OCR Text |
Show MATHEMATICAL MODELING OF SWIRLING PULVERIZED COAL FLAMES: WHAT CAN COMBUSTION ENGINEERS EXPECT FROM MODELING? Roman Weber, Andre A. F. Peters and Peter P. Breithaupt International Flame Research Foundation Umuiden.The Netherlands Martien B. Visser Gasunie Research Groningen.Tne Netherlands ABSTRACT The present study is concerned with mathematical modeling of swirling pulverized coal flames. The attention is focused on the near burner zone properties of high- and low-NOx flames issued from an Acrodynamically Air Staged Burner of 3.4 M W thermal input. The swirling combusting flows are calculated using the k-e model and second-order models of turbulence. The Eulerian balance equations for enthalpy and mass fractions of oxygen, volatilcs, carbon monoxide and final combustion products (CO2+H2O) are solved. The Lagrangian particle tracking is accompanied by appropriate models of coal devolatilization and char combustion. Nitric oxide emissions are calculated using a N O x post-processor for thermal-, prompt- and fuel-NO. The objective of this paper is to examine whether the engineering information required for designing industrial burners is obtainable through the mathematical modeling. T o this end. the flame computations, including N O emissions, are compared with the measured in-flame data. The guidelines as to the combination of physical sub-models and model parameters needed for quality predictions of different flame types are given. I N T R O D U C T I ON Manufacturers and users of coal combustion equipment are making significant efforts to design efficient and reliable combustion systems which would meet the present emission standards or even more stringent forthcoming regulations. Modifications to and retrofitting of the existing equipment need to be carried out without reduction in overall process efficiency and reliability. Both tasks require careful considerations of the combustion chamber, burners, heat extraction unit and flue gas cleaning equipment. A n integral part of the combustion system is either a single flame or more frequently a row of flames. Engineering information required for designing flames suitable for a particular process can be classified into: First-order information: - knowledge of flame shape and length, and an estimate of flame temperatures with accuracy of around 200"C. - estimate of heat fluxes to the heat extraction unit with accuracy of around 30-40%, - location of regions of high- and low-mixing intensities; Second-order information: - knowledge of temperature distribution (accuracy within 100'C), oxygen concentration (ace. 0.3%) and unburned fuel (ace. 0.3%), - identification of furnace and burner zones of slagging potential; Tliird-order information: ; A ^ '' : >-•r' - I - - knowledge of flue emissions of nitrogen oxides, carbon monoxide, sulphur oxides, soot, Polycyclic Aromatic V Hydrocarbons (PAH) and char burnout, - knowledge of in-flame temperatures (ace. 50°C) and detailed chemistry including identification of regions of high pollutant formation rate. Computational methods in fluid flow, heat transfer, combustion and pollutant processes have advanced to the point that such methods arc of an assistance to combustion engineers searching for confidence in the design. Simulation of pulverized coal combustion involves modeling of a number of complex. interdependent and simultaneous processes (Smoot, 1993; Wall, 1987). The computational results depend on the quality of the numerical solver and the physical sub-models used, their implementation and coupling. Of primary importance is the selection of parameters of the physical sub-models. Until now there is no universal model of pulverized coal combustion, which would predict with the same good accuracy, all types of flames. The objective of this paper is to examine whether the three-level engineering information listed above, is obtainable through mathematical modeling. The intention is to provide guidelines as 10 the combination of sub-models and model parameters needed to -77 1 1 - 11 |