Assembling and testing of PIV system

Chih-Hsiang Tsao

Master's thesis

Abstract:

In this paper, the main experiment is done by stereoscopic PIV and the result is analysed, including the formation of vortices that are observed, and the reconstruction images using Proper Orthogonal Decomposition (POD) method. Moreover, the planned future research is shortly described in the last chapter. Appeared around 35 years ago, Particle Image Velocimetry (PIV) has been an essential measurement technique in fluid mechanics. Considerable efforts have been made to determine instantaneous velocity vector fields in a plane of the flow, which is the major asset of PIV. This technique measures the displacement of particles which are assumed to flow with the local flow within a given period of time. For evaluating the recorded images, the digital image is divided into smaller areas called "interrogation windows", sometimes overlapping between neighbouring areas, on which the cross-correlation function will be applied. However, the performance is greatly influenced by the preparation work in advance. Generally, the calibration plays an important role in this part. It requires taking several images of the calibration target within few parallel planes. According to these images, the "third axis" is computed by mapping function. By all means, the placement of calibration target should be aligned with the light sheet to minimize the possible error. The evolution of PIV is still on-going as to capture dynamic velocity flow fields more precisely. Thanks to the progress in laser technologies and electronic imaging systems, the development of PIV technique is pushed further into next stage where measurement of velocity fields in three dimensions can be conducted. These 3D techniques include: holographic PIV, dual plane PIV, tomographic PIV. Tomographic PIV is built based on the concept lying in stereoscopic PIV; thus, the comparison between them is elaborated. The idea in calibration is similar, yet in 3D PIV polynomial calibration is used. Tomographic PIV is of course superior to stereoscopic in providing fully volumetric and truly instantaneous images and less sensitive to background noise. The entire targeted volume is focused by equipping a relatively small lens aperture under the Scheimpflug condition. In evaluation, the images recorded with two laser pulses are divided into small areas called "interrogation volume", on which 3D cross-correlation is applied. Also, 2C PIV is more sensitive to background noise and contamination on optical components such as dirt on vessel because of the lack of recoding cameras which can be used to compensate the image distortions. However, the drawbacks of tomographic PIV are the occurrence of ghost particles and huge amounts of data that require large storage capacity. Thus, some algorithms are proposed to "digest" the enormous amounts of data. The "digestion" way varies in different producers for instance Dantec, LaVision, and TSI. These producers emphasized on different features. The tomographic PIV is produced by LaVision, so the DaVis software will be introduced in detailed. On realizing the fundamental concept in PIV, the future work will be continued right after the delivery of tomographic PIV.

Abstract:

In this paper, the main experiment is done by stereoscopic PIV and the result is analysed, including the formation of vortices that are observed, and the reconstruction images using Proper Orthogonal Decomposition (POD) method. Moreover, the planned future research is shortly described in the last chapter. Appeared around 35 years ago, Particle Image Velocimetry (PIV) has been an essential measurement technique in fluid mechanics. Considerable efforts have been made to determine instantaneous velocity vector fields in a plane of the flow, which is the major asset of PIV. This technique measures the displacement of particles which are assumed to flow with the local flow within a given period of time. For evaluating the recorded images, the digital image is divided into smaller areas called "interrogation windows", sometimes overlapping between neighbouring areas, on which the cross-correlation function will be applied. However, the performance is greatly influenced by the preparation work in advance. Generally, the calibration plays an important role in this part. It requires taking several images of the calibration target within few parallel planes. According to these images, the "third axis" is computed by mapping function. By all means, the placement of calibration target should be aligned with the light sheet to minimize the possible error. The evolution of PIV is still on-going as to capture dynamic velocity flow fields more precisely. Thanks to the progress in laser technologies and electronic imaging systems, the development of PIV technique is pushed further into next stage where measurement of velocity fields in three dimensions can be conducted. These 3D techniques include: holographic PIV, dual plane PIV, tomographic PIV. Tomographic PIV is built based on the concept lying in stereoscopic PIV; thus, the comparison between them is elaborated. The idea in calibration is similar, yet in 3D PIV polynomial calibration is used. Tomographic PIV is of course superior to stereoscopic in providing fully volumetric and truly instantaneous images and less sensitive to background noise. The entire targeted volume is focused by equipping a relatively small lens aperture under the Scheimpflug condition. In evaluation, the images recorded with two laser pulses are divided into small areas called "interrogation volume", on which 3D cross-correlation is applied. Also, 2C PIV is more sensitive to background noise and contamination on optical components such as dirt on vessel because of the lack of recoding cameras which can be used to compensate the image distortions. However, the drawbacks of tomographic PIV are the occurrence of ghost particles and huge amounts of data that require large storage capacity. Thus, some algorithms are proposed to "digest" the enormous amounts of data. The "digestion" way varies in different producers for instance Dantec, LaVision, and TSI. These producers emphasized on different features. The tomographic PIV is produced by LaVision, so the DaVis software will be introduced in detailed. On realizing the fundamental concept in PIV, the future work will be continued right after the delivery of tomographic PIV.

Keywords

stereoscopic PIV, tomographic PIV

Language used: English

Date on which the thesis was submitted / produced: 21. 5. 2018

Thesis defence

• Supervisor: Ing. Petra Dančová, Ph.D.

Technical University of Liberec

Faculty of Mechanical Engineering

Master programme / field:
Mechanical Engineering / Machines and Equipment Design