| OCR Text |
Show In addition to the flame stability issue, flame geometry was of an equal importance to this application because there are differences on burner arrangement as well as furnace exit configuration . between Grayson Units 3 and 4. Therefore, as pan of the development wor~ a number of fuel injector configurations were also briefly examined. As discussed above, LFG was simulated by a blend of natural gas and nitrogen, such that the heating value of the fuel blend was approximately 300 Btu/scf. 5.1.1 Screening Test The screening tests consisted primarily of flame observations. The initial test using natural gas as support fuel resulted in a total disappointment. The -as-installed - injector configuration (Configuration A) failed to ignite at any amount of support fuel. Furthermore, it induced instability in the natural gas support flame. In view of these results, the injectors were rotated such that the fuel discharge faced tangentially to the primary swirler (Configuration B). This setting resulted in a much improved performance at the light-up condition. Marginal flame stability was achieved throughout the entire load range, including operation with an FGR rate of approximately 25 percent. Configuration B required a minimum support fuel of approximately 8.6 MMBtulhr to maintain flammability. This minimum amount of support fuel represented approximately 16 percent of total heat input. In spite of the stable flames achieved by this configuration, the flame geometry was not totally acceptable. In general, the flame shape could be characterized as -fingery- with an appearance which resembled a jet of hazy iDcandescent gases. Moreover, flame tails impinged frequently OD the side walls and floor of the furnace causing the CO to spike. As a result of these observations, a second set of injectors wu placed in service (Configuration C). The flame impingement previously seen was no longer evident. In addition, this configuration appeared to enhance the overall flame stability, since it -14- |
