OCR Text |
Show An example of a challenging distribution problem is the utility boiler shown in Figure 7. Many utility boilers are designed with operating temperatures in the lower furnace region much higher than suitable for the NOxOUT chemistry. Flue gas velocities for these units can be more than 30 meters per second, providing very little residence time for droplet distribution and evaporation, and for chemical reactions to occur. One possible solution is to use a lance containing a number of nozzles as shown in Figure 8. According to model predictions, the lance can be used to inject droplets in gaps between tube banks where the flue gas temperature is in the appropriate range for the NOxOUT reactions. If the droplets are very small, they evaporate quickly, but are unable to penetrate very deeply due to the high flue gas velocities. Larger droplets penetrate farther, but do not evaporate quickly enough to allow residence time for reactions. Fine droplets injected at high velocities meet the design requirements. Field trials with these lances are examining their suitability in these stringent conditions. Another furnace type is the municipal solid waste combustor, shown in Figures 9 and 10. Figure 9 shows the flue gas velocities in a vertical cross section. Secondary air was included at the left and right indentations, and the effects on flue gas velocity are clearly visible near the left indentation. The secondary air produced small recirculation zones above the indentations. These zones tend to have lower temperatures and higher chemical concentrations according to model predictions. It is recognized that the model predicts probable steady state conditions, and turbulence and flow variations in the actual unit would tend to disrupt these zones. A characteristic of waste incineration is large variations in fuel quality and heating value. Figure 10 shows temperature profiles for the hypothetical case of poor combustion on one section of the grate. In this case, the lower temperature gas mixes with the surrounding high temperature gas, but a cool zone is predicted to extend around the right indentation. If too little heat is released in this poor combustion zone, the gas temperature at the injection zone could be too low, reducing the effectiveness of NOx reduction and potentially producing ammonia slip. This example demonstrates the utility of a model in predicting performance under varying operating conditions. In the final example, the flow and temperature in a stoker-fired wood furnace are shown in Figures 11 and 12. The model predictions showed significantly higher temperatures and flue gas velocities within the furnace than was thought to be present. When field performance was not as effective as desired, the actual temperatures were measured using suction pyrometry. These tests showed the temperatures to be considerably higher than previously known, and the NOx levels achieved were consistent with those predicted for the higher temperatures. In this and other cases, the CFD model has been used to improve performance by identifying potentially undesirable operating conditions, and in this particular case the field measurements gave excellent model verification. These examples are a representation of the many types of processes for which the NOxOUT Process can be applied. Most units are unique in their design and operation, so models are routinely performed as part of the engineering design for a project. Model improvements are continually being sought in order to improve -8- |