Bulletin of the American Physical Society
65th Annual Meeting of the APS Division of Fluid Dynamics
Volume 57, Number 17
Sunday–Tuesday, November 18–20, 2012; San Diego, California
Session G14: Experiments: PIV I |
Hide Abstracts |
Chair: Barton Smith, Utah State University Room: 27B |
Monday, November 19, 2012 8:00AM - 8:13AM |
G14.00001: A super-resolution approach for uncertainty estimation of PIV measurements Bernhard Wieneke, Andrea Sciacchitano, Fulvio Scarano A super-resolution approach is proposed for the a posteriori uncertainty estimation of PIV measurements. The measured velocity field is employed to determine the displacement of individual particle images. A disparity set is built from the residual distance between paired particle images of successive recordings. Within each interrogation window, the disparity set is treated with a statistical analysis to infer the measurement uncertainty: the mean disparity is ascribed to bias errors due to poor particle image sampling or spatial modulation effect; the dispersion of the set is related to precision errors, mainly due to random noise in the recordings and to errors in the PIV interrogation. The performance of the estimator is assessed on a synthetic uniform flow field with varying out-of-plane displacement. The uncertainty is accurately estimated in optimal imaging condition, but underestimated for very small particle images. Experiments are conducted on a water jet experiment, where the actual measurement error is computed as the difference between measured and a reference displacement field estimated from the time redundancy of highly oversampled data. The uncertainty is quantified accurately within 0.1 px. [Preview Abstract] |
Monday, November 19, 2012 8:13AM - 8:26AM |
G14.00002: Mean-flow reconstruction by data-assimilation techniques from PIV-measurements of flow over an idealized airfoil Nicolas Dovetta, Dimitry P.G. Foures, Denis Sipp, Peter J. Schmid, Beverley J. McKeon Measurements (experimental or numerical) typically contain a low-dimensional representation of a high-dimensional flow field. The methodology that extracts the non-measured components of the flow by matching a parametrized model to the data is referred to as data-assimilation. Flow around an idealized airfoil ($Re=20000$) is measured using time-resolved PIV which produces the mean velocity field by averaging over a sequence of snapshots, as well as the velocity fluctuations. The mean flow is assumed to be known in only part of the flow domain. The assumed relationship between a mean flow measurement and the Reynolds Averaged Navier-Stokes equations is used together with a data-assimilation strategy in order to recover the mean flow everywhere from artificially limited input data. The estimated and measured mean flows are compared to illustrate the potential and effectiveness of the data-assimilation technique. [Preview Abstract] |
Monday, November 19, 2012 8:26AM - 8:39AM |
G14.00003: Volumetric PIV with a Plenoptic Camera Brian Thurow, Tim Fahringer Plenoptic cameras have received attention recently due to their ability to computationally refocus an image after it has been acquired. We describe the development of a robust, economical and easy-to-use volumetric PIV technique using a unique plenoptic camera built in our laboratory. The tomographic MART algorithm is used to reconstruct pairs of 3D particle volumes with velocity determined using conventional cross-correlation techniques. 3D/3C velocity measurements (volumetric dimensions of 2.8'' x 1.9'' x 1.6'') of a turbulent boundary layer produced on the wall of a conventional wind tunnel are presented. [Preview Abstract] |
Monday, November 19, 2012 8:39AM - 8:52AM |
G14.00004: Autocorrelation based estimate of particle image density in PIV Scott Warner, Barton Smith, Pavlos Vlachos In Particle Image Velocimetry (PIV), the number of particle images per interrogation region, or particle image density, impacts the number of valid vectors and, especially in regions of shear, can impact the uncertainty of PIV. Therefore, any estimate of the uncertainty of PIV requires knowledge of particle image density. An auto-correlation-based method for estimating the local, instantaneous, particle image density is presented. The method is applied to synthetic images to provide an initial estimate of how the autocorrelation peak magnitude varies with known values of particle image density, particle image diameter, illumination intensity, interrogation region size, and background noise. From the synthetic image results, an empirical relationship is developed such that the particle image density is a function of the autocorrelation peak height, particle image diameter, illumination intensity, interrogation region size. Results are also obtained using images from two experimental setups with different seeding particles and flow mediums. The experimental results are compared to manual particle counts and are found to be robust. The effect of varying particle image intensities is also discussed and found to effect the particle image density. [Preview Abstract] |
Monday, November 19, 2012 8:52AM - 9:05AM |
G14.00005: PIV Uncertainty Roadmap Barton Smith, Steven Beresh, Pavlos Vlachos The error of PIV measurement has dozens of sources that can broadly be categorized as stemming from the particle motion, particle images, the processing algorithm, post processing, or hardware. Some cause bias errors while others cause random errors. In this talk, we will present our current thinking on the most important of these and how their impact can be assessed. While the theory of PIV is quite advanced, for most PIV uncertainty sources, such an assessment has yet to be performed. Hundreds of studies have focused on the accuracy of PIV, but these studies often cannot be generalized in such a way as to provide a 95{\%} confidence uncertainty band for every instantaneous vector in a PIV data set. Methods for assessment include use of synthetic images, image processing of real images, concomitant measurements, or deviation from best practices. Other assessment may be able to be automated and become part of the vector processing. In order for the PIV community to begin producing meaningful uncertainty estimates for PIV data, it is essential that uncertainty estimation for PIV becomes accessible and affordable. \newline [Preview Abstract] |
Monday, November 19, 2012 9:05AM - 9:18AM |
G14.00006: Method to Determine the Minimum Random Uncertainty in PIV Based on Real Images Kyle Jones, Barton Smith The noise floor, or minimum random uncertainty, of PIV based on the actual acquired PIV images can be determined by generating image pairs with known displacement. Image pairs are acquired with sufficiently small dt such that there is zero displacement between the images. A second image is then shifted by a prescribed amount. By computing a vector field based on these image pairs and calculating the standard deviation of the errors, a random uncertainty can be computed that incorporates the effects of camera noise, particle density, particle images, and displacement. The resultant values are helpful in determining whether the image quality is sufficient to achieve the desired uncertainty. A common PIV experimental setup with seeded water in a glass tank was used. The aperture of the camera lens was varied to achieve a range of particle image diameters. It was found that it is critical to filter the images prior to shifting in order to prevent smearing of the particle images. A Matlab code was written to shift the images by a prescribed, sub-pixel displacement, which were then imported into DaVis and correlated, resulting in displacement vector images. The random and bias errors of the DaVis and PRANA SCC algorithms are also compared for multiple sub-pixel displacements. [Preview Abstract] |
Monday, November 19, 2012 9:18AM - 9:31AM |
G14.00007: 3D reconstruction and velocity fields of a flame Jonathon Pendlebury, Dale Tree, Tadd Truscott We present three-dimensional internal velocity and shape measurements of an axi-symmetric partially premixed natural gas diffusion flame using Synthetic Aperture particle image velocimetry (SAPIV). This is the first step in demonstrating the technique for fully turbulent premixed flames. It has been shown that there is significant 3-dimensional motion in turbulent flames and knowing the internal fluid structures of flames is vital to understanding the flame properties. For example, fluid strain is directly related to flame extinction and recent measurements are showing strain is related to regions of soot formation and oxidation. Thus, the 3D strain velocity field and strain tensor are needed to both understand flame physics and validate flame theory and models. The current state of the art in flame imaging is Stereoscopic-PIV (2D-3C). Alternatively, SAPIV allows the complete time resolved 3D reconstruction of a measurement volume. The technique refocuses multi-camera viewpoints (8-10) into a refocused image at several depths within a scene, reconstructing a focal stack of images for each time step. [Preview Abstract] |
Monday, November 19, 2012 9:31AM - 9:44AM |
G14.00008: Microscopic Light Field Particle Image Velocimetry Tadd Truscott, Bryce McEwen, Jesse Belden This work presents the development and analysis of a system that combines the concepts of light field microscopy (Levoy 2006) and particle image velocimetry (PIV) to measure 3D velocities within a micro-volume. Flow at Reynolds numbers in the range of 0.02 to 0.03 was seeded with fluorescent particles and pumped through a micro-channel. The images were post processed to render a stack of 2D refocused images resulting in a 3D focal stack. Subsequently, a multi-pass, 3D PIV algorithm was used to measure channel velocities. Results from PIV analysis were compared with an analytical solution for fully developed cases, and with CFD simulations for developing flows. The relative error and advantages / disadvantages of this system will be presented. [Preview Abstract] |
Monday, November 19, 2012 9:44AM - 9:57AM |
G14.00009: Applying Tomographic PIV to Turbulent Taylor-Couette Flows Daniel Borrero-Echeverry, Donald Webster, Michael Schatz Over the years many techniques have been used to measure velocity fields in Taylor-Couette flows. However, these have been limited to measurements at discrete points (i.e., LDV or hotwire measurements) or planar sections of the flow (i.e., planar or stereo PIV). Tomographic PIV is a strong candidate for extending these measurements to three component volumetric velocity fields. Applying tomographic PIV to Taylor-Couette flows poses some serious challenges (curved interfaces, mechanical vibration, and moving surfaces). We discuss how these issues may be resolved and present temporally and spatially resolved measurements of the structures that form when finite-amplitude perturbations are applied to linearly stable Taylor-Couette flow and trigger the transition to turbulence. [Preview Abstract] |
Monday, November 19, 2012 9:57AM - 10:10AM |
G14.00010: Propagation of Instantaneously Varying Systematic and Random Uncertainties into the Measurement Mean, Variance, and Covariance Brandon M. Wilson, Barton L. Smith Particle Image Velocimetry (PIV), has numerous error sources, introducing non-linear, instantaneously varying systematic and random errors. Traditional uncertainty quantification (UQ) methods assume linear or constant uncertainty on the time-averaged quantity and may be inadequate for methods with this sort of error behavior. The Uncertainty Surface, and other methods can provide the instantaneous systematic and random errors of a PIV measurement. Equations to propagate these non-linear, instantaneously varying uncertainties into measurement statistics (e.g. mean, variance, and covariance) are derived using the Taylor series uncertainty equations and presented. The mean and variance uncertainty equation validity is verified using Monte-Carlo simulations for various instantaneously varying errors. The effects of the relative magnitude, asymmetry, and variance of random and systematic errors on the measurement uncertainty are demonstrated. The mean and variance uncertainty equations are demonstrated with actual PIV measurements for two experiments. Instantaneous uncertainties are estimated using the Uncertainty Surface method for four error sources: particle image density, diameter, and displacement and gradients. The first experiment consists of a rectangular jet with known error sources. Hot wire measurements are compared to the PIV measurements to assess the accuracy of the mean and variance uncertainties. These uncertainties are also demonstrated for flow through a confined bank of cylinders. [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