1. The Transport Process
2. The droplet transport processcanstronglyinfluence the droplet heat and mass transfer rates due to background gas gradients in temperatures and velocities. Droplet spatial and velocity distributions in the flow fieldwilldepend upon dropletsize, droplet-background gas relative velocity, and the temperature-sensitive c e efficient of viscosity of the background gas. In Sec. 5, whereresults of theflowmodelingarediscussed,itwill beseen that droplet distributions may be differentdue to these three properties with regions of droplets accumulation or spreading observed. For these reasons, it isimportant to establishthe accuracy of thedroplet transport model and its region of applicability for the flowconditionsunder consideration. First,weconsider the trajectory modelingof 10p-sized dropletsinjected into a background gasof argon ataconstanttemperatureof 2000Kand 1atmpressure, movingwithonlyan axialvelocityforanaxisymmetricgeometry. Notethat the properties of the background argon atoms,suchas position and velocity, do not change in the simulation. Although this is a highly simplified flow condition we can determine the accuracy of the proposed transport model with an analytic result. We then use Singlesphere DSMC calculations to choose the best model for the drag force of micron-sized droplets in the twophase flow. 2.1 Comparison of Statistical and De
3. The DSMC calculation was performed using a computational domain of 0.03 m x 0.004 m with 300 x 40 cells. With the use of both radial weightingfactors and aspeciesweightingfactorforthe droplet particles, about2,225,580argongasparticlesand 15,000droplets were modeled to simulate the transport process. The radialweighting factorsforaxisymmetricflowsarediscussedin Ref. [ll]. The species weights imply the use of differentreal-to-simulated moleculeratios, FnUm, for different gas species. In the calculations discussed in this subsection, speciesweights were used in the statistical transport model due to the largedifferencebetween the argon and droplet densities. The number density of argon ison the order of 3x particle/m3 and the number density of water droplets at the inflow boundary is 4.12 x 1013 particle/m3. Values of 0.6 x lo'* and 0.6 x 10' were therefore taken for the argon and droplet F, respectively. To decreasethe number of argon-droplet collisionsand increase the efficiencyof these collisions (i.e. momentum transfer in a single collisions), collision weights were also used in the DSMC transport calculations. The use of these weights implies the application of ascale factor to increase the momentum transfer per collision, and proportionallydecreasethecollisionfrequencybetweenargon and droplets. Different collision weights were assumed in these calculations depending on the size of the droplets, with the general rule of having a virtual argonmassof about 0.1percent of the droplet mass.(2] Several different values for the collision weights were tried and it was found that the results remained the same, whilethe computational efficiency improvedsignificantly. Droplets have an average initial velocity of 10m/s in theaxialdirectionand the background argon atomshave afixed averagevelocityof 500m/s, apressure of 1atm, and a temperature of 2000 K. A time step of 1 xlo-' s was used toensure that the droplet displacement per time step was lessthan one cell size. Results were sampled at the 5,000 th time-step. Solutions were alsoobtained with a finer grid and larger numbers of samples to ensure that the result were accurate and grid independent.
4. Process
5. The Coalescence Process