ABSTRACT
A complete combustion system may be extremely complex and can include a wide range of physical processes that are often highly interactive and interdependent. A given combustion system may include:
• Turbulent fluid dynamics in the flame with laminar fluid dynamics in the bulk of the combustor
• Multi-dimensional flows that could include swirl • Multiple phases that could include gases, liquids, and solids, depending on the fuel
composition
0-8493-2000-X/00/$0.004.50 © 2001 by CRC Press LLC 33
• Combustion of solid fuels that may include processes like vaporization, devolatilization, and char production and destruction
• Very high temperature, velocity, and species gradients in the flame region with much lower gradients in the bulk of the combustor
• Large material property variations caused by the wide range of temperatures, species, and solids present in the system
• Multiple modes of heat transfer, especially radiation that is highly nonlinear and may include wavelength dependence
• Complex chemistry involving numerous reactions and many species, most of which are in trace amounts
• Porous media • Catalytic chemical reactions in some limited applications • Complex, nonsymmetrical furnace geometries that may include an intermediate surface
to separate the combustion products from the load • Multiple flame zones produced by burners that may be operated at different conditions
and whose flames interact with each other • A heat load that may be moving in a complicated manner and interacting with the
combustion space above it in a nonlinear manner • A heat load that may produce volatile species during the heating process • A heat load whose properties may vary greatly with temperature, physical state, and even
wavelength (for radiation) • A transient heating and melting process that may include discrete material additions and
withdrawals
There are many challenges caused by this complexity, including inadequate physics to properly model the problem, large numbers of gridpoints requiring large amounts of computer memory, and long computation times. The simulation results may be difficult to validate as many of the experimental measurements are difficult, time-consuming, and costly to make in industrial combustors. Therefore, in most combustion simulations, simplifying assumptions must be made to get costeffective solutions in the amount of time available for a given problem. The actual simplifications depend on many factors, including the level of accuracy required, the available amount of computing power, the skill and knowledge of the modeler, the experience with the given system being simulated, and the time available to get a solution. These simplifying approaches are briefly discussed here. More detailed information on each aspect of modeling is given later in this chapter. Spalding (1963) discussed simplifying approaches to combustion modeling and noted that the main concern is which modeling rules can be ignored to simplify the problem and then to estimate the errors in the resulting predictions.' Also noted was the difficulty in matching all the dimensionless groups in a large-scale problem with small-scale experiments. Weber et al. (1993) classified models for designing industrial burners into three categories.2 First-order methods give rough qualitative estimates of heat fluxes and flame shapes. Second-order methods give higher accuracy results than first-order methods for temperature, oxygen concentration, and heat flux. Third-order methods further improve accuracy over second-order methods and give detailed species predictions in the flame that are useful for pollutant formation rates. The order used will in large part depend on the information and accuracy needed.
