Bulletin of the American Physical Society
72nd Annual Meeting of the APS Division of Fluid Dynamics
Volume 64, Number 13
Saturday–Tuesday, November 23–26, 2019; Seattle, Washington
Session P11: Experimental Techniques: Quantitative Flow Visualization I |
Hide Abstracts |
Chair: Sam Grauer, Georgia Tech Room: 3B |
Monday, November 25, 2019 5:16PM - 5:29PM |
P11.00001: Three dimensional velocity and pressure measurements in a turbulent shear layer behind a backward-facing step. Karuna Agarwal, Omri Ram, Jin Wang, Yuhui Lu, Joseph Katz This study characterizes the unsteady pressure field generated by quasi-streamwise vortices that develop between the main spanwise vortices in the near field of a shear layer behind a backward facing step. Our objective is to understand cavitation inception, which occurs in the core of quasi-streamwise vortices located between 45 to 75{\%} of the reattachment length forming 1-2 mm wide and 5-7 mm long cavities. Tomographic imaging followed by 3D particle tracking using the Shake-the-Box method is used for calculating the instantaneous velocity and acceleration fields. These results are interpolated using singular value decomposition, and then refined using Constrained Cost Minimization to generate divergence free velocity and curl-free material acceleration at a spatial resolution of 250 \textmu m. The pressure is obtained by spatially integrating the material acceleration. The measurements are performed at Reynolds numbers based on separating boundary layer height Re$=$7100 and 17700. Analysis examines the effects of Re on the frequency, time evolution, size, strength and the pressure in the quasi streamwise vortices. The preferential location of pressure minima moves from the spanwise vortices at low Re to the quasi-streamwise structures at the higher Re. [Preview Abstract] |
Monday, November 25, 2019 5:29PM - 5:42PM |
P11.00002: 3D Particle Reconstruction of Volumetric Particle Image Velocimetry with Convolutional Neural Network Shaowu Pan, Qi Gao, Qijie Li, Hongping Wang, Runjie Wei, Jinjun Wang 3D particle reconstruction of volumetric Particle Image Velocimetry (PIV) is an under-determined inverse problem of which exact solution is difficult to obtain. Traditionally, approximated solutions can be obtained via optimization, e.g., multiplicative algebraic reconstruction technique (MART). Despite its popularity in recent years, the performance of MART-like algorithms deteriorates when the particle concentration becomes high, e.g., particle per pixel (ppp) $\approx$ 0.3. In this work, a particle reconstruction method based on convolutional neural network (CNN) is proposed. The method consists of two steps: first, an initial particle field is generated from a plurality of two-dimensional particle images get by camera using multiplied line-of-sight (MLOS) method; second, we use CNN to take the aforementioned initial particle field as input and output the 3D reconstructed particle fields. The data is artificially generated by randomly placing particle in 3D space then projected to four cameras to obtain 2D particle images. Compared to the traditional MART algorithm, the proposed method is not only significantly improving the accuracy of 3D particle reconstruction especially at high particle concentration but also eight times faster than the traditional MART algorithm. [Preview Abstract] |
Monday, November 25, 2019 5:42PM - 5:55PM |
P11.00003: A High-Spatial-Resolution Three-Dimensional Three-Component Velocimetry Method Based on Divergence-Free Polynomials Keishi Kumashiro, Adam Steinberg, Masayuki Yano We present a physics-constrained method of inferring three-dimensional three-component (3D3C) velocity fields in constant-density flows from reconstructed 3D Mie-scattering fields of tracer particles. The proposed method is based on the representation of the estimated velocity field as a linear combination of divergence-free polynomial basis functions; the piecewise constant representation of the estimated velocity field that is inherent to tomographic particle image velocimetry (T-PIV) is replaced by a smooth representation that automatically satisfies conservation of mass. The appropriate linear combination is determined using a non-regularized motion estimation framework that is influenced by optical flow estimation. We provide a detailed evaluation of the proposed method in terms of accuracy and spatial resolution, treating 3D constant-density DNS data as the ground truth. We show that the proposed method (a) yields significant improvements in accuracy and spatial resolution compared to an idealized implementation of T-PIV and (b) achieves comparable accuracy and spatial resolution to an idealized estimate that corresponds to the theoretical one-vector-per-particle limit. [Preview Abstract] |
Monday, November 25, 2019 5:55PM - 6:08PM |
P11.00004: ABSTRACT WITHDRAWN |
Monday, November 25, 2019 6:08PM - 6:21PM |
P11.00005: Direct background-oriented schlieren tomography Samuel Grauer, Adam Steinberg We present a novel approach to background-oriented schlieren (BOS) tomography that combines the deflection sensing and reconstruction algorithms. BOS imaging is a refraction-based flow visualization technique. Simultaneous BOS measurements from multiple cameras can be reconstructed by computed tomography to estimate the fluid's 3D refractive index field, which is post-processed to obtain local densities. Each camera is focused on a textured background pattern that is positioned behind the fluid. Density gradients cause distortions in the image; the deflected light trajectories are typically determined using an optical flow algorithm. These deflections constitute the projection data for reconstruction. Deflection sensing is itself a complex inverse problem and a primary source of error in BOS tomography. We propose an alternative measurement model for BOS tomography that incorporates the optical flow equation. The deflection model is extended to calculate image gradients, directly, such that the refractive index field is reconstructed from the distorted images. As a result, reconstructions must satisfy observed gradients instead of inferred deflections, which are prone to error. The talk describes our measurement model and presents a numerical assessment of direct BOS tomography. [Preview Abstract] |
Monday, November 25, 2019 6:21PM - 6:34PM |
P11.00006: Volumetric Velocimetry in the Rotating Frame of Reference using a Plenoptic Camera Abbishek Gururaj, Mahyar Moaven, Zu Puayen Tan, Brian Thurow, Vrishank Raghav Understanding the spatio-temporal behavior of flow separation on rotating surfaces is essential to characterize the performance of wind turbines, helicopters and more recently for drones. These require time-resolved measurements to quantify the unsteady flow fields during flow separation over rotor blades. Shortcomings in the use of conventional velocimetry techniques have limited researchers to comprehensively understand the flow over rotating wings. This study presented a novel concept for the development of a rotating frame of reference flow field measurement technique. A single plenoptic camera mounted coaxially above a hub mounted mirror that rotates with the rotor enables instantaneous and three-dimensional flow field measurements over a wide range of azimuth angles. The design and implementation of this concept as an experimental test facility to conduct flow field and force measurements over a rotating wing is discussed. In the presentation, an overview of the methodology along with the challenges, details of the test facilities and some preliminary flow field and force measurement results will be discussed. [Preview Abstract] |
Monday, November 25, 2019 6:34PM - 6:47PM |
P11.00007: Application of Plenoptic Camera 3D PIV in a Rotating Frame of Reference Mahyar Moaven, Abbishek Gururaj, Zu Puayen Tan, Brian Thurow, Vrishank Raghav A major limitation to the thorough understanding of a rotating flow's aerodynamic characteristics is the difficulty in measuring data within the rotating frame of reference. Rotating 3D Velocimetry (R3DV) is a concept designed to acquire volumetric flow field measurements over a rotor, within the rotating frame of reference, using a single camera. A submerged wing rotates in conjunction with a mirror that reflects light traveling through flow over the wing to a stationary plenoptic camera. The plenoptic camera's added array of microlenses between the aperture and image sensor gives it the ability to capture the incident angle of light rays, allowing for 3D reconstruction of volumes. Traditionally, plenoptic camera calibration has been explored for a stationary field of view. In current development is a method to incorporate rotation into the calibration such that images from all azimuth angles can be calibrated by interpolation, thereby eliminating the need to calibrate each distinct angle. Other challenges include achieving sufficient depth resolution and obtaining an adequate depth of field at a distant focal plane. Preliminary results demonstrating the efficacy of the R3DV method will be presented along with explanations of how the aforementioned problems will be overcome. [Preview Abstract] |
Monday, November 25, 2019 6:47PM - 7:00PM |
P11.00008: Technique for Characterization of 3D Unsteady Fluid-Structure Interactions via a Single Plenoptic Camera Brian Thurow, Zu Tan, Bipin Tiwari, Vrishank Raghav Due to flexible boundaries, many flows are unsteady, three-dimensional and subjected to fluid-structure interactions (FSI). Conventional studies of FSI nominally treat the flow and boundary surface measurements separately, resulting in a loss of true FSI physics, especially for aperiodic flows. More recently, simultaneous multi-camera approaches using tomographic-PIV/3D-PTV and Digital Image Correlation (DIC) have been explored. However, this approach is cumbersome to arrange in confined spaces, and challenging to align with each camera's shallow depth-of-field. Here, we propose an alternative method of simultaneous FSI measurement using a single plenoptic camera. Equipped with a microlens array, plenoptic cameras capture a volume's light-field in 4D, which enables single-camera 3D measurements. In this study, a newly developed \textit{kHz}-rate plenoptic camera is applied to characterize the FSI of a flexible 50x 30 x 30\textit{mm}$^{3}$ tube model. Macroscopic painted dots were used to track surface motion, while micron-sized particles were used as the flow tracer. A pulsatile flow was applied to intermittently collapse the tube, during which the dots and seeds were simultaneously imaged. In post-processing, dots and seeds are separated by diameters and/or shape, after which their 3D trajectories are tracked separately via a plenoptic-PTV algorithm. Preliminary proof-of-concept results are presented, along with further analyses to optimize particles/dots sizes and densities. [Preview Abstract] |
Monday, November 25, 2019 7:00PM - 7:13PM |
P11.00009: Refraction correction for 3D optical measurements inside cylinders Tommaso Astarita, Gerardo Paolillo The present work proposes a calibration camera model to precisely express the projection from 3D world to 2D image coordinates for measurements inside transparent cylinders. Snell's law is used to model the refraction of the optical rays at the external and internal surfaces of the cylindrical wall and a standard pinhole camera is used to describe the lens behavior. Making use of the perspective and refraction laws, the mapping function consists of a relatively small number of parameters and all of these have a clear geometrical or physical meaning. A calibration procedure for the proposed camera model is also outlined. Finally, the innovative model is comparatively assessed against the classical pinhole camera model and different polynomial-based models by using experimental data from an investigation of Rayleigh-B\'{e}nard convection. Although the model is tested for tomographic particle image velocimetry measurements in ?ows occurring inside cylinders, it can be used also in a variety of other imaging applications, as well. [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