Font Size: a A A

Liquid-gas phase coupling by means of Lagrange polynomials for the CFD simulation of high-velocity dense sprays

Posted on:2008-04-04Degree:Ph.DType:Dissertation
University:Michigan Technological UniversityCandidate:Stalsberg-Zarling, KristaFull Text:PDF
GTID:1452390005979986Subject:Engineering
Abstract/Summary:
High-velocity, dense sprays are encountered in many fields and industries such as engines, foods, pharmaceuticals and agriculture. One of the key numerical issues in the simulation of high-velocity, dense sprays is the mesh dependence resulting from the coupling between the liquid and gas phases. This mesh dependence was investigated using an enhanced version of the open-source software package KIVA-3. Particular emphasis was placed on diesel engine sprays. In order to address some of the problems that arise due to mesh dependence, a second-order accurate model using Lagrange interpolation polynomials was developed for evaluating gas properties at droplet locations and for distributing spray source terms onto the gas phase. The model was implemented into KIVA-3 for momentum, mass, energy and turbulence coupling. Computations were performed primarily for non-evaporating and evaporating sprays. Additionally, the method was validated for a real engine case. The results were compared with the standard liquid-gas coupling method of KIVA-3, i.e., the nearest neighbor method, and with experimental data when possible.; The application of the Lagrange method to momentum coupling in the simulation of non-evaporating sprays resulted in reduced mesh dependence on polar and cartesian meshes. This included improvements in spray penetrations, droplet sizes and spray structure, and good agreement was achieved between the simulated and experimental results. Additionally, CPU times were reduced on cartesian meshes, by more than half on the finest mesh, due to an increase in computational stability. For evaporating sprays, the Lagrange method was systematically applied to momentum, mass, energy and turbulence coupling and then to a combination of these couplings. A significant reduction in mesh dependence was achieved using the Lagrange method for mass coupling. When the Lagrange method was applied to all forms of liquid-gas coupling combined, the mesh dependence of liquid and vapor penetrations and gas velocities was diminished. CPU times were reduced by more than half for mass coupling alone due to increased computational stability, and by nearly half on all meshes when the Lagrange method was applied to mass coupling combined with any other type of liquid-gas coupling. Simulated results gave reasonable agreement with experimental liquid penetrations.; Finally, the Lagrange method was validated successfully for a real engine case. Good agreement with experimental results was achieved, and improvements in CPU times were observed for increasing mesh refinements.
Keywords/Search Tags:Sprays, Coupling, CPU times, Lagrange, Dense, Mesh, Gas, Engine
Related items