4PIV FundamentalsPIV measures are taken very quickly (milliseconds in total)At time t1Pulsed laser sheet illuminates a planar region of the flowParticles are imaged on the camera (Frame A)At time t1 + DtA second image (Frame B) is taken of a second light sheetStatistical (Cross-Correlation) methods are used to determine the particle displacement over the time Dt, and thus the local velocity
8Measurement at one point over a period of time PIVLDVMeasurement at one point over a period of timeProvides time history of the flow and hence time averaged statistics at one point. Flow field mapped by traversing the measuring pointVelocity obtained by measuring time to travel a known distanceMeasurement at many points at one instant of timeProvides instantaneous vector fields. Time averaged statistics obtained by averaging several image fieldsVelocity obtained by measuring image displacement in a known timeLDV provides the time history of the flow (velocity components) at the measurement location. The measuring region or the measuring volume is the intersection region of the laser beams.In PIV, measuring the displacement of particle images in the time between two laser flashes (multiple flashes can also be used) that illuminate a plane (in the flow field) provides the velocity information.While the measurement volume size determines the spatial resolution in LDV, the maximum image displacement gives a measure of spatial resolution in PIV. Seeding (adding particles that follow the flow) requirements are comparable for both techniques.
9Cross CorrelationAll particles look alike, so it is hard to find the ‘same’ particle in both Frame A and B (PTV)Instead, PIV uses a statistical approach to find the most likely displacement of a group of particlesFrame A is broken up into a grid of ‘interrogation regions’The group of particles in the interrogation region creates a unique ‘fingerprint’ that we can look for in both frames (PIV)
10Cross Correlation Fame B Search Area (“Spot B”)The relative location of the interrogation region in Frame A is knownA search area is defined in Frame B.Frame A InterrogationRegion(“Spot A”)
11Original Particle Positions in Frame A Cross CorrelationOriginal Particle Positions in Frame A
12Location of particles in Frame B Cross CorrelationLocation of particles in Frame B
13Cross CorrelationA ‘Spot Mask’ (above) created from Frame A, is scanned across the search area of Frame B to form a ‘correlation map’
14Cross CorrelationThe correlation map will have a peak relative to the location where the ‘fingerprint’ of the Spot Mask is identified in Frame B
15Cross CorrelationThe correlation values for different locations of the spot mask can be represented on a “correlation map”In the ideal case, there is a distinct, single, round peak in the correlation map
16Cross Correlation DY DX The X and Y displacements of the particles are determined by the offset of the interrogation regionsDX
17Cross CorrelationSince we know the time separation (Dt) between the 2 laser pulses very accurately, we can measure the ‘group’ velocity as the displacement / Dt, and assign a single velocity vector
18Cross Correlation A) B) NOTE: Particles and Interrogation region are typically larger than represented here in this simplified exampleFrame ASpot Mask
19Cross Correlation A) B) (1*1) (1*0) + (1*0) 1 Cross Correlation Map Displacement (x,y) = (0,0)1
20Cross Correlation A) B) (1*0) + (1*0) Cross Correlation Map Cross Correlation MapDisplacement (x,y) = (0,1)1
21Cross Correlation A) B) (1*1) (1*0) + (1*0) 2 2 Cross Correlation Map Displacement (x,y) = (0,2)1
22Cross Correlation A) B) (1*0) + (1*0) 2 Cross Correlation Map 2Cross Correlation MapDisplacement (x,y) = (1,2)1
23Cross Correlation A) B) (1*1) + (1*1) 5 2 5 Cross Correlation Map Displacement (x,y) = (2,2)1
24Cross Correlation A) B) Cross Correlation Map Displacement (x,y) = (2,2)Peak in the Correlation Mapis at (+2,+2), indicating that theparticles moved in this directionfrom Frame A) to Frame B).
25Processing Crosscorrelation Spot B Interrogation region Spot A Frame A particledisplacementCrosscorrelationFrame BVector field
26Cross CorrelationThe process is repeated for each interrogation region in Frame A, resulting in a 2-Dimensional velocity field for the imaged region
27Hardware Principles 2-component System 3-component System PIV component technology is continuously evolving and it is a challenge for both the users as well as manufacturers to keep pace with the latest developments. The user is particularly concerned that the PIV system acquired today is not obsolete tomorrow. TSI has factored this need into our PowerView PIV system design and the result is a very versatile and flexible system that can incorporate future developments with minimum cost.TSI’s philosophy has always centered around systems and processing approaches that can extract maximum information from the experiment. In PIV this would require that the raw image information is never lost and is available along with the corresponding vector file. The benefits of such a data management approach are numerous. Having the raw image would mean never having to repeat an experiment. The same image can be processed in several different ways to obtain detailed flow information not extracted during preliminary analysis. Comparing the vector field to the corresponding particle image field offers insight into the nature of the flow, on-line validation and enhances the confidence level in the velocity vector data. Availability of the raw image field offers the ability to get detailed information about the scatterers such as particle image size, concentration etc.The flexible approach offered by the TSI’s PowerView PIV system ensures that the system will grow as camera, interface and processing technologies evolve.2-component System3-component System
28System ComponentsImaging Subsystem (Laser, Beam delivery system, light optics)Illuminate a plane in the flow (seeded) using a pulsed laserPulse energy, duration, and repetition rateTypically Nd:YAG laser operating at 532 nmImage Capture Subsystem (CCD Camera, Camera Interface, Synchronizer-Master control unit)Master Timing devise triggers illumination and camera capturingCamera captures particle images and records themAnalysis and Display SubsystemCalculates and displays a two dimensional vector field from the particle image fieldsCapable of processing higher-order statistics of flow fieldA complete PIV system can be viewed as a combination of three sub-systems. The first and second subsystems (in the list above) are responsible for obtaining good particle image fields, and the third for processing and displaying corresponding velocity fields and their derivatives. While imaging involves on-line set up and optimization, analysis/display can be done on-line or off-line.Laser has become the choice illumination device since it can produce high energy short duration pulses. The laser pulse is delivered to the measurement location using flexible beam guides with mirrors or through optical fibers. Optical elements can convert these pulsed laser beams into pulsing light sheets. CCD cameras are used to record the particle image field. A master control system referred to as a Synchronizer system is used as the timing device to generate the control signals and synchronize the camera with the laser pulses and the camera interface.Camera interface is used to ensure very fast data transfer from the camera to the computer. Image acquisition/analysis/display software is the umbrella program that allows the user to set and optimize the hardware and the experiment, acquire images, process them and display velocity fields.
29Nd:YAG Laser 15 mJ - 400 mJ per pulse 4 ns - 20 ns pulse duration Acts like strobe lightfreezes the particle imagesWide range of Tto measure flow velocities from mm/s to supersonic speedsHz Pulse Repetition Rate532 nm Wavelength (Frequency Doubled)Most visible wavelengthThe dual Nd:YAG laser is the preferred light source for most PIV experiments. This integrated laser houses two separate laser heads whose pulses are combined with built-in beam combination optics to produce a coaxial pulse train. The very short pulse duration is fast enough to “freeze” even particle images in supersonic flow without image streaking. The high pulse energy is able to illuminate small particles for either air or water flows. Because two lasers are used any time between pulses is possible, with the full energy in each pulse.Lower power (10 to 100 mJ ) Nd:YAG lasers are now available for water flow experiments or small air flow experiments. These small Nd:YAG lasers are able to deliver much higher pulse energy than even a 10 Watt argon laser.Dual Nd:YAG s can come with variable repetition rate from Hz. A dual YAG laser with 15 Hz rep rate per laser works well with the special 30 Hz cross correlation camera, when two images are captured on separate frames. This would give 15 velocity fields per second.
30Energy vs. Q-switch delay HiMedLow100%80%60%Pulse energy - % of maximum40%20%Nd:YAG lasers use two trigger signals in generating a pulse of light. The first trigger fires a flash lamp. Photons from the flash lamp are absorbed by the Nd:YAG rod where this energy is stored. The second trigger opens the Q-Switch releasing the stored energy as a laser pulse. The light discharge from the flash lamp takes some time to reach maximum brightness and then decreases. The pulse energy depends on flash lamp brightness when the Q-Switch is opened. The figure above shows the relationship between pulse energy and Q-Switch delay.By setting the Q-Switch delay time the pulse energy is selected. This allows a minimum energy pulse to be used for alignment and setup, and a high energy pulse to be used for the experiment. The Synchronizer and LaserPulse software set the laser pulse energy by setting the Q-Switch delay time.The pulse energy can also be set (on some laser models) by setting the flash lamp voltage. Adjusting the pulse energy with Q-Switch delay is recommended because it keeps the laser at the operating temperature it was tuned for. This maintains the best beam quality.0%50100150200250300Q-Switch delay (microsec)
31Light sheet Optics “Fan” laser light to create 2D measurement plane Combination of cylindrical and spherical lensesCylindrical lens diverges (fans) light in one directionDictates measurement areaSpherical lens waists (thins) light sheetAchieves “2D” sheet (very thin)Cylindrical lensSpherical lenswaistLaser light sheet optics are used to control the dimensions of the illuminated area. A cylindrical lens controls the light sheet height. A spherical lens is used to control the light sheet thickness.The top view shows the thickness of the light sheet. The cylindrical lens has no effect on the light sheet thickness. The spherical lens forms a waist in the light sheet at its focal point. The bottom view shows the light sheet height. The beam diverges in the vertical direction after passing through the negative focal length cylindrical lens. The positive focal length spherical lens reduces this height divergence only a little. Typically the cylindrical lens has a much shorter focal length than the spherical lens.Typically the camera views the light sheet near the thickness waist, where the intensity is highest, but any place along the light sheet can be used. When selecting the light sheet optics and camera viewing position the light sheet intensity should not be allowed to vary too much over the field of view.Fan
32PIV Image Capture Synchronizer controls timing with high precision Controls laser power and “que” timingInitiates camera captureCamera captures 2 images “back to back” upon triggerFrame A exposure Independent (user defined)Frame B exposure Dependent (fixed)Laser light intensity determines particle image brightnessNOT exposure timeTiming parameters chosen so that one laser pulse appears in frame A and another in frame B
33Synchronization Camera Exposures Camera Image Readout Pulse separation 4/14/2017Image 1 ExposureImage 2ExposureCamera ExposuresCameraImage ReadoutImage 1 ReadoutImage 2 ReadoutPulse separationLaserPulsesFrame straddling technique allows very short time between laser pulses and thus provides the capability to measure higher and higher speed flows. Frame Straddling requires a close synchronization between the camera exposure and the pulsing of the laser.Key camera specification is the time from the end of one exposure to the start of the next exposure, the minimum frame straddle time. When trying to make the highest flow velocity measurement the first pulse is delayed until the last moment of exposure 1, and the second pulse happens at the first moment of exposure 2.Actual frame straddle time requires that Image1 be removed from the pixels before the second exposure starts. The time to move the image from the pixels depends on the image brightness. Actual frame straddle times can vary with experimental setup.Camera Trigger: The Synchronizer Triggers the camera to start a double exposure sequence. The Frame Grabber is also triggered at this time to capture the next two images. An external trigger can initiate the camera trigger.Camera Shutter Feedback: After a short time the Shutter Feedback signals the start of the first frame exposure.Camera Exposure: The two frame exposure includes a short frame 1 exposure and a longer frame 2 exposure. The first exposure is just long enough to Q-Switch a Nd:YAG laser. During the second exposure image in frame 1 is readout.Camera Digital Video Image output: The first image(frame 1) moves quickly from the light sensitive pixels to the readout registers. It then takes one frame time to transfer the image out of the camera and to the frame grabber. The second image cannot end until all of the first image has been read out of the camera.Laser Pulses: The first laser pulse must end before the end of exposure 1. The second pulse must start after the beginning of exposure 2
34Synchronization-Example Timing 4/14/2017400 ns2 msCamera Exposures400 usCameraImage Readout2 ms500 usLaserPulsesFrame straddling technique allows very short time between laser pulses and thus provides the capability to measure higher and higher speed flows. Frame Straddling requires a close synchronization between the camera exposure and the pulsing of the laser.Key camera specification is the time from the end of one exposure to the start of the next exposure, the minimum frame straddle time. When trying to make the highest flow velocity measurement the first pulse is delayed until the last moment of exposure 1, and the second pulse happens at the first moment of exposure 2.Actual frame straddle time requires that Image1 be removed from the pixels before the second exposure starts. The time to move the image from the pixels depends on the image brightness. Actual frame straddle times can vary with experimental setup.Camera Trigger: The Synchronizer Triggers the camera to start a double exposure sequence. The Frame Grabber is also triggered at this time to capture the next two images. An external trigger can initiate the camera trigger.Camera Shutter Feedback: After a short time the Shutter Feedback signals the start of the first frame exposure.Camera Exposure: The two frame exposure includes a short frame 1 exposure and a longer frame 2 exposure. The first exposure is just long enough to Q-Switch a Nd:YAG laser. During the second exposure image in frame 1 is readout.Camera Digital Video Image output: The first image(frame 1) moves quickly from the light sensitive pixels to the readout registers. It then takes one frame time to transfer the image out of the camera and to the frame grabber. The second image cannot end until all of the first image has been read out of the camera.Laser Pulses: The first laser pulse must end before the end of exposure 1. The second pulse must start after the beginning of exposure 2
35Synchronizer System External trigger -or- Insight3G Software Capture interfaceComputerThe computer controlled Synchronizer is at the heart of all PIV systems. The primary function of a Synchronizer system is that of a precise timer and a master system controller. It synchronizes the timing of image capture with the pulsing of the laser. Acting as the master controller for system components, it automates control of the timing between laser pulses, camera, camera interfaces and any external device during system set-up and image acquisition. The Synchronizer enables the system to be completely computer controlled via a serial interface. Special crosscorrelation cameras are employed in PIV which enable pairs of images be taken with short time between pulses. Control of such cameras is built into the synchronizer. Signals for the laser flash lamps and Q-switches, the camera and the frame grabber are generated and automatically synchronized for effortless image acquisition. For PIV applications, pulse delay time and the time between pulses necessary to collect frame-straddling images, are controlled visa TSI’s INSIGHT software. The synchronizer has an auxiliary output for controlling various devices in an experimental rig. For periodic flows, where phase locked velocity measurements are desirable, the Laserpulse synchronizer can be externally triggered using a TTL signal from the experimental apparatus. For example, in an IC engine experiment, images are captured at a certain crank angle position to be ensemble averaged. In bio-fluid mechanics relating to heart valve flows, images are captured at certain events in the systolic/diastolic cycle. The synchronizer system provides all the needed control signals – and does not need adjustments – for the cameras, lasers and other devices supplied as part of the TSI PIV system.Lasercontrol
36PIV Rules of Thumb for great results The key to good measurements is good raw dataThe raw data of PIV measurements are particle imagesIn the Ideal CaseSeed particle images are 3 – 5 pixels in diameter5 – 15 particles per interrogation regionMaximum particle displacement approximately 25% of size of interrogation regionIt is not always possible to satisfy these in all measurements, but they are good experimental goals
37Experiment Considerations Important considerations:What size field of view?What desired spatial resolution?Are the above 2 realistic together?Appropriate seedingOptical accessCamera 90 degrees with respect to laser sheetView is not distortedWhat does it look like when you view the measurement region from the perspective of the camera?0.2 m0.2 mField of ViewUsing 4MP (2k x 2k) Camera……apply 32pxl x 32pxl interrogation region…apply 64pxl x 64pxl interrogation region
38Experimental Setup General Setup Setup Tips Laser illuminates plane of interestCamera 90 degrees to light sheetGood optical access (no distortions)Camera is focused on ‘waist’ of thelight sheetSetup TipsInitial setup: steps should be performed in ‘free; continuous’ mode of the Insight 3G softwareUse room light, no laserFinal setup: the best practice is to focus camera on seed particles in the flowThis assures that the focus is optimized to the laser sheet
39Experimental SetupWhen setting up the cameras to view laser light scatteredfrom particles:Start on a large f-number (small aperture) and low laser energyIncrease alternately until seed particles are well-illuminated with minimal pixel saturationWARNING: IF LARGE REGIONS OF SATURATED (PINK)PIXELS APPEAR, STOP CAPTURE IMMEDIATELYLaser light can damage camera pixelsReduce laser energy and/or increase f-number
40Calibration What are we ‘calibrating’? We need to show the software how to convert from pixel units to the physical units relevant to our flow spaceWe are considering the simplest case in 2D PIV, where we view the light sheet at 90 degrees with good optical accessOff-axis viewing and optical distortions are sometimes unavoidable, and appropriate corrections can be madeIn these cases, we need to give the software more information to be able to convert pixels to mm over the entire image regionAddressed in future webinars
41Calibration Procedure Assumption: Cameras focusedon laser light sheetCalibration StepsCapture and save image of “Ruler” in plane of laser light sheetCreate a calibration file with the saved imageUse the 2D-Calibration software tool to measure across a known distance in the calibration image
42Seed Density Considerations for Seeding Density: The true test of appropriate seed particle density is examining the Particle ImagesConsiderations for Seeding Density:Field of View is knownDesired Spatial ResolutionknownInterrogation Region SizeDo I have enough particles ineach interrogation region?5-15 particles per region
43Optimizing DtIn addition to imaging the flow field, we must also determinethe timing between Frame A and Frame B.This parameter is called ΔtWe must select a Δt so that the displacement follows our “rule of thumb” of 25% of our intended interrogation regionEnsure consistency with experimental objectivesField of viewSpatial resolutionAppropriate limits on particle displacementMaximum particle displacement should be approximately 25% of the interrogation region
44Optimizing DtThe first step in optimizing Dt is developing “an eye for it”Can you see the displacement?Although qualitative, this is a critical step in optimizing a PIV measurementIf displacements appears random, reduce DtIf there is little / no displacement, increase Dt
45Optimizing DtThere are several ways to ‘measure’ the displacement for a more quantitative assessmentZoom into individual particlesPerform ‘Point Processing’ and assess displacement of individual spotsRemember, this process is iterative. Expect to alternate between experimental optimization and processing adjustments
46Optimizing Dt Timing Setup The Timing Setup Window in Insight 3GDt
47Basic PIV Processing Start with “general purpose” settings Basic processing lets you know what the flow is doing and how well your images processA good foundation for further optimizationIteratively optimizeThis may include making experimental changes and acquiring new images
48PIV Rules of Thumb Review Seed particle images are 3 – 5 pixels in diameter5 – 15 particles per interrogation regionMaximum particle displacement approximately 25% of size of interrogation region3-5 pixel particles allows for a Gaussian ‘fit’ to determine location to subpixel accuracy, but still allows for a high seeding densityMultiple particles per interrogation region strengthen the correlation25% displacements keep most particles within the interrogation region, allowing the search to match with most of the spots
49Basic PIV ProcessingInsight3G’s default settings are a good starting pointClassic PIV produces 1 vector field from 1 image64x64 pixel size for both spot A and spot B works well for moderately-seeded flows with similar x and y velocity rangesThe Nyquist grid setting gives us 50% overlap between neighboring spotsSince we are using NyquistGrid, the starting spots are the final spotsThe FFT correlator is a good general-purpose choice.This enforces our 25% rule of thumbThe Gaussian peak engine is normally the best choice for obtaining accurate velocities.
50Evaluating Processing The easiest way to evaluate your processing settings is to process a vector field.Adjust the scale of the vectors for a clear viewDo the generated vectors appear physically reasonable?Are there many “red” vectors that failed validation?
51Evaluating Processing Vector Statistics provide more useful informationMaximum U and V pixel displacementsNumber and percentage good vectors
52Evaluating Processing This image is a good candidate for increasing the spatial resolution by decreasing interrogation region size99.6% valid vectorsmaximum particle displacements of >5 pixelsHigh seeding density suggests 5-15 particle rule of thumb possible for smaller interrogation regionWhat to consider:How much smaller?Spot A? Spot B?
53Spot DimensionsThe Spot A size is used to determine spatial resolutionVelocity of all particles in Spot A is averaged into a single vectorThe Spot B size and maximum dx, dy determine search area sizeLarger sizes slow down processing, only help with large displacementsSpot ASpot B
54Adjusting spot dimensions Dimensions of Spot A determines:How many particles in the interrogation region (5-15)Smallest resolvable flow features (spatial resolution)Maximum allowed displacement (dx, dy):Maximum measurable particle displacements (dynamic range) in X and Y25% spot width, heightAdjusting these features can accommodate processing smaller interrogation regions.
55Adjusting Spot Dimensions Reprocessed with 32x32Now we are starting to get more invalid vectors.Why? Let’s look closer:Uneven seeding leaves some spots empty
56Adjusting Spot Dimensions Point Processing lets us evaluate the outcome of processing at a given spot.Here we see how low seeding density affects the correlation mapMultiple peaksLumpy Background
57Point Processing Point Processing shows us other information as well Peak RatioProcessing AlgorithmsInput and as-processed spotsLocation of the correlation peak within the mapNear the center is bestDx, Dy
58Adjusting Spot Dimensions For some flows, square spots may not be optimalFor example, a boundary layerLarge displacements, low gradients in X (need large spot size)Small displacements, sharp gradients in Y (need small spot size)Can use rectangular spots to optimize both dimensionsRectangular spot A averages velocities over a rectangular regionRectangular spot B combined with larger allowed displacements looks for a correlation over a larger area
59Adjusting spot dimensions This flow is sufficiently seeded for 32x32 (or even 24x24) interrogation, but we still get many invalid vectors.Dx is too large for 32x32 spot A and spot B sizeTry different spotA and spotB sizesHere we could also offset the search region in spot BRequires knowledge of the flow fieldOnly works in flows with a dominant flow direction
60Adjusting spot dimensions 64x24 Spot A64x24 Spot B0.25 Maximum dx32x32 Spot A64x32 Spot B0.49 Maximum dx
61Single-Pass drawbacks With a fixed spot size and single pass, smaller interrogation regions give higher spatial resolution but reduced dynamic rangeThere are workaroundsWe can relax the 25% rule of thumb and use a larger spot B if the image quality is highWe can use rectangular interrogation regions if the flow geometry permitsWe can offset spot BThere is a better option – Multi-pass processingCan optimize for both spatial resolution and dynamic rangeImproves SNR for better quality vector fieldsOperates like a dynamic spot offset
62Multi-Pass PIV Processing Using the RecursiveNyquist-Grid Multi-pass PIV processing lets us increase our spatial resolution
63Multi-pass PIV Processing Process PIV image multiple times in sequenceSuccessive passes use the initial results as a starting guessThe search area in Spot B is “offset” relative to Spot A – centered on the displacement “guess”Spot A and Spot B will contain only the “fingerprint”Higher SNR, more and better vectors25% rule of thumb only applies to first passSpot BSpot A
64Vector Post Processing Vector ValidationBasic PIV processing checks only SNR to validate vectorsNeed a way to “weed out” invalid vectorsIn final dataIn between passesVector ConditioningInterpolation to replace discarded vectorsSmoothingEspecially important for intermediate passes in multi-pass techniquesAvoids using a bad vector as a guess for additional processing
65Vector Validation Global Validation Check velocity against specified range or number of standard deviations from the meanGood when you know what range of velocities to expect
66Vector Validation Local Validation More flexible, compares a vector with its neighborsDiscriminates based on difference between the vector and the median or mean displacement (in pixels), or dimensionless “Universal” medianDecrease neighborhood size for flows with high gradientsOption to replace discarded vectors with local median or secondary correlation peak
67Vector Conditioning Apply After discarding invalid vectors Replace discarded vectors with the local median or mean velocityOptionally apply a low-pass filter to the vector field for smoothingKeep in mind that smoothing changes all vectors – modifying your data
68Deformation Grid Processing Means to sharpen peak of correlation mapAccounts for high rotation or velocity gradients within interrogation regionsAnticipates displacement of particles, and deforms Spot B to account for such gradientsSurrounding velocities used to determine deformationSpot ASharpen peak of correlation functionDeformed Spot B“Traditional” Spot B(Overlaid Spot Masks)
69Deformation Grid Processing Stronger correlation peak and better accuracy with high gradientsHigher computational cost