입자 이미지 Velocimetry 에코

The goal of this experiment is to acquire two dimensional, instantaneous fields of velocity in Hagen Zoi flow, also known as laminar pipe flow. Using echo particle image LOC symmetry or E-P-I-V-E-P-I-V validation measurements are demonstrated in a recirculating pipe flow of a 50 50 mixture of water and glycerin. The pipe flow system used here employs a constant pressure head maintained by an aquarium pump to drive the flow.

A phased array ultrasound probe is mounted to the pipe wall of the flow system and B mode ultrasound images are streamed. The fluid is seated with hollow glass spheres or tracer particles that faithfully follow the motions of the flow. Ultrasound images are then acquired, transferred to a PC and converted to an image format compatible with the commercial particle image veloc symmetry software.

Cross correlation algorithms are applied to successive ultrasound B mode images to compute two dimensional fields of velocity. Ultimately, the vector fields are analyzed to compute flow quantities of interest, such as ensemble average sheer stress, and vorticity. Though this technique can provide insight into basic fluid dynamics, it can be and is often used in practical flow systems, including biomedical applications, for example, arterial or interventricular flows.

Our ongoing experiments using liquified biomass and fluid is what interested us in this technique. To set up the EP IV system, start by powering on the pumps. This will start the recirculating pipe flow at a constant rate.

Next, apply a water-based topical gel to the ultrasound probe. The gel minimizes loss of transmission of the ultrasound beam between the probe face and the pipe. Then using a specially constructed probe mount with through wall pipe fittings, attach the ultrasound probe to the exterior pipe wall power on the ultrasound machine.

Once all systems load, a live stream of images will appear on the ultrasound screen. 2D mode is the default setting for the linear probe to acquire EPIV measurements. First, measure out the appropriate dry weight of nominal 10 micrometer hollow glass spheres so that their final concentration when added to the flow system is about 17 weight parts per million.

Next, extract a volume of fluid from the basin and add the particles to the fluid to form a concentrated particle solution. The particles when added to the flow system will serve as ultrasound contrast agents or tracer particles. Add the concentrated particle solution to the recirculating pipe flow system by stirring it into the water basins.

The glass spheres can then be viewed on the ultrasound monitor. After a few minutes, the glass spheres will be evenly distributed throughout the system, So one of the most difficult parts of this procedure is obtaining clear high resolution ultrasound images. In order to maximize the quality of these images, we adjust the gain focal points and dynamic range based on apriori estimates of the flow velocity.

We further optimize these parameters on the fly by analysis of the ultrasound images. Use the depth control knob on the control panel of the ultrasound to set the image depth to three centimeters. Next, using the 2D gain knob, adjust the total image gain to increase the brightness of the image such that the seed particles are clearly visible in the instrument panel.

Adjust the time gain compensation sliders to attenuate scatter from the pipe walls and to compensate for depth related attenuation of the ultrasound signal. This will remove excess image at the top and bottom of the pipe walls in 2D mode. The knobs on top of the control panel from left to right correspond to width, focus frequency and frame rate.

Use these knobs to adjust the image further to achieve the highest possible physical resolution, frequency and frame rate for analysis. Then adjust the probe operating frequency to 10 megahertz and set the frame rate as 49.5 frames per second. Note that these four parameters are inherently coupled.

Consequently, for a given ultrasound image scan, there is a trade-off between spatial and temporal resolution. Due to limited lateral resolution, the glass spheres will be smeared in the lateral direction and appear as ellipsoids in the image. Once the parameters are optimized, it is time to collect data on the ultrasound control panel on the instrument.

Press the new exam button to start a new experiment. Under patient. Enter pipe flow as the last name and the date as the first name.

Enter the test number in the patient ID field. The ultrasound scan will then begin automatically when the preset maximum of 1000 to 1, 500 images is reached. A new scan loop begins as scanning continues.

Make adjustments to the imaging parameters until the seed particle is sharply in focus with about 10 particles per interrogation area. To restart the scanning recording loop, press the freeze button on the ultrasound control panel. Once a sufficient number of ideal images have been acquired, press the freeze button.

Next, press the C loop button on the ultrasound control panel. Select all images to include all ultrasound images in the analysis set. Once the images for analysis have been selected, press the image store button to save the selected set of ultrasound images.

Once the images have been saved, press the archive button on the ultrasound control panel. When prompted, select the syn loop of interest from the small window to be saved to the local hard drive. Then use the mouse cursor to select end exam.

Press the archive button and use the mouse cursor to first select more and then select disc management. This will transfer the saved cyl loop or cyl loops to the PC running the particle image velocity symmetry or PIV software. Once the images are captured and saved, the ultrasound image must be converted from a digital imaging communications in medicine or DICOM file to a joint photographic experts group or JPEG picture file for analysis.

Use a MATLAB script running DICOM to jpeg dot m to convert the DICOM files to JPEGs. This script was made in-house and can be obtained for educational purposes from the web address shown here. Once the file has been converted, open Davi software from Law Vision in the software.

Double click on the davi icon, select new project, then select PIV. In the toolbar, select import images and choose import via numbered files. Then in the pull down menu, locate the folder where the JPEG ultrasound images are stored and double click on the first image of the set.

This will import all ultrasound images in this numbered set to define a region of interest for analysis that includes only the fluid. Apply a mask to create the mask, enter the coordinates, a rectangular area using two x and y coordinate points based on the information from the DICOM file and knowledge of the pixel dimensions. Next in the main control panel in dvu, click on the tab located under current project containing the imported images.

Select the table labeled batch processing. This enables the vector processing window of Davi for batch processing from the operations list using the PIV plus PIV time series tree select vector calculation parameters and choose multipass with a decreasing interrogation size of 64 pixels by 64 pixels to 12 pixels by 12 pixels with an overlap of 50%Set the relative vector range restriction to all and then absolute vector range restriction to five pixels. Then apply a median filter to suppress noise and smooth the vector fields.

Next for vector processing, check the box data range equals use masked area in the vector calculation parameter menu. Note that optimal selection of vector calculation parameters is dependent on flow geometry, flow properties, image resolution, tracer particle density, and desired quantitative flow analysis. Once all of the desired parameters have been set at the left of the batch processing screen, select the total number of images to be processed.

Click on start processing. This will compute the displacement field between successive ultrasound images using cross correlation algorithms. To analyze the processed data, export the UCV vector fields from DAVO as txt files.

To do this under the JPEG image branch in the project screen, select the vector displacement branch. In the toolbar, select the export tab. Select file type.

Ask e txt. Choose create an export folder and select export. Next, open the file in MATLAB by running the MATLAB script.

The exported vector fields are named B-X-X-X-X-X do TXT where X represents a number in increasing from one to 99, 999. Each file contains four data columns, which can be viewed using notepad one x location of the vector in the image two Y location of the vector in the image, three X component of displacement or stream wise displacement, and four y component of displacement, which describes the wall. Normal displacement use to compute the velocity vector field U as a function of X and Y, where X and Y correspond to spatial coordinates in the ultrasound image, this is done by first converting the displacement field D of X and Y measured in pixels to a displacement field measured in meters using the image scaling parameter M given in units of meter over pixel.

Next, the displacement field is divided by the sweep corrected temporal separation between images delta T, where delta T is equal to one over the frame rate given in frames per second, plus the displacement field divided by the time it takes for the ultrasound image to sweep across the image width. In summary, U of X and Y equals M times D of X and Y divided by delta T.DICOM inherently stores a file structure that provides the information needed to compute the image scaling parameter M and sweep corrected temporal separation. Delta T in the present study M equals 77 microns per pixel FPS equals 49.5 and B equals 25, 047 pixels per second.

Finally, compute ensemble average velocity vector fields while normal profiles of mean velocity and any other flow quantities of interest. To demonstrate the usefulness of EPIV and to assess measurement uncertainty, two dimensional, instantaneous fields of velocity and laminar pipe flow were acquired as described in this video, this instantaneous vector plot shows velocity vectors every fourth column, and the background color contour map corresponds to velocity magnitude. The two dimensional spatial location of the velocity vector is denoted by D over D and X over D, where D is the radial position measured from the upper wall.

D is the pipe diameter, and X is the stream wise position measured from the pipe entrance. The apparent parabolic shape of the velocity vectors along columns indicates that the measurements are consistent with the expected velocity profile for pipe flow. The ensemble averaged vector plot computed by averaging over 1000 instantaneous vector plots provides a representation of the mean velocity field.

It will also average out random noise errors in the instantaneous vector fields. The velocity vectors are primarily in the stream wise direction. The largest velocities occur at the pipe center line.

The velocities decreased to zero at the pipe walls and the flow is roughly symmetric the mean stream wise velocity profile along the pipe radius obtained by averaging the ensemble averaged vector plot along the rows in the horizontal direction is shown here. Also shown is the expected mean velocity profile for laminar pipe flow. Given the experimental conditions.

The agreement between the EPIV measurements and the expected higgin pozo profile is best near the pipe center line and worst near the pipe walls. The large differences near the wall are likely due to strong reflection and refraction of the ultrasound waves at the curved surface of the pipe wall that produce high image intensities in these regions, the high intensities at the wall obscure particle intensities leading to measurement error. With the development of this technique, researchers studying fluid dynamics in engineered or biological flow systems can now acquire spatial temporal variations of the velocity field in optically opaque fluids or through optically opaque geometries.

After watching this video, you should have a good understanding of how EPIV works, its limitations and how to build and operate an EPIV system using a commercial ultrasound machine.