Experimental uncertainties associated with particle image velocimetry (PIV) based vorticity algorithms

Luff, J. D.; Drouillard, Thomas F.; Rompage, A. M.; Linne, Mark; Hertzberg, Jean

doi:10.1007/s003480050263

Cited by 90 publications

(43 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For the swirl view measurements, the Laser beam was converted to a sheet of about 0.5-1 mm thickness by a combination of spherical and cylindrical lenses, fired horizontally through the pentroof of the engine in full optical setup (quartz liner and sapphire piston window); the camera was positioned in front of the 45° mirror that was situated within the Bowditch piston arrangement, see Figure 2 Several aspects of the PIV technique, including practical application, precision and uncertainties were optimised according to [23] in consultation with seminal publications on the specifics of PIV [24][25][26][27][28][29][30][31] and previously published PIV studies focused on internal combustion engine measurements [6][7][8][9][32][33][34]. In this context, Table 3 summarises the system's main settings and several other details of the PIV measurements presented in this paper, e.g.…”

Section: Test Procedures Pivmentioning

confidence: 99%

Modulation of integral length scales of turbulence in an optical SI engine by direct injection of gasoline, iso -octane, ethanol and butanol fuels

Aleiferis

Behringer

2017

Fuel

View full text Add to dashboard Cite

In-cylinder air flow structures are known to play a major role in mixture preparation and engine operating limits for spark-ignition engines. In this paper PIV measurements were undertaken in an optical spark-ignition at 1500 RPM, part-load with 0.5 bar intake plenum pressure. One of the PIV planes was vertical, cutting through the centrally located spark plug (tumble plane). The other plane was horizontal 1 mm below the spark plug ground electrode (swirl plane). The effect of engine head temperature was also examined by using engine-head coolant temperatures of 20 °C and 80 °C. The flow field was examined late in the compression stroke at typical ignition timing. The study was conducted under air-only motoring engine conditions but also under fuelled conditions in the early intake stroke using direct injection of gasoline, iso-octane, ethanol and butanol fuels. The flow field under air-only motored conditions showed velocities between 3-5 m/s predominantly from intake to exhaust. Little differences were observed between hot and cold engine-head temperature; typically ~10% larger mean velocity and turbulent kinetic energy was seen on the intake side. Integral length scales were on the tumble plane between 2-4.5 mm in the vertical and 4-7 mm in the horizontal direction. The swirl view showed scales between 4-10 mm, larger at cold than at hot engine-head conditions. The vertical length scales appeared to be limited by the clearance height, scaling typically by about 10-15%. The horizontal components scaled to the cylinder bore diameter by about 5-12%. The fuel injection process in the early intake stroke led to little differences in the general mean flow structure at ignition timing, except for a small increase in the maximum velocities, ~10%. The turbulent kinetic on the tumble plane was highest towards the exhaust side of the engine under non-fuelled engine conditions but fuel injection resulted in highest values on the intake side. The integral length scales with fuel injection were of the same order of magnitude to those of air only measurements on the tumble plane, but showed distinctly larger areas with length scales up to 9 mm on the swirl plane. Differences between fuels for their fuel specific injection duration were small, with average length scales between 4-6 mm. Ethanol exhibited typically largest scales and butanol smallest.3

show abstract

Section: Test Procedures Pivmentioning

confidence: 99%

Modulation of integral length scales of turbulence in an optical SI engine by direct injection of gasoline, iso -octane, ethanol and butanol fuels

Aleiferis

Behringer

2017

Fuel

View full text Add to dashboard Cite

show abstract

“…The combination of Gaussian smoothing with this finite difference scheme has been used by Ge et al 27 , among others, for similar measurements. It has been shown by Luff et al 41 that this approach is superior to more complex finite differences. In tests on synthetic PIV data, is has been demonstrated that the smoothing step can reduce noise in the gradients from 68% to about 4% 41 .…”

Section: Summary Of Data Processingmentioning

confidence: 99%

High-Resolution Measurement of the Unsteady Velocity Field to Evaluate Blood Damage Induced by a Mechanical Heart Valve

Bellofiore

Quinlan

2011

Ann Biomed Eng

View full text Add to dashboard Cite

“…Consequently, if uncertainties within 1% are desired with turbulence intensity of 10%, 100 vector maps would be required for the mean velocity but 20,000 vector maps for a component of the Reynolds normal stress. Luff et al (1999) conducted a computational and experimental study to investigate the typical experimental uncertainties that could be expected in the use of PIV systems to measure vorticity (which is calculated via measured velocity gradients or their equivalents). They showed that average uncertainties up to 37% and peak uncertainties of the order of 550% could result from the direct numerical processing using autocorrelation techniques.…”

Section: Particle Motionmentioning

confidence: 99%

“…With appropriate processing, average uncertainties could be reduced to about 4% and peak uncertainties were on the order of 20%. Various approaches to improving the situation are provided in the Luff et al (1999) article. Forliti, Strykowski, and Debatin (2000) examined bias and precision uncertainties of various peak-finding algorithms with TSI software.…”

Section: Particle Motionmentioning

confidence: 99%

Estimated Uncertainties in the Idaho National Laboratory Matched-Index-of-Refraction Lower Plenum Experiment

McEligot¹,

McIlroy²,

Johnson³

2007

View full text Add to dashboard Cite

The purpose of the fluid dynamics experiments in the MIR (Matched-Index-of-Refraction) flow system at Idaho National Laboratory (INL) is to develop benchmark databases for the assessment of Computational Fluid Dynamics (CFD) solutions of the momentum equations, scalar mixing, and turbulence models for typical Very High Temperature Reactor (VHTR) plenum geometries in the limiting case of negligible buoyancy and constant fluid properties. The experiments use optical techniques, primarily particle image velocimetry (PIV) in the INL MIR flow system. The benefit of the MIR technique is that it permits optical measurements to determine flow characteristics in passages and around objects to be obtained without locating a disturbing transducer in the flow field and without distortion of the optical paths. The objective of the present report is to develop understanding of the magnitudes of experimental uncertainties in the results to be obtained in such experiments. Unheated MIR experiments are first steps when the geometry is complicated. One does not want to use a computational technique, which will not even handle constant properties properly. This report addresses the general background, requirements for benchmark databases, estimation of experimental uncertainties in mean velocities and turbulence quantities, the MIR experiment, PIV uncertainties, positioning uncertainties, and other contributing measurement uncertainties.The accuracy of the velocity field determination is ultimately limited by the ability of the scattering particles to follow the unsteady motion of the continuous phase. For the particles in the MIR VHTR experiment, the estimated cutoff frequency is about 40 kHz. The frequency of shedding from the posts in the MIR model is estimated to be of the order of 1 Hz and, therefore, the particles should follow the flow.Based on review of the literature, we estimate that the random uncertainty in measured particle displacement for our system and experiment is about 0.3 pixel for a single realization away from the walls. As an example, we estimated the uncertainties in measurements for the maximum velocity in an inlet jet flow operating at about its minimum flow rate. Consequently, this example would provide worst-case estimates in some senses. For the uncertainty in the mean velocity at this location, one could report With the current MIR model and the 3D PIV system, McIlroy and colleagues examined the effects of sample size for a typical set of measurements. Trends were generally as observed by Uzol and Camci. Preliminary conclusions were that (for this view and processing parameters) about 750 images should be collected to reduce the scatter in mean velocity statistics to between 0.4 and 10% v for velocities greater than about 2 m/s. The trends predicted by an uncertainty analysis are consistent with the observations, but, for the most part, the expected values are less than the scatter observed. Preliminary conclusions from the comparisons are that (1) more samples should be collected than predict...

show abstract

Experimental uncertainties associated with particle image velocimetry (PIV) based vorticity algorithms

Cited by 90 publications

References 10 publications

Modulation of integral length scales of turbulence in an optical SI engine by direct injection of gasoline, iso -octane, ethanol and butanol fuels

Modulation of integral length scales of turbulence in an optical SI engine by direct injection of gasoline, iso -octane, ethanol and butanol fuels

High-Resolution Measurement of the Unsteady Velocity Field to Evaluate Blood Damage Induced by a Mechanical Heart Valve

Estimated Uncertainties in the Idaho National Laboratory Matched-Index-of-Refraction Lower Plenum Experiment

Contact Info

Product

Resources

About