OCR Text |
Show the combustion chamber. These properties are either conserved in the form in which they were supplied to the combustion chamber through burners and other inlets or they are converted into other forms by chemical reactions, dissipation, emission/absorption, etc., depending on spatial location and time. The balances are numerically solved for a more or less large numbers of control elements in which the furnace divided, and for a more or less number of dependent furnace variables like temperatures, concentration of chemical species, velocities, pressure etc. representative for each control element. The fineness of subdivision into control elements is frequently used to classify available furnace models into two groups, namely: • the zone models and • the finite - difference models. The zone models use with respect to the whole furnace a macroscopic subdivision whereas the finite-difference models utilize almost microscopic control volumes. Zone models are usually so-called decoupled heat transfer models whereas in finite-difference models, fluid flow, combustion and heat transfer is modeled in a coupled manner. The advantages and disadvantages of both approaches are extensively discussed in (4). Finite-difference models allow a finer resolution of flame temperatures and other furnace variables. They also need a much smaller amount of input data than zone methods since velocity distribution in the furnace is calculated simultaneously with heat transfer. However, because of the small size of control volumes, sophisticated turbulence models are necessary to predict turbulent exchange. 3-D applications of coupled finite-difference methods to boiler furnaces require still an enormous computational effort, yet a stable converged solution of the balance equations is not always guaranteed. The major disadvantage of available finite-difference models for boiler furnace performance predictions lies in the fact that it is notably the important radiative heat transfer which is approximated in a most approximate manner. In almost all of the finite-difference models reviewed, the geometrical and physical description of radiative transfer is simplified by use of so-called flux methods in such a way that levels and profiles of predicted temperatures can be doubtful. Thus, more sophisticated radiation models have to be coupled to finite-difference furnace models to make them a reliable tool for boiler performance predictions. This would increase computing times which are already large because of the fully coupled numerical solution procedure. An interesting new development is the discrete transfer method which promises a fast and fairly accurate computation in combination with finite-difference modeling (5). Zone models allow a more realistic simulation of directional radiative heat exchange in the furnace enclosure especially for complex 3-D boiler furnace geometries. This, of course, is achieved on the expense of resolution of the temperature distribution because of use of coarser control volumes. In most of zone methods the flow pattern necessary to solve total 2 |