Scientific enquiry has aligned itself, in recent times, to the understanding of the flow physics encountered in micro- and nano-scales. The prominence of this line of enquiryis due to its recurring influence in the fields of MEMS and porous media. These flows occur in pathways whose width is comparable to the mean free path and thus, are classified as rarefied gas flows. Assuming continuum and resorting to traditional CFD techniques leads to results with severe reservations while, deterministic kinetic theory based approaches, such as DVM, implement liberal approximations such as restricting the velocity space to a limited set determined through pre-cognizance or trials. Even stochastic particle approaches, such as the DSMC, usually adept at resolving such flows, proves to be prohibitively expensive owing to the small signal to-noise ratio necessitating a large number of samples to obtain appreciably accurate results. The present research is aimed at recovering the inherent advantages of particle methods through the development of parallel kinetic particle-based solver founded on the low variance ideologies. The solver is validated against classical fluid dynamics problems prior to application to simple, yet practical, MEMS applications such as isothermal flow through long ducts and past infinite arrays. The final objective of the current research is aimed at understanding and simulating the transport of unconventional gases in subterranean micro-porous networks which is of great significance in the oil and gas industry.To this end, the simulation of porescale flows for the entire rarefaction spectrum through idealized porous media such as Sierpinski carpets, Menger sponge etc. along with image-reconstructed porous media such as 2D Berea sandstone and 3D rock samples such as Gambier, Castlegate and Fayetteville shale are carried out and analysed. The unique approach enables the solver to obtain novel, substantial and reliable insights at the pore level while employing appropriate averaging to predict macroscopic properties such as apparent permeability and tortuosity with a previously unprecedented computational effciency.
|Date of Award||8 Oct 2019|
- University Of Strathclyde
|Sponsors||University of Strathclyde|
|Supervisor||Yonghao Zhang (Supervisor) & Monica Oliveira (Supervisor)|