In-fiber doppler velocimeter for velocity measurements of moving surfaces

An optical fiber Doppler velocimeter using a 3×3 coupler with a large path imbalance Mach-Zehnder interferometer for passive demodulation of Doppler wavelength shifts is demonstrated. The particle velocity on the free surface end of the Hopkinson bar is directly obtained using the all-fiber Doppler velocimeter. These data show good agreement with the particle velocity predicted using strain gage data combined with one-dimensional stress wave propagation theory.     

Velocity measurements of wake-induced unsteady flow in a linear turbine cascade

Extensive velocity measurements have been taken in a linear turbine cascade with unsteady oncoming wakes. The unsteady wakes were generated by moving cylinders on a squirrel cage device. The Reynolds number was 1.1 × 10⁵, and the Strouhal number varied from o to 7.36. The blade-to-blade flow and the boundary layers on the suction side were measured with a hot-wire anemometer. The results were obtained in ensemble-averaged form so that periodic unsteady processes can be studied. Of particular interest was the transition of the boundary layer. The boundary layer remained laminar in the case without wakes. The passing wakes caused transition, and the beginning of transition moves forward as the wake-passing frequency increases. Unlike in the flat plate study of Liu and Rodi (1991a) the boundary layer state hardly changed with time, although the turbulence level in the boundary layer showed clear periodic response to the passing wakes. The work reported here was sponsored by the German Federal Ministry of Research and Technology through program TURBOTHERM under contract no. 0326501D. The authors should like to thank Mr. D. Bierwirth for his excellent technician work on this project, Dr. N. H. Cho for his help with the preparation of the plots and Mrs. R. Zschernitz for her expert typing of the text.     

Error estimation of laser-Doppler anemometry measurements in fluids with spatial inhomogeneities of the refractive index

We report a combined experimental and theoretical investigation of the influence of spatial non-uniformities of the refractive index on the accuracy of laser Doppler anemometry (LDA) measurements in transparent fluids. One LDA beam is guided through heated air of a thermal boundary layer near a heated vertical flat plate. It is found that the hot air is deflecting the beam because of a modification of the refractive index n in the fluid. This deflection causes three effects: (1) spatial displacement of beam intersection, (2) waist mismatch in the measurement volume and (3) variation in interference fringe distance. With the help of a rotating disk calibration system the resulting displacement of the LDA measurement volume and the Doppler frequency variation is systematically studied at different temperatures. Using a simple model of beam propagation under the influence of well-defined temperature inhomogeneities, the displacement of measurement volume and change in Doppler frequency are calculated and are found to be in agreement with the experimental observations. The results provide a rational framework for an assessment of the accuracy of LDA data in arbitrary transparent fluids with non-uniform refractive index.     

Constituency measurements in the mixing region of a cross flow jet using a laser velocimeter

The velocities in the mixing region of a cross flow jet injected into a freestream were studied in detail with a laser velocimeter. Three jet to freestream momentum ratios were used (3.1, 8.1, 16.2). By purposely seeding the jet and freestream separately (as well as both simultaneously), marking the fluid was feasible. Thus, determining the velocities that emanated from the different streams was possible. By methodically analyzing the three sets of dependent data, the size and location of the mixing region was determined. The mixing regions for the three momentum ratios were found to be of different sizes and at different locations. By proper scaling, however, the regions for the three momentum ratios were found to collapse to one scaled region. Because of the intermittent behavior of the mixing, conventional turbulence models for such mixing may not be applicable; however, detailed velocities and turbulence quantities are included for benchmarking predictions.List of symbols B slot width - H channel height - MR momentum ratio, jet to free stream = _j V _j ²/ U ² - Re _H Reynolds number, U H/v - U free stream velocity - u axial velocity - u rms of axial velocity fluctuation - v transverse velocity - v rms of transverse velocity fluctuation - V _j slot exit transverse velocity - x axial direction (Fig. 3) - x _c x-center of mixing region - scaled value of x, = x/B - y transverse direction (Fig. 3) - y _c y-center of mixing region - scaled value of y, = y/ MRB - _x mixing region width in x-direction - _y mixing region width in y-direction - scaled mixing region width in x-direction, = _x /B - scaled mixing region width in y-direction, = _y / MRB - free stream density - _j slot exit density - v kinematic viscosity of freestream This research was sponsored in part by the Fulbright Commission (Bonn, Germany), the Institut für Thermische Strömungsmaschinen, Universität Karlsruhe (Karlsruhe, Germany), and the Rotating Machinery and Controls Industrial Research Program, University of Virginia (Charlottesville, VA, USA)     

Turbulence measurements in the inlet plane of a centrifugal compressor vaneless diffuser

Detailed flow measurements at the inlet of a centrifugal compressor vaneless diffuser are presented. The mean 3-d velocities and six Reynolds stress components tensor are used to determine the turbulence production terms which lead to total pressure loss. High levels of turbulence kinetic energy were observed in both the blade and passage wakes, but these were only associated with high Reynolds stresses in the blade wakes. For this reason the blade wakes mixed out rapidly, whereas the passage wake maintained its size, but was redistributed across the full length of the shroud wall. Peak levels of Reynolds stress occurred in regions of high velocity shear and streamline curvature which would tend to destabilize the shear gradient. Four regions in the flow are identified as potential sources of loss - the blade wake, the shear layers between passage wake and jet, the thickened hub boundary layer and the interaction region between the secondary flow within the blade wake and the passage vortex. The blade wakes generate most turbulence, with smaller contributions from the hub boundary layer and secondary flows, but no significant contribution is apparent from the passage wake shear layers.     

The effect of spatial wandering on experimental laser velocimeter measurements of the end-wall vortices in an axial-flow pump

In a high Reynolds number axial-flow pump, laser velocimeter (LV) measurements were made to study the size and structure of the end-wall vortex. The time mean measurements show that the core size of the end-wall vortex increased with decreasing tip clearance, which is contrary to existing theory. Observations of cavitation in the vortex showed that the flow was unsteady. The vortices emanating from the smaller clearances were observed to wander or meander spatially and to develop kinks more than the vortices emanating from the larger tip clearances. This observed unsteadiness has a significant effect on the time mean size and velocity distribution of the vortex as measured with the LV employing the field point measurement technique. In order to obtain an estimate of the true size and velocity distribution, computational experiments were conducted which modelled a periodically wandering vortex and the LV measurement process. The computational and experimental results show good agreement, including a broadened and reduced tangential velocity distribution. In this paper, the end-wall vortex LV measurements are presented, and the method of analyzing the vortex wandering is described.     

Control of the corner separation in a compressor cascade by steady and unsteady plasma aerodynamic actuation

This paper reports experimental results on using steady and unsteady plasma aerodynamic actuation to control the corner separation, which forms over the suction surface and end wall corner of a compressor cascade blade passage. Total pressure recovery coefficient distribution was adopted to evaluate the corner separation. Corner separation causes significant total pressure loss even when the angle of attack is 0°. Both steady and unsteady plasma aerodynamic actuations suppress the corner separation effectively. The control effect obtained by the electrode pair at 25% chord length is as effective as that obtained by all four electrode pairs. Increasing the applied voltage improves the control effect while it augments the power requirement. Increasing the Reynolds number or the angle of attack makes the corner separation more difficult to control. The unsteady actuation is much more effective and requires less power due to the coupling between the unsteady actuation and the separated flow. Duty cycle and excitation frequency are key parameters in unsteady plasma flow control. There are thresholds in both the duty cycle and the excitation frequency, above which the control effect saturates. The maximum relative reduction in total pressure loss coefficient achieved is up to 28% at 70% blade span. The obvious difference between steady and unsteady actuation may be that wall jet governs the flow control effect of steady actuation, while much more vortex induced by unsteady actuation is the reason for better control effect.     

General framework for the identification of constitutive parameters from full-field measurements in linear elasticity

In this paper, several approaches available in the literature for identifying the constitutive parameters of linear elastic materials from full-field measurements are presented and their sensitivity to a white noise added to the data is compared. The first investigated approach is the virtual fields method (VFM). It is shown that the uncertainty of the parameters identified with the VFM when a white noise is added to the data depends on the choice of a relevant set of virtual fields. Optimal virtual fields exist, thus minimizing the uncertainty and providing the “maximum likelihood solution”. The other approaches investigated in this paper are based on finite element model updating (FEMU). It is proved that FEMU approaches actually yield equations similar to the ones derived from the VFM, but with nonoptimal sets of virtual fields. Therefore, the FEMU approaches do not provide the “maximum likelihood solution”. However, the uncertainty of FEMU approaches varies dramatically with the cost function to minimize. On one hand, the FEMU approach based on the “displacement gap” minimization yields equations which are very close to the ones of the VFM approach and therefore, its uncertainty is almost the same as the VFM one. On the other hand, it is shown that other approaches based on the “constitutive equation gap” minimization or the “equilibrium gap” minimization provide biased solutions. For all the approaches, very fast algorithms, converging in only two iterations, have been devised. They are finally applied to real experimental data obtained on an orthotropic composite material. Results confirm the success of two methods: the VFM approach which provides the “maximum likelihood solution” and the FEMU approach based on the “displacement gap” minimization.     

Full-field strain measurements of the inside of an agricultural tyre using digital image correlation

This paper investigates strain measured on the inside of an agricultural tyre with large tread-blocks during a series of static tests using a novel measuring system. The full field strain measurements may be used in the development of a tyre which is capable of estimating tyre forces from strain measurements. The strain measurement system makes use of a calibrated stereo camera system on a mechanical stabilizing system that keeps the cameras pointed at the inside surface of the tyre in contact with the road while the wheel rotates. A static tyre test rig is used to displace the road surface relative to the tyre in the vertical and longitudinal direction. The large tread-blocks caused strain concentrations on the inner surface as the tyre deforms to comply with the road surface. Vertical and longitudinal tests each produce unique strain patterns in the contact patch region. Relationships between the applied forces and strain measurements were developed and showed that these relationships are near linear with R² values above 0.97. The strain measurements also show that the location where strain gauges, for single point strain measurements, are placed inside the tyre is very important on large lugged tyres.     

Discrepancies between linear viscoelastic measurements and double-reptation predictions for commercial polystyrene samples

We have found discrepancies between the predictions of two double-reptation models and the corresponding linear viscoelasticity measurements for commercial, polydisperse polystyrene samples with weight-average molecular weights ranging from 140 to 300kg/mol. The accuracy of the experimental data has been confirmed by conducting viscoelasticity measurements in different laboratories using different types of rheometers and by verifying that small distortions in chromatographic measurements cannot account for the discrepancies seen in the viscoelastic response. In addition, we show that the discrepancies between predictions and measurements are systematic, suggesting that gaps in the theory are responsible for the mismatch. Is it concluded that commercial polystyrene resins may pose additional challenges for rheological modeling because of their relatively high polydispersity and low degree of entanglement. The experimental data given here can be used to validate future modeling efforts.     

PIV measurements of the flow field inside an enclosed cubical cavity in natural convection

The natural convective velocity field in an enclosed air-filled cubical cavity with two opposing isothermal faces and the remaining four sides having a well-defined linear temperature rise from the cold to the hot face has been measured at different physical orientations and Rayleigh (Ra) numbers. In particular, two components of the velocity at the mid-plane have been measured by using particle image velocimetry (PIV) at Ra = 10⁶ and 6 × 10⁶ at each of two different physical orientations: heating-from-the-side (HFS), and heating-from-below (HFB). The 95% confidence limit uncertainties in the measured velocity vectors are about 2% for laminar flow. The accuracy and integrity of the experiments were validated by the comparison to some well-established CFD results at the HFS orientation at Ra = 10⁶. It was concluded that the experimental method is sound and so findings at other orientations and at other values of Ra should have an accuracy consistent with the findings of the uncertainty analysis. Therefore, the other results can be confidently used as benchmark data for testing CFD codes. The turbulence intensities at the mid-plane are also presented.     

Buffeting of large telescopes: Wind-tunnel measurements of the flow inside a generic enclosure

The design of future large optical telescopes must take into account the wind-induced vibration of the telescope structure that is caused by large-scale flow structures and turbulence inside the telescope enclosure when the aperture in the enclosure dome is open. However, estimating the resulting degradation in image quality is difficult due to our relatively poor understanding of the flow inside the enclosure. Data has been collected in a wind-tunnel test of an empty telescope enclosure to understand the flow-field around the region near the dome opening where the secondary mirror and supporting structure would be subjected to wind loads. Digital particle image velocimetry (DPIV) data was collected in a vertical plane near the dome opening to obtain mean velocity and fluctuation kinetic energy, and hot-wire data was collected along the telescope axis to obtain temporal spectra of the velocity. Data was collected both with and without additional venting, in order to understand its influence on the flow. The temporal spectrum of the broadband turbulence is of von Karman type. The flow field also includes tonal shear layer modes, with the mode selection influenced by the enclosure Helmholtz mode, and the overall amplitude strongly influenced by the vented area. In addition to its direct use in telescope modelling and design, the data presented herein is of particular value in validation of computational fluid dynamic (CFD) analyses, so that CFD can be used with confidence in future design work.     

A three-dimensional blockage theory for the stability of an axial compressor row

The concept of ‘blockage’ between the blades of an axial compressor row as a theoretical device to replace loss was used by Stenning and Kriebel (1) to derive their well-known stability criterion for the stability of two-dimensional cascades. In their work they ignored the influence of that blockage upon the term describing the inertia of the fluid between the blades. The first part of the present work includes this influence and a modified stability criterion is derived that links the stability parameter with ‘stall cell’ speed and number. In the second part the blockage principle is extended to the problem of an isolated rotor row of arbitrary hub/tip ratio, treated as a ‘semi-actuator disc’, and perturbation of the upstream potential arising from changes of blockage are studied. The method proceeds by analysing the response of the flow to an imposed pressure discontinuity within the blade row. Rotating stall modes as such are not predicted but the theory demonstrates regions of the span which are undamped and on this basis a comparison can be made with two-dimensional ‘strip’ theory and it is shown that blades of low aspect ratio have broader undamped regions when zero damping is first encountered.     

多重网格法求解轴流压气机内部流场

提出了一种能够与显式时间推进法、有限体积差分格式、粘性体积力法有效地配合使用的有效地配合使用的多重网格法来解轴流压气机内部流场。作者对美国NASA37^#跨音速轴流压气机转子内部流场的数值计算表明,多重网格法的计算结果与试验结果吻合, 同时该多重网格法的收敛速度比不用多重网格法加快了1．5倍。     

Experimental study on the effect of unsteadiness on boundary layer development on a linear turbine cascade

 The results from an experimental investigation of unsteady boundary layer behavior on a linear turbine cascade are presented in this paper. To perform a detailed study on unsteady cascade aerodynamics and heat transfer, a new large-scale, high-subsonic research facility for simulating the periodic unsteady flow has been developed. It is capable of sequentially generating up to four different unsteady inlet flow conditions that lead to four different passing frequencies, wake structures, and freestream turbulence intensities. For a given Reynolds number, two different unsteady wake formations are utilized. Detailed unsteady boundary layer velocity. turbulence intensity, and pressure measurements are performed along the suction and pressure surfaces of one blade. The results display the transition and development of the boundary layer, ensemble-averaged velocity, and turbulence intensity. Received: 23 September 1996/Accepted: 19 February 1997