| OCR Text |
Show suitable for glass furnaces (Ref. 12). The arrangement of the burners is shown in Fig. 4. Model ° Genmetry for End-Section Fer reasons described in an earlier section, jt: wa~ necided to conduct the 3-D modeling study of the combustion space for only one section of the furnace, namely the end-section. The space covered by the model, which is indicated in Fig. 3, extends from the center-plane between the fourth and the fifth ports to the bridge-wall. Thus, the model assumes that heat transfer in the combustion space is symmetrical with respect to the vertical center-plane. In the graphical display of the results following later, this symmetry feature will be utilized by extending the plots of the furnace variables up to the center-plane of the fourth port. The 3-D zone arrangement used in the heat transfer model of the end-section is displayed in Fig. 4. The zone arrangement utilized basically consists of a 9x6x7 rectangular volume zones with some of the near roof zones distorted or omitted in order to simulate the curvature of the furnace roof. Thermal Boundary Conditions for End-Section The thermal boundary conditions considered for the endsection are summarized in Table 1. A uniform value of 0.5 was assumed for the emissivity of all refractory surfaces. The effective ratios of conductance to wall thickn~3s (k/s)eff as well as the outside surface temperatures for var10US furnace wall sections were deducted from ~ heat loss analysis carried out for air operation by a furnace manufacturer. In particular, the ratios (k/s)eff were determined so that the current zone model will predict the same overall heat losses, if the flame side surface temperatures coincide with those used in the heat loss analysis mentioned above. Table 1 also lists the view factors and effective regenerator temperatures used in Eq. (5) to calculate the port losses for air and 02 combustion, respectively. The mean values for Trq eff used in the air case are also supported by measurements of gas and refractory temperatures in the upper regenerator carried out over the whole firing cycle. The boundary conditions at the glass surface were prescribed as discussed in the general description of the approach. In particular, two sets of calculations were carried out. The first set of calculations was carried out assuming a uniform effective 9 |
