Finite-element analysis of turbulent diffusion flames

Abstract

A numerical procedure for the computation of confined, axisymmetric, turbulent diffusion flames is developed. The governing elliptic partial differential equations are solved by the finite-element method to generate the flexible mesh in flow regions with physical and geometrical complexity. This algorithm adopts a sequential velocity-pressure formulation that allows the uncoupling of pressure and velocity. In order to minimize the numerical diffusion, the streamline upwind/Petrov-Galerkin formulation is employed. Turbulence is represented by the k-e model, and the combustion process involves an irreversible one-step reaction at an infinite rate. The time-averaged mixture properties are obtained by weighting the property function with a beta probability density function (pdf). Predictions are made for turbulent reacting and nonreacting flow systems with recirculation, and the results are compared with experimental data. Nomenclature a,b -parameter in /3 function pdf C p = specific heat C^ = constants in the turbulent model / = mixture fraction g = square of fluctuations of mixture fraction h = enthalpy of mixture Hf U = heat of reaction k = kinetic energy of turbulence M = molecular weight P = pressure R = universal gas constant r = distance in radial direction s = stoichiometric mass of oxidant T = temperature u = axial velocity component v -radial velocity component W = weighting function x = axial coordinate distance j8 = conserved scalar £ = nondimensional mixture fraction e = rate of dissipation of turbulence energy fji = molecular viscosity Ht = turbulent viscosity p = mixture density $ = continuous weighting function = discontinuous weighting function Subscripts A = airstream F = fuel stream fu -fuel ox = oxidant pr = product Superscripts = correction * = estimation

A simple and efficient error analysis for multi‐step solution of the Navier–Stokes equations

“…Another advantage is the Ladyzhenskaya-Babuska-Brezzi (LBB) [5; 6] condition can be relaxed, and an equal interpolation for the velocities and pressure can take place. In this work the method of Comini and Del Giudice [1; 2] and Kim and Chung [3] is compared with that of Dukowicz and Dvinsky [7]. The result of this comparison shows that the method of Comini and Del Giudice is a subset of that presented by Dukowicz and Dvinsky.…”

“…The method chosen to solve the incompressible Navier-Stokes equation is the method presented by Comini and Del Giudice [1; 2] and used later by Kim and Chung [3]. This method has several advantages over direct methods, one major advantage is less computer memory is required to obtain a solution [4].…”

Investigation of Radial Swirler Effect on Flow Pattern inside a Gas Turbine Combustor

“…In the present simulation, k-ε turbulence model with the default setting were used. Turbulence is represented by the standard k-ε model, which provides an optimal choice and economy for many turbulent flow (Kim, 1989). Menzies (2005) had studied the behavior of five k-ε variants in modeling the isothermal flow inside a gas turbine combustor and compared the results with the experimental data of Da Palma (1988) for the velocity and turbulence fields.…”

Algebraic multigrid methods for the solution of the Navier-Stokes equations in complicated geometries

“…After the points N L were divided into the sets F L and C L , we could de®ne the interpolation as given in (25). In the algorithm of Ruge and Stu Èben 11,12 a little more sophisticated interpolation is used, which gives better results in numerical experiments:…”