigid plate. Existing analytical models are used to estimate the unsteady wall pressure and shear stress from 10–1000 Hz for a flat plate boundary layer with zero pressure gradient. A new model is developed for the transmission of unsteady shear stress through the viscoelastic layer. The model is used to predict the unsteady pressure fluctuations, or flow noise (due to the unsteady shear stress), which would be seen by a finite size sensor embedded under the elastomer layer. The calculated unsteady pressure and shear stress levels are in good agreement with recent experimental measurements. The unsteady shear stress transfer functions are found to have a peak at the acoustic wavenumber.

Removing unwanted signals from wall pressure and vibration measurements on a structure subjected to turbulent boundary layer excitation.

Fluid‐structure interaction experiments typically involve measurements of the excitation force (or pressure) and the corresponding vibration response to the excitation. In addition to desired flow information, measured turbulent boundary layer wall pressure data often include unwanted signals such as acoustic pressures and vibration induced pressures. Measured vibration data on a structure can also include unwanted electrical noise and vibration energy from adjacent structures. A noise removal technique is presented, which allows one to remove an unlimited number of unwanted correlated signals from a set of measured data. In its simplest form, this technique is related to the coherent output power. However, the more general technique provides an ability to remove multiple signals and to retain complex values (magnitude and phase). These advantages can yield significantly greater information of the flow field and structure under investigation. This technique is demonstrated using measurements from an alumi...

Low‐wave‐number turbulent boundary layer wall pressure measurements from vibration data on a cylinder in pipe flow

Low‐order vibration modes of a cylinder subjected to a fully developed turbulent boundary layer (TBL) in a pipe are measured and used to determine the low‐wave‐number content of energy in the boundary layer. The experiment is conducted using a 6‐in.‐diameter, thin‐walled, aluminum test section filled with water flowing at 20 ft/s. Coupling of certain cylinder modes in the axial and radial directions allows use of the inverse method to determine the wave‐number levels of TBL wall normal pressure and wall shear stress in the flow direction. These measurements are used to evaluate commonly used empirical models of the TBL wave‐number pressure spectrum attributed to Corcos (1963) and Chase (1987) at lower wave numbers than have previously been reported. Preliminary results from several higher order cylinder modes confirm prior experimental findings that the Corcos model overestimates the low‐wave‐number pressures for these conditions by nearly 20 dB. The Chase model provides estimates within 3 dB of the measu...

Predicting the surface wall pressure frequency spectra of submerged cylindrical bodies

This paper addresses the modeling of structural excitations caused by low speed turbulent boundary layer flows. When the Corcos‐type cross‐spectrum or its corresponding wavevector‐frequency spectrum is used as the forcing function in a homogeneous turbulent flow, the point frequency spectrum is the common multiplier to the factors representing either the cross‐spectral terms or the wavevector terms. The capability of accurately predicting the point frequency spectrum is the premise of obtaining an accurate cross‐spectrum or wavevector‐frequency spectrum. Several point‐frequency spectral models have been evaluated against the measured data obtained from buoyancy propelled cylindrical bodies. It was found that predictions using the most recently published point‐frequency spectral models by Smolyakov [Acoust. Phys. 46(3), 342–347; translated from Akusti. Zh. 46(3), 410–407 (2000)] and Goody [AIAA‐2002‐2565, 8th AIAA/CEAS Aeroacoustics Conference and Exhibit, Breckenridge, CO, (2002)] provide a reasonably goo...