6th Int conference on numerical combustion -- New Orleans; March 4-6, 1996
Moti L. Mittal
Program for Computational Mechanics
Ohio Supercomputer Center
1224 Kinnear Road
Columbus, OH 43212
Department of Mechanical Engineering
The Ohio State University
Columbus, OH 43210
We present a description of the coal particle behavior as influenced by the flow fluctuations in a hot combustion atmosphere. The particle reactions in the combustion chamber depend on type of coal, particle size, rate of heating, and gaseous atmosphere surrounding the particle. The gaseous atmosphere surrounding the particle is responsible for particle motion as well as heat flux to the particle surface by convection. Flow fluctuations of the surrounding gaseous atmosphere change the particle flights and the heating rate by altering its path.
For small Mach number M a flows (for combustion flows M a << 0.1), the accoustic wave propagation is eliminated from the system of equations. The pressure field is separated into a thermodynamic part Pr and a dynamic part for the momentum equation. The thermodynamic part Pr is constant in space. The flow field in the combustion enclosure is calculated by directly solving the time dependent Navier-Stokes equations in a sudden-expansion axisymmetric geometry. Starting from an initial flow, the time dependent Navier-Stokes equations are solved for sufficient length of time to achieve a quasi-steady state flow condition.
A vorticity stream function formulation, instead of one based on the primitive variables, is used for the flow simulation. The formulation is based on global conservation of mass, momentum, and energy. A clustered conformal coordinate procedure establishes a surface-oriented orthogonal coordinate system with grid clustering/stretching characteristics and a novel grid-point placement in the physical flow domain.
The governing equation for vorticity transport is solved by an alternating-direction fully implicit (ADI) method in time, using a forward time marching scheme. To maintain numerical accuracy and minimize roundoff errors, we use GMRES iterative method to solve the resulting systems of linear equations. The corresponding stream function distribution is obtained by a fully implicit solution of the elliptic stream function equation. High performance computing techniques are used for the numerical solutions.
On this flow field, we have imposed experimental temperature and gas concentration fields measured in pulverized coal flames in the experimental furnace of the International Flame Research Foundation. Particles of different sizes are injected into these combined fields. Lagrangian equations are used to predict the particle location, point by point through the hot gas field, with prediction of particle motion governed by standard viscosity and drag relations for particles in a flow field with relative velocity. With each incremental calculation, the particle path is traced out through the gas field, and, knowing the location at each point, we also know the gas temperature and reactive gas concentrations.
A standard energy balance equation based on the rate of reaction and rate of energy loss by relevant heat transfer methods describes the particle temperature. The associated reduction of particle diameter and density is then incorporated in the drag relations for the particle motion in the next time interval calculation.
Reaction is calculated point by point using the relevant values of the imposed field parameters. Injected particles first pyrolyze. After the pyrolysis, the resultant char particle is reacting. The gaseous combustion of the volatile matter is considered and the production and profiles of carbon dioxide CO2 and nitric oxide NO are calculated. By this procedure, the predicted behavior of single particles in the turbulent flow is followed, giving the particle path, the changes in particle temperature along the path, the changes in volatile loss, the production of pollutants from particle combustion, and the changes in particle diameter to burn-out.
Examining the behavior of single particles of different sizes, injected into the flow and followed individually, we show that particle trajectories vary substantially, not only with particle size and injection location but also with time sequence. The consequence is that particles can pyrolyze and burn out at very different rates and with different temperature histories, depending on the size and input location. The difference in the particle heating rate changes the particle burning rate and the emission. Our particular focus is to emphasize the effects of flow fluctuations on the behavior of particles followed individually through the flame.