OCR Text |
Show Paper No. 27 Coal-Fired Pilot Plant Furnace Computer Simulations Philip J. Srruth, Paul A. Gi1lis, and Knute R. Chrjsten~en Combustion Computations Laboratory 75 erB, Brigham Young University Provo, Utah 84602 Coal combustion mathematical model simulations an presented for pilot plant and laboratory furnace tEStS. ThE. furnace gEomdries rtprLsent full-scale wall-firtd and tangentially-firtd utility boiler applications. Laboratory furnaCE studies explort thE near burner performance of industrial operation. ThE mathematical modeling calculations dEmonstrate the ability of currtnt computer and engineering technologies to perform a priori pf'fdictions of local, in situ furnaCE properties. Comparisons of three cloSUri schemes for fluid turbulencE shows the three dimensional k-f model to bE significantly better than a constant eddy diifusivity or simple mixing length rnode! but still somewhat inadequate for 3-D furnace applications. FinE scale numerical grid resolution is shown to be neccesary to simulate obsfrved flow patterns in furnace geometries, even for relatively simple fU11lace configurations. ThE tight coupling bdween local heat transfer and other physico-chemical processes occurring in coal combustion applications is emphasized. Differences of 50-70% are shown to be obtained if parlicltgas convective/conductive heat exchange is ignored or if turbulent fluctuations are not accounted for. Industrial furnace geometries are shown to rt,qu2're advanced numerical treatment to achieve accurate and efficient solutions. A 12 increase in efficiency of two orders of magnitude is sho'U.'Tl to be achieved by using multi-grid mdhods. 1 Introduction \Vith con1putational abilities increasing each year due to technological ad\'ances in comput.er hardware and software, and with impro\ed numerical capabilities: t.he oportunity of obtaining detailed local information from computer sin1ulators of industrial coal-fired furnaces seems within reach. Although computations haye been demonstrated and evaluat.ed over the past ten )ears for sn1aller laboratory furnaces the ability to demonstrate full comprehensive calculations of multiple-burner, industrial geon1etries has only been investigated quite recently [1,2,3,4]. Even these calculations have often been performed with simplified su bmodels for turbulence, reaction, and heat transfer and have employed very coarse grid structures. Very little eyaluation of accuracy of these predictions has been done to date. ~1any questions ren1ain regarding nunlerical accuracy as welJ as the impact of various su bprocesses and the coupling between them on the con1putations. Of particular interest is the coupling between the turbulence and the chemistry and heat tran sfer processes. 1 |