High-speed Particle Image Velocimetry Near Surfaces

The overall goal of this video is to demonstrate how to set up and operate a benchtop particle image veloc symmetry experiment. First, estimate the recording parameters needed to resolve the velocities and gradients expected in the flow. The second step is to adjust the seating density to achieve eight to 10 particles per interrogation window.

Next, adjust the camera frame rate and the time between laser pulses To achieve a particle shift between image pairs of less than one quarter of the interrogation window size, the final step is to record data and process the results with appropriate processing algorithms. Ultimately, particle image vela symmetry may be used to study the structure and dynamics of any flow with sufficient optical access. The main advantages of this technique over existing methods of flow measurement like hot wire anemometry or PTO tubes, are that it is non intrusive and multidimensional.

This method can help answer key questions about the velocity flow field in many applications such as the structure and dynamics of transient boundary layer flow formed in an internal combustion engine. These experiments use lasers, so be sure to review laser safety and obtain the correct safety equipment before proceeding. Every experimenter should wear appropriate safety goggles.

Hang laser safety drapes. To isolate the laser experiment, use a warning light to alert others when a laser is in operation. As a first step, determine an appropriate camera lens, a long distance microscope lens for this setup.

In addition, determine the camera frame rate and time delay based on practical guidelines for particle image V symmetry. In this video flow near the surface of a plate will be studied. The frame rate and time delay are based on estimates of flow velocities in the near wall region.

Ensure the laser is level and the beam forming optics are in place. In this demonstration, light is scattered from silicone oil droplets created. Using an oil atomizer, turn on the flow and adjust the seating density using the atomizer jets and the bypass valves on the atomizer.

Begin optimization by ensuring the high speed controller and software are properly configured. Set the camera to continuously record. Turn on the laser and the atomizer.

Focus the long distance microscope to clearly image the particles. Also, make sure the intensity of the particle images is not saturating the camera. Once focused particle images are achieved, turn off the continuous acquisition mode.

Record several hundred images of the flow when finished, check the recorded images for focus appropriate particle seating and particle drift of around eight pixels between images. Here is an example of sparse seeding and a particle drift of about one pixel in a 32 by 32 pixel window to correct particle seeding density. Increase or decrease the number of atomizer jets to adjust the drift between images.

Increase or decrease the time interval between the laser pulses. This example is close to ideal with eight to 10 particles per 32 by 32 pixel interrogation window, and the particles shift less than eight pixels. If it is difficult to track groups of particles through a series of images, as with these images, there may be too much out of plane motion.

Adjust the camera offset from the focal point or increase its working distance from the light sheet to compensate for too much out of plain motion. Finally, after all adjustments again, record several hundred images and check the particle. Shift sating density and focus.

Repeat adjustments until the criteria are met. Before each run of the experiment, perform a camera intensity calibration with the cap on the camera assembly to set a reference for intensity. After calibration is finished, remove the cap.

Set the laser to the repetition rate and current that was determined as optimal. Before switching the laser to external mode, make sure the laser receives a continuous trigger signal that matches the set frequency. Turn the laser on record.

A sequence of background images of just the light sheet grazing the surface of the plate. There is no significant background scattering and therefore most of the images dark. Save these images for later processing.

Now start the flow and allow it to stabilize with the camera set to continuously acquire images, verify that it is collecting focused particle images. Once this is done, turn off the continuous acquisition mode. For each run, enter the desired number of images to be collected.

Then press record. Once the recording is finished, turn off the flow and the laser. Review the sequence of images to check particle drift seating density and particle image focus.

If satisfied, save the recording. Perform as many runs as needed to guarantee proper setup or for statistics. Each time the camera assembly or focus is changed, take images to help determine the location of the surface in later analysis.

First, increase the exposure time of the camera. Next, set a calibration target on the light sheet plane and make sure it is in contact with the plate. Illuminate the target from behind with a light source with the camera continuously recording images.

Adjust the target so the recorded image is in focus and not distorted. Ensure the contact point between the plate and the target is visible in the image. This is crucial for determining the location of the plate in the image record 10 images of the calibration target Data processing is done in software.

First average each set of 10 calibration. Target images use the resulting average image in the calibration dialogue. This will determine the true world dimensions of the acquired images.

Apply each calibration image to the corresponding set of images from this. Determine the location of the plate in the calibrated images. Next, average the background images.

Compare the intensity counts of the average background image to those of the seating particles to determine if laser reflection from the surface contributes significantly to the background noise. In this experiment, reflections did not contribute significantly to the background. Pre-process the calibrated flow images using a high pass filter to remove large intensity fluctuations in the background, this will not affect the particle signals in this case.

Now define a region in which the particle image of all asymmetry analysis is to take place in this software, a geometric mask is used. Next, choose a procedure to calculate the vectors. The software option here does two initial passes with 64 by 64 square pixel interrogation windows with 50%overlap.

This is followed by three passes using 32 by 32 square pixel interrogation windows with 50%overlap in software. Apply subroutines to remove velocity vectors that do not satisfy criteria established to ensure the quality of cross correlation results. These sample data showing red the velocity vector field near the surface of a plate as a function of distance along the surface and height above the surface.

Immediately after calculation. First, apply a subroutine to redefine the region for analysis, the subroutines first determine the peak ratio, which compares the highest correlation peaks to the common correlation. Background vectors with peak ratios near one indicate a false peak and are removed reflected here by the white boxes and missing arrows.

Next, the routines reject vectors that have a component that is not within one root mean square of the median for that component as calculated from a subset of vectors, the routines also eliminate spurious vectors that arise from a large overlap in the velocity vector calculation. As a last step, the routines fill up empty spaces in the field by interpolating between non-zero neighboring vectors. These vectors are shown in magenta.

As a final step, assess the quality of the results flow near the surface of a flat plate was measured and data on the instantaneous velocity and fluctuation velocity fields were obtained in this plot. A subset of the instantaneous velocity field at time T equals 0.1 milliseconds is shown at left and the fluctuation velocity field is at right note. The vectors provided for scale at the upper left of each plot at a later time, step in the middle of recording sequence, the instantaneous velocity field on the left remains stable, but the fluctuation field has changed dramatically.

The magnitude of the vectors has decreased and there is less uniformity in direction near the end of the recording sequence. The instantaneous velocity field still shows the overall left to right direction of the flow. The fluctuation field has reversed direction.

This figure shows the horizontal velocity profile at different times during the flow. The error bars shown on the T equals 0.1 milliseconds profile are representative of the error bars for all other times. Note, the time history shows a decrease in flow over time.

The time averaged profile is shown with circles. After watching this video, you should have a good understanding of how to adjust and optimize particle image os symmetry parameters to adapt the procedure to any flow investigation. Don't forget that working with lasers can be extremely hazardous and precautions such as wearing protective eyewear and using warning signs to alert others should always be taken while performing this procedure.