Digital Holographic Method And Application To Measurement Of Particle Combustion And Droplet Atomization

Geometric and dynamic parameters of fuel particles and droplets change dramatically during combustion and atomization.This directly affects the combustion efficiency and air pollutant emissions.The underlying mechanism of the combustion of solid and liquid fuels are not completely clear due to the lack of quantitative,in-situ,time and space resolved characterization of fuel particles.Most of the optical diagnostic techniques are only able to measure limited parameters at a single point or two-dimensional(2D)plane.Examples are phase Doppler anemometry(PDA)for single-point size and velocity,and particle imaging velocimetry(PIV)for 2D velocity.Digital holography(DH),as an emerging three-dimensional(3D)imaging technique,has demonstrated its great potential to quantify 3D,multi-parameter characterizations of particles in multi-phase flows.However,there are some drawbacks of DH: z location in DH has relatively low accuracy.The result is easily affected by the flame when burning particles are measured.And 3D morphologies of ligaments can not be extracted.To solve the three problems above,the present work is devoted to the development and application of digital holographic method to the diagnostics of combustion and spray via theoretical,numerical,and experimental study.It enables us to investigate pulverized coal combustion and droplet atomization in a 3D,quantitative,and dynamic way.To improve the measurement accuracy of DH,a general model to describe the formation of digital in-line hologram with different recording conditions such as plane wave illumination,spherical wave illumination,and condition with imaging lenses is established.A model to describe the formation of digital off-axis hologram with arbitrary spherical reference wave is also proposed.Self-adaptive grayscale thresholds based on the intensity and gradient map of reconstructed images are adopted to detect particles.A new criterion for depth location using local variance of gradient in the vicinity of particle border is proposed which can alleviate the influences of neighboring particles and separate overlapping particles.The uncertainty of depth location decreases to 30 ?m(about 5 pixels),about 10% ~ 40% of the current level.To study and reduce the influences of flame on burning particle measurement,a model to describe hologram formation of particles in the flame is proposed.Holograms of particles in flames with different temperature profiles are simulated and reconstructed according to this model.It is found that when a particle is located in a region with high temperature gradient,the reconstructed image of its in-line hologram has astigmatism,lower contrast,and higher location error.However,reconstructed image of an off-axis hologram does not endure these drawbacks because the reference wave does not pass the flame.The influences of flame on both in-line holography an off-axis holography are relatively small when particles are located at a place with low temperature gradient.Experimental study on burning single droplet and droplet cloud of ethanol validates the theory and simulation.Furthermore,off-axis holography is proved able to measure the approximate temperature profile of a symmetric flame surrounding a single droplet.Burning coal and biomass particles in the jet flame and flat flame burner are investigated with digital holographic particle tracking velocimetry(DHPTV).Motions,rotations,and disintegrations of particles,volatile ejection and combustion,and soot accumulation are observed directly.Particle sizes increase slightly,and particle morphologies become more irregular in devolatilization.A secondary peak in size distribution appears at around 40 ?m after combustion which is due to particle disintegrations and soot accumulation.The soot wake moves faster than its parent particle and the relative velocity between them is close to the gas-solid slip velocity.To measure 3D liquid ligaments,an automatic algorithm to extract the 3D structure and volume of stringy objects is developed.Stringy objects are first recognized in the binary image according to the aspect ratio and solidity of connected regions.Then a stringy object is segmented into small sections along its skeleton and edge.The sections are sized and 3D located as dispersed particles before they are stitched together to be a whole object.A spring with known parameters is used to calibrate the accuracy which indicates the size error of 25 ?m and z location error of 200 ?m.Dynamic evolution of ligaments and droplets generated in the secondary atomization of ethanol drops 2.34 mm in diameter is characterized via 20 kHz high-speed digital holography.In the case of Weber number being 11,bag breakup happens in which a large number of smaller droplets(<300 ?m)are generated in the early stage and the newly generated droplets have larger size and smaller velocity over time.Then a toroidal rim forms and disintegrates into several droplets whose diameters are about 650 ?m and 1.1 mm,leading to trimodal distribution of the droplet size.In the case of Weber number being 25,multi-mode breakup happens and a multibranch ligament is produced.Both the toroidal rim and the multi-branch ligament occupy more than 90% of the initial volume.