Bulletin of the American Physical Society
67th Annual Meeting of the APS Division of Fluid Dynamics
Volume 59, Number 20
Sunday–Tuesday, November 23–25, 2014; San Francisco, California
Session M20: Experimental Techniques: PIV II |
Hide Abstracts |
Chair: Alexandra Techet, Massachusetts Institute of Technology Room: 2008 |
Tuesday, November 25, 2014 8:00AM - 8:13AM |
M20.00001: An experimental database for evaluating PIV uncertainty quantification methods Scott Warner, Douglas Neal, Andrea Sciacchitano Uncertainty quantification for particle image velocimetry (PIV) data has recently become a topic of great interest as shown by the publishing of several different methods within the past few years. A unique experiment has been designed to test the efficacy of PIV uncertainty methods, using a rectangular jet as the flow field. The novel aspect of the experimental setup consists of simultaneous measurements by means of two different time-resolved PIV systems and a hot-wire anemometer (HWA). The first PIV system, called the ``PIV-Measurement'' system, collects the data for which uncertainty is to be evaluated. It is based on a single camera and features a dynamic velocity range (DVR) representative of many PIV experiments. The second PIV system, called the ``PIV-HDR'' (high dynamic range) system, has a significantly higher DVR obtained with a higher digital imaging resolution. The hot-wire was placed in close proximity to the PIV measurement domain. All three of the measurement systems were carefully set to simultaneously collect time-resolved data on a point-by-point basis. The HWA validates the PIV-HDR system as the reference velocity so that it can be used to evaluate the instantaneous error in the PIV-measurement system. [Preview Abstract] |
Tuesday, November 25, 2014 8:13AM - 8:26AM |
M20.00002: New in-situ, non-intrusive calibration Heather Zunino, Ronald Adrian, Liuyang Ding, Kathy Prestridge Tomographic particle image velocimetry (PIV) experiments require precise and accurate camera calibration. Standard techniques make assumptions about hard-to-measure camera parameters (i.e. optical axis angle, distortions, etc.)--reducing the calibration accuracy. Additionally, vibrations and slight movements after calibration may cause significant errors--particularly for tomographic PIV. These problems are exacerbated when a calibration target cannot be placed within the test section. A new PIV camera calibration method has been developed to permit precise calibration without placing a calibration target inside the test section or scanning the target over a volume. The method is capable of correcting for dynamic calibration changes occurring between PIV laser pulses. A transparent calibration plate with fine marks on both sides is positioned on the test section window. Dual-plane mapping makes it possible to determine a mapping function containing both position and angular direction of central rays from particles. From this information, central rays can be traced into the test section with high accuracy. Image distortion by the lens and refraction at various air-glass-liquid interfaces are accounted for, and no information about the position or angle of the camera(s) is required. [Preview Abstract] |
Tuesday, November 25, 2014 8:26AM - 8:39AM |
M20.00003: Iterative Particle Image Velocimetry Algorithm for Rotating Flows Matthew Giarra, John Charonko, Pavlos Vlachos Particle image velocimetry (PIV) can fail to reliably estimate fluid velocities in flows with large spatial velocity gradients because the shearing, stretching, and rotation of particle image patterns decreases the signal-to-noise ratio of cross correlations (CCs). We present a new PIV correlation algorithm called the Fourier-Mellin correlation (FMC) that accurately measures particle pattern displacements in flow regions with large rotation (like vortex cores) compared to traditional correlations by measuring rotation and then aligning particle patterns before performing Cartesian CCs. FMC reliably measures particle displacements between interrogation regions with up to 180 degrees of angular misalignment compared to 6-8 degrees for traditional correlations. We combined our FMC algorithm with iterative discrete window offset (DWO) to measure velocity and vorticity fields in synthetic PIV images of counter-rotating vortex cores and an experimental vortex ring in water. FMC with DWO reduced the errors in velocity and vorticity estimates by an order of magnitude compared to traditional correlations with DWO, increased the correlation peak height ratios in synthetic and experimental images, and accelerated the convergence of iterative image deformation algorithms. [Preview Abstract] |
Tuesday, November 25, 2014 8:39AM - 8:52AM |
M20.00004: Uncertainty Estimation in Stereoscopic Particle Image Velocimetry Sayantan Bhattacharya, Pavlos Vlachos In Stereoscopic Particle Image Velocimetry (Stereo-PIV) particle images are recorded using two viewing directions and the projected velocity components obtained in each view are combined to predict the three component velocity vector in the plane of measurement. The accuracy of the method depends on precise determination of viewing angles, measurement plane location and estimation of projected velocity components. However, the complex measurement chain with non-linear combination of errors make uncertainty estimation in Stereo-PIV challenging. Here we consider the overall uncertainty stemming from various error sources involved in the measurement process. The uncertainty in the absolute particle locations due to mismatch in the overlapping camera views are combined with the uncertainty in individual camera velocity components to predict the uncertainty in the reconstructed velocity field. The mapping function uncertainty and viewing angle uncertainty are also considered in the propagation equation. Present framework is tested with both simulated random field and experimental vortex ring image set. The RMS error and predicted uncertainties are compared for different viewing angle camera pairs. A sensitivity analysis of the individual uncertainty contributions to the overall uncertainty coverage is also presented. [Preview Abstract] |
Tuesday, November 25, 2014 8:52AM - 9:05AM |
M20.00005: Comparative assessment of four a-posteriori uncertainty quantification methods for PIV data Pavlos Vlachos, Andrea Sciacchitano, Douglas Neal, Barton Smith, Scott Warner Particle Image Velocimetry (PIV) is a well-established technique for the measurement of the flow velocity in a two or three dimensional domain. As in any other technique, PIV data are affected by measurement errors, defined as the difference between the measured velocity and its actual value, which is unknown. The objective of uncertainty quantification is estimating an interval that contains the (unknown) actual error magnitude with a certain probability. In the present work, four methods for the \textit{a-posteriori} uncertainty quantification of PIV data are assessed. The methods are: the uncertainty surface method (Timmins \textit{et al.}, 2012), the particle disparity approach (Sciacchitano \textit{et al.}, 2013; the peak ratio approach (Charonko and Vlachos, 2013) and the correlation statistics method (Wieneke 2014). For the assessment, a dedicated experimental database of a rectangular jet flow has been produced (Neal \textit{et al.} 2014) where a reference velocity is known with a high degree of confidence. The comparative assessment has shown strengths and weaknesses of the four uncertainty quantification methods under different flow fields and imaging conditions. [Preview Abstract] |
Tuesday, November 25, 2014 9:05AM - 9:18AM |
M20.00006: Direct Estimation of Particle Image Velocimetry Measurement Uncertainty from Cross-Correlation Plane Moments Zhenyu Xue, Sayantan Bhattacharya, John Charonko, Pavlos Vlachos Particle Image Velocimetry is a non-invasive measurement technique in which images of flow tracers are correlated to estimate flow velocity. The coupled effect of error sources including particle image size, velocity gradient, out of plane motion, and seeding density poses a challenge in quantifying the uncertainty. Here we establish a method to quantify PIV uncertainty by extracting the Probability Density Function (PDF) of all possible displacements from the cross-correlation plane. The PDF is obtained by deconvolving particle image size from the correlation plane, and approximating its shape and standard deviation by an elliptic Gaussian least squares fit. The PDF variance is then scaled by a normalized estimate of the number of correlated particles between the image pairs to obtain the standard uncertainty. The method takes into account the peak stretching due to velocity gradients and also includes an estimate of bias error. The calculated uncertainty is compared with the RMS error for synthetic and experimental images, including a vortex ring and the recent uncertainty benchmark jet flow cases. Results show reasonable uncertainty coverage. Thus, the current framework provides a direct approach to quantify PIV uncertainty from the correlation plane. [Preview Abstract] |
Tuesday, November 25, 2014 9:18AM - 9:31AM |
M20.00007: Benchmark measurements for evaluation of PIV uncertainty method Stamatios Pothos, Sayantan Bhattacharya, Pavlos Vlachos, Dan Troolin, Wing Lai PIV combines a series of instruments, algorithms and user inputs in order to quantify the displacement of flow tracer patterns in complex flows. Each of these components is bound to introduce uncertainty in the resulting measurement, and often these uncertainties are coupled or difficult to estimate. Recent developments have now presented a series of methods for quantification of uncertainty in planar PIV measurements, however each of these methods appears to offer different advantages or disadvantages and their strengths and weaknesses are not well understood. Moreover, there is a need for extensive testing of these methods against a variety of real experimental data and flow conditions. In this work we execute a benchmark experiment of a flow over a cylinder using time resolved PIV with simultaneous LDV measurements to serve as a comparison benchmark, and we use these data to compare the different uncertainty quantification methods and assess their reliability. The presented comparisons will include signal to noise ratio methods, image disparity methods and correlation plane statistics and the estimated uncertainties will be assessed using error probability distributions, time series analysis, and coverage factors. [Preview Abstract] |
Tuesday, November 25, 2014 9:31AM - 9:44AM |
M20.00008: Measurement uncertainty of mean velocity fields acquired by PIV Sven Scharnowski, Christian J. K\"ahler Particle Image Velocimetry (PIV) has become a standard tool for the investigation of various flow fields. In order to compare the mean velocity distributions or higher order statistics from experiments and numerical predictions, it is essential to know the uncertainty of the estimated values. However, due to the complex evaluation procedure of PIV the error cannot be estimated with standard methods. Many parameters, including particle image size, particle image density, turbulence level, noise level, velocity gradients, number if PIV image pairs \textellipsis affect the accuracy. This work systematically analyzes the effect of several parameters on the random and bias errors of the estimated mean velocity by using single-pixel ensemble-correlation as well as window-correlation based PIV. To have full control of all parameters, synthetic PIV images are generated and analyzed, while identifying the most relevant error sources. The different parameters can be determined from the raw data by generating a multidimensional uncertainty hyper-surface that allows for determining the random error of the shift vectors. Furthermore, the knowledge about the dependency on the different parameters enables to identify the bottleneck and thus, to optimize the measurement setup and evaluation procedure to improve the accuracy. [Preview Abstract] |
Tuesday, November 25, 2014 9:44AM - 9:57AM |
M20.00009: Image preprocessing method for particle image velocimetry (PIV) image interrogation near a fluid-solid surface Yiding Zhu, Lichao Jia, Ye Bai, Huijing Yuan, Cunbiao Lee Accurate particle image velocimetry (PIV) measurements near the moving wall are a great challenge. The problem is compounded by the very large in-plane displacement on PIV images commonly encountered in measurements of the high speed flow. An improved image preprocessing method is presented in this paper. A wall detection technique is used first to qualify the wall position and the movement of the solid body. Virtual particle images are imposed in the solid region, of which the displacements are evaluated by the body movement. The estimation near the wall is then smoothed by data from both sides of the shear layer to reduce the large random uncertainties. Interrogations in the following iterative steps then converge to the correct results to provide accurate predictions for particle tracking velocimetries(PTV). Significant improvement is seen in Monte Carlo simulations and experimental tests such as measurements near a flapping flag or compressor plates. The algorithm also successfully extracted the small flow structures of the 2nd mode wave in the hypersonic boundary layer from PIV images with low signal-noise-ratios(SNR) when the traditional method was not successful. [Preview Abstract] |
Tuesday, November 25, 2014 9:57AM - 10:10AM |
M20.00010: Two-colour micro-PIV and high speed shadowgraphy measurements for liquid-liquid plug flows Maxime Chinaud, Dimitrios Tsaoulidis, Panagiota Angeli Two-colour micro-Particle Image Velocimetry (micro-PIV) is a relatively new technique that provides velocity fields simultaneously in both phases of a two-phase flow system. In this work, a laser emitting at two different wavelengths was used to excite two different types of particles, each added in one of the liquid phases of a two-phase, oil-water, system. The two types of particles emitted signals at separate wavelengths that were captured simultaneously by two different cameras. Instantaneous velocity fields could thus be obtained in both phases at the same time. This technique was used to study liquid-liquid plug flows in microchannels. Both plug propagation in the main channel and plug formation in the T-shaped inlet junction have been investigated. During plug propagation analysis of the velocity fields reveals recirculation patterns inside the dispersed plug and the continuous slug. These will be related to dimensionless numbers. The results on plug formation will be discussed against current models on plug size. [Preview Abstract] |
Follow Us |
Engage
Become an APS Member |
My APS
Renew Membership |
Information for |
About APSThe American Physical Society (APS) is a non-profit membership organization working to advance the knowledge of physics. |
© 2024 American Physical Society
| All rights reserved | Terms of Use
| Contact Us
Headquarters
1 Physics Ellipse, College Park, MD 20740-3844
(301) 209-3200
Editorial Office
100 Motor Pkwy, Suite 110, Hauppauge, NY 11788
(631) 591-4000
Office of Public Affairs
529 14th St NW, Suite 1050, Washington, D.C. 20045-2001
(202) 662-8700