A numerical study of entropy generation in nanofluid flow in different flow geometries.

Abstract

This thesis is concerned with the mathematical modelling and numerical solution of equations for boundary layer flows in different geometries with convective and slip boundary conditions. We investigate entropy generation, heat and mass transport mechanisms in non-Newtonian fluids by determining the influence of important physical and chemical parameters on nanofluid flows in various flow geometries, namely, an Oldroyd-B nanofluid flow past a Riga plate; the combined thermal radiation and magnetic field effects on entropy generation in unsteady fluid flow in an inclined cylinder; the impact of irreversibility ratio and entropy generation on a three-dimensional Oldroyd-B fluid flow along a bidirectional stretching surface; entropy generation in a double-diffusive convective nanofluid flow in the stagnation region of a spinning sphere with viscous dissipation and a study of the fluid velocity, heat and mass transfer in an unsteady nanofluid flow past parallel porous plates. We assumed that the nanofluids are electrically conducting and that the velocity slip and shear stress at the boundary have a linear relationship. We also consider different boundary conditions for all the flow models. The study further analyzes and quantifies the influence of each source of irreversibility on the overall entropy generation. The transport equations are solved using two recent numerical methods, the overlapping grid spectral collocation method and the bivariate spectral quasilinearization method, first to determine which of these methods is the most accurate, and secondly to authenticate the numerical accuracy of the results. Further, we determine the skin friction coefficient and the changes in the heat and mass transfer coefficients with various system parameters. The results show, inter alia that reducing the heat transfer coefficient, the particle Brownian motion parameter, chemical reaction parameter, Brinkman number, thermophoresis parameter and the Hartman number all lead individually to a reduction in entropy generation. The overlapping grid spectral collocation method gives better computational accuracy and converge faster than the bivariate spectral quasilinearization method. The fluid flow problems have engineering and industrial applications, particularly in the design of cooling systems and in aerodynamics.