Mark J. Moeller

MEMS pressure sensor array for aeroacoustic analysis of the turbulent boundary layer

The second stage of the design, fabrication, and characterization of a surface micromachined, front-vented, 64 channel (8 8), capacitively sensed pressure sensor array is described. The array was fabricated using the MEMSCAP PolyMUMPs R process, a three layer polysilicon surface micromachining process with an additional fabrication step using Parylene-C. An acoustic lumped element model was used to model an individual microphone and then applied to the array as a whole. The computational results for the design, including mechanical components, environmental loading, uid damping, and other acoustic elements are detailed. Theory predicts single element sensitivity of 0.65 mV/Pa at the gain stage output in the 100-40,000 Hz band. A laser Doppler velocimetry (LDV) system has been used to map the spatial motion of the elements in response to electrostatic excitation. A strong resonance appears at 410 kHz for electrostatic excitation, in agreement with mathematical models. Static stiness measured electrostatically using an interferometer is 0.1 nm/V 2 , similar to the expected stiness.

A Microphone Array on a Chip for High Spatial Resolution Measurements of Turbulence

A microelectromechanical systems-based microphone array on a chip has been developed and applied to aeroacoustic measurements. The array is designed to measure the fluctuating pressures present under a turbulent boundary layer (TBL). Each chip measures 1 cm2 and contains 64 individually addressable capacitively sensed microphones, with a center to center pitch of ~1.25 mm. Surface topology, including the packaging, is kept to less than 0.13 mm. Element-to-element sensitivity variation in the array is less than ±2.5 dB from least to most sensitive, and phase variation is less than ±6.5° (at 1 kHz). The microphone 3-dB bandwidth is 700 Hz to 200 kHz, and the microphones are linear to better than 0.3% at sound pressure levels up to 150-dB SPL. A unique switched architecture system electronics and packaging method are employed to reduce data acquisition channel count requirements, and to maintain a low surface roughness. The array has been applied to the measurement of single point turbulence spectra under a flat plate TBL in a flow duct at Mach numbers up to 0.6 and Reynolds numbers based on plate length of 107.

MEMS Microphone Array on a Chip for Turbulent Boundary Layer Measurements

A MEMS microphone array has been designed and applied to the measurement of wall pressure spectra under the turbulent boundary layer in flow duct testing at Mach numbers from 0 to 0.6. The array was micromachined onto a single chip in the PolyMUMPS surface micromachining process, allowing high spatial resolution and low surface roughness. The chip measures 1 cm by 1 cm, and is flush mounted into the wind tunnel wall. Individual elements are 0.6 mm in diameter, with element to element spacing of 1.11 mm in the crossflow direction and 1.26 mm in the flow direction. The 64 element array has 59 working elements, 58 of which are matched to within ± 2.5 dB at 1 kHz. Phase matching between the 59 elements is ± 6.5 degrees at 1 kHz. The array has been calibrated from 100 Hz to 4 kHz in a plane wave tube. The transducer bandwidth is greater than 400 kHz as determined by laser vibrometry measurements. Sensor nonlinearity of less than 0.36% is observed at a sound pressure level of 150 dB SPL. Board level electronics allow the array to be reconfigured on the fly using computer controlled CMOS switches. Multipoint wall pressure spectra were measured in 38 array configurations at the wall of a 6 inch by 6 inch flow duct at Mach numbers from 0.0 to 0.6. The array shows excellent agreement with Kulite and Bruel & Kjaer microphone measurements in the 300 Hz to 10 kHz band, and appears to be able to measure turbulent pressure spectra at frequencies as high as 40 kHz.

Investigation of Low-Frequency Single Point Wall Pressure Spectrums

A review of existing turbulent boundary layer single point wall pressure spectrum models for acoustic analysis is presented. Important features of the single point wall pressure spectrum models are shown along with the boundary layer parameters that have been used for model generation. The more recent models use a mixture of scaling variables. The models of Efimtsov and Goody are compared to experimental results from the Spirit 6x6 duct at Mach numbers ranging from 0.1 to 0.6. The single point wall pressure spectrums are developed from the experimental data which shows that the mixed variables do the best job of scaling. Little Mach number dependence is noted in the scaled spectrums. The coherence is shown at three microphone separation distances and behaves as expected. The cutoff/corner frequency occurs at the same location for the coherence and single point wall pressure spectrum indicating where the signal begins to drop off. The Spirit 6x6 data is bounded by the two models of Efimtsov and Goody. The spectrum at low frequencies rolls off similar to the Goody model. This analysis indicates that the Goody model is the appropriate single-point, wall pressure spectrum model for subsonic aircraft applications.

Effect of Developing Pressure Gradients on TBL Wall Pressure Spectrums

The effects of favorable and adverse pressure gradients on the shape and pressure spectra of turbulent boundary layers are investigated. A contracting and then expanding wind tunnel test section is used for this investigation. Favorable (negative) pressure gradients are found to cause a modest reduction in the pressure spectrum levels compared to equilibrium flows. Adverse (positive) pressure gradients are found to dramatically increase the pressure spectrum levels and decrease the phase velocity.

An Assessment of SEA Model Quality

Statistical Energy Analysis (SEA) models are routinely being adopted in up-front automotive sound package design. SEA models serve two important functions. First they provide a means of assessing noise and vibration performance relative to absolute targets. Secondly, they are used to assess various alternative designs or changes required to meet targets. This article addresses how to objectively evaluate both the absolute and relative predictive capability of SEA models. The absolute prediction is assessed using a hypothesis test to determine membership of the analytical prediction relative to a set of test data. The relative prediction is assessed using hardware-designed experiments to estimate design sensitivities. Both have been found useful to drive model improvement efforts. Being able to objectively document model capability also improves the credibility of SEA model predictions and the design information they deliver. Competitive cost pressures are forcing design/manufacturing concerns to move from hardware prototype based design strategies to Computer Aided Engineering (CAE) based alternatives. Statistical Energy Analysis (SEA) has been widely adopted as a means of assessing high frequency noise, vibration and harshness (NVH) concerns in the automotive industry. This article addresses how to objectively evaluate the predictive capability concerns assessing the feasibility of using SEA models for this purpose. The quality of CAE models is becoming increasingly important as availability of test information decreases. Several activities have grappled with this issue. The most extreme example would be the nuclear weapons arena, which is being forced to move to a completely CAE based process due to the voluntary nuclear test ban in place. There are active discussions in the literature on determining the quality of many modeling technologies. Computational Fluid Dynamics has an especially active discussion ongoing as evidenced by Roache. 1 Structural safety has similar model quality discussions that can be found in the NAFEMS SAFESA material 2,3 for finite element analysis. Han 4 discusses crash safety model quality. Some organizations have adapted ISO 9001 as a quality system for CAE in their organizations. This article is a continuation of the SEA model quality discussion started by Thomas. 5 It will incorporate efforts toward adapting NVH CAE Quality Metrics, described by Moeller, 6 to high frequency models. The desired result is a framework for objectively assessing suitability to task for high frequency modeling technology. This can then be integrated into an overall quality framework for SEA modeling. The first section develops a metric for assessment of baseline model correlation for transfer function prediction. The next section applies the metrics to current SEA model capability. SEA model design capability is examined next. The last section presents the discussion and conclusions.

Application of Statistical Energy Analysis (SEA) to the Development of a Light Truck Sound Package

This paper describes the application of Statistical Energy Analysis to the development of a light truck sound package. The development concerns addressed included airborne engine noise, tailpipe exhaust noise and rear road noise. SEA provided a framework to investigate different sound package configurations quickly and easily. SEA helped prioritize design alternatives to maximize benefit/cost ratio and to shorten development time. The project resulted in significant contributions to the development and optimization of the sound package and added value to the vehicle.

Wall Pressure Phase Velocity Measurements in a Turbulent Boundary Layer

The turbulent boundary layer that forms on the outer surfaces of vehicles can be a significant source of interior noise. In automobiles this is known as wind noise, and at high speeds it dominates the interior noise. For airplanes the turbulent boundary is also a dominant noise source. Because of its importance as a noise source, it is desirable to have a model of the turbulent wall pressure fluctuations for interior noise prediction. One important parameter in building the wall pressure fluctuation model is the convection velocity. In this paper, the phase velocity was determined from the streamwise pressure measurements. The phase velocity was calculated for three separation distances ranging from 0.25 to 1.30 boundary layer thicknesses. These measurements were made for a Mach number range of 0.1 < M < 0.6. The phase velocity was shown to vary with sensor spacing and frequency. The data collapsed well on outer variable normalization. The phase velocities were fit and the group velocity was calculated from the curve fit. The group velocity was consistent with the array measured convection velocities. The group velocity was also estimated by a band limited cross correlation technique that used the Hilbert transform to find the energy delay. This result was consistent with the group velocity inferred from the phase velocities and the array measured convection velocity. From this research, it is suggested that the group velocity found in this study should be used to estimate the convection velocity in wall pressure fluctuation models.Copyright

High frequency modeling of vehicle noise performance of aluminum body structure.

This presentation is dedicated to Curt Holmer’s contribution to acoustics at Ford Motor Co. The presentation material is based on the work he was involved with during the last part of his life. The results of the 1992 Taurus SEA model are used to assess the effect of an aluminum body structure. The noise performance comparison includes airborne and structure‐borne paths. The effect of structural damping is also simulated.

Micromachined reconfigurable microphone array for wind tunnel testing.

A surface micromachined, front‐vented, 64 channel (8×8), capacitively sensed microphone array‐on‐a‐chip devices for aeroacoustic testing is described. The arrays are fabricated using the MEMSCAP PolyMUMPs polysilicon surface micromachining process, with a Parylene‐C passivation layer. The devices are packaged with low profile interconnects, presenting a maximum of 100 μm of surface topology. The array electronics allow the microphone outputs to be redirected to one of two channels, allowing dynamic reconfiguration of the effective transducer shape in software. Measured microphone sensitivity is 0.15 mV/Pa for an individual microphone and 8.7 mV/Pa for the entire array, in close agreement with model predictions. The microphones and electronics operate over the 200–40 000 Hz band. The dynamic range extends from 60 dB SPL in a 1 Hz band to greater than 150 dB SPL. Element variability is ±0.05 mV/Pa in sensitivity with an array yield of 95%. Off‐chip electronics provide 80 dB off isolation. Preliminary wind t...