Development of a scalable real-time Lagrangian particle tracking system for volumetric flow field characterization

Barker, Douglas

Permalink

https://hdl.handle.net/2142/42395 Copy

Description

Title

Development of a scalable real-time Lagrangian particle tracking system for volumetric flow field characterization

Author(s)

Barker, Douglas

Issue Date

2013-02-03T19:37:17Z

Director of Research (if dissertation) or Advisor (if thesis)

Zhang, Yuanhui

Doctoral Committee Chair(s)

Zhang, Yuanhui

Committee Member(s)

• Gates, Richard S.
• Jacobi, Anthony M.
• Thomas, Brian G.
• Sun, Yigang

Department of Study

Engineering Administration

Discipline

Agricultural & Biological Engr

Degree Granting Institution

University of Illinois at Urbana-Champaign

Degree Name

Ph.D.

Degree Level

Dissertation

Keyword(s)

• Lagrangian particle tracking
• experimental fluid mechanics
• particle tracking velocimetry
• parallel processing
• real-time processing
• flow field visualization

Abstract

Lagrangian particle tracking (LPT) is important for the study of turbulence and the inertial behavior of particles in natural and industrial flow fields. This work describes the development of a new scalable real-time LPT system which is able to track particles in large scale 3D turbulent flow fields. A statistical accumulator grid concept was developed to attribute the measured Lagrangian velocities and accelerations of particles to a Cartesian framework for comparison with Eularian based measurements and simulations. Real-time processing was achieved through development of parallel frameworks based on fine and coarse grain problem decomposition for heterogeneous computing architectures. The first framework is based on exposing task and data parallelism through a streaming pipeline within each multi-core processor node. The second framework is based on pipelining temporal data through a unique message passing algorithm in order to utilize large clusters containing hundreds of processor nodes. A sensitivity analysis was completed based on the derivation of measurement uncertainty. It was shown that by utilizing groups of four cameras in the 3D reconstruction process, the overall sensitivity to camera location, image noise and propagated uncertainty could be reduced significantly. A six camera prototype system was developed and an experimental analysis was conducted to assess the uncertainty in the 3D position, velocity and acceleration measurements of observed particles. The 3D position combined standard uncertainty was 0.16 mm and accuracy was comparable to a caliper in measuring distances between static particles. The velocity measurements were shown to be less than 1% of the calculated value for an object rotating with constant angular velocity. Acceleration was accurate to within 1% for low frame rates but diverged from the calculated value at higher frame rates. The system was used to characterize the dynamic motion of neutrally buoyant helium filled soap bubbles in an unconfined round turbulent jet. The results for axial velocity decay and transverse velocity profile all matched well with widely accepted models. In addition, the profiles of Reynolds shear stress and axial turbulence intensity were in good agreement in both profile and magnitude of those found in literature. The validated LPT system was then applied to a turbulent forced air vortex and particles were successfully tracked with complex 3D paths.

Graduation Semester

2012-12

Permalink

http://hdl.handle.net/2142/42395

Copyright and License Information

Copyright 2012 Douglas Barker

Owning Collections