OCR Text |
Show furnace code is also available. This version can be applied to cylindrical furnaces fired with an axisymmetrical flame. The model is basically a zone method. The flow pattern in the furnace necessary to solve the total heat balance of volume zones is prescribed in decoupled manner. The distribution of heat release from gaseous fuel components or volatile matter is calculated automatically from the prescribed flow pattern and from estimated mixing (burning) times. Burnout of char particles and corresponding heat release is calculated from mass balances of fixed carbon which are also based on the prescribed flow pattern. The zone arrangement can have a variable fineness. The volume zones themselves may have a general prismatic shape. Thus, almost any furnace geometry and firing pattern encountered in practice may be investigated. Typical zone arrangements used in the application of the model to a pilot-scale and two boiler furnaces are shown in Figure 1. This figure also shows how the geometry of actual radiant superheater and reheater sections in the upper furnace of a boiler can be simulated in the furnace model. The radiation model in the furnace heat transfer code is derived from Monte Carlo calculation techniques. Some features of the radiation model are indicated in Figure 2. The emissive power of each furnace zone is distributed among a discrete number of beams of unit radiation, which are traced through the arrangement of furnace zones. The energy fluxes of the beams are gradually attenuated and redirected due to wall reflection and eventual scattering until final absorption occurs. At the end of the radiative exchange calculation a balance is set up for each zone. The difference between the sum of energy fluxes emitted by a zone and the accumulated absorbed energy fluxes is the net radiative heat flux necessary to solve the total heat balance of this zone. In pure Monte Carlo methods all steps in the history of a radiation beam are determined with help of random numbers. However, in order to reduce the statistical error for comparative calculation times, some random steps of such a pure Monte Carlo approach were replaced in our model by deterministic decisions, resulting in a so-called semistochastic model for furnace heat transfer predictions (14). The accuracy of the semistochastic method is increased as more beams^re traced. It is possible to obtain an acceptable accuracy of overall heat transfer with less than 10000 beams (5). Radiating species considered in our current model are H2O, COo, soot, cha7, and ash particles. Nongray radiation of the gaseous species H2O and CO2 and of soot is simulated with weighted gray gas approach. Radiation of char and ash particles is assumed to be gray. The absorption coefficient of char particles is calculated from local size distributions predicted by the char combustion model. The present furnace model can take anisotropic scattering of radiation at ash particles optionally into account. In this case, effective scattering and absorption efficiencies as well as phase functions of ash particle clouds are calculated from measured or assumed ash particle size distributions and from complex refractive indices. The model of radiative exchange is directly coupled with total heat balance for all gas and surface zones of the furnace of unknown temperatures, and, if required, for all radiant superheater and reheater surfaces. Radiative and total heat balances are based on locally prescribed thickness of ash deposits and values of emissivity and thermal 4 |