press room
   
 
CFD creates shapely flames for industrial boilers
 
   

François McKenty, Brais Malouin & Associates, Montreal, Canada
N. Brais, V-P. Product Development, Cleaver-Brooks, USA

 

Why do industrial flames look the way they do? If burner designs are symmetrical, then why do the flames have such irregular shapes? We know that the answer is of course that burners are fired within a furnace, and the geometry of this enclosure will have a great influence on the resulting flowfield developing inside it. Obviously the character of the flowfield will influence flame shape, but how strong this influence will be is always difficult to assess a priori. Most operational problems can be corrected by adjustments to the burners in the field, but the trade-off is often a loss of efficiency and/or increased pollutant formation.

 

This article aims to show how flame shape can be significantly affected by internal furnace aerodynamics. To illustrate this point we have chosen a study that was carried out using 125-MMBTU/hour NATCOM Low-NOx Hyper-Mix burners (figure1) fired in two 100,000-lbs/hour D-Type boilers. These D-Type boilers are mirror images of each other (figures 2a and 2b), but both burners are identical. The burner (figure 1) has a central low-velocity pilot gas injector, which is surrounded by a central air-core. The central air-core is itself surrounded by a variable-pitch blade air-swirler. Natural gas is injected through 8 high velocity gas spuds located around the swirler within the annular section of the axial-air by-pass. Four of the gas spuds are axial injectors surrounded by Hyper-Mix steam injectors and recirculated flue gas injectors (FGR). The other four spuds are fired at an angle in order to induce a counter-swirling motion of the gas.

Figures 2a and 2b show the flame shapes resulting from firing the burner in each of the two D-Type boilers. For two such boiler geometries that are mirror images of each other, one would also expect the flames to be mirror images. But this isn’t the case because the burners are identical rather than mirror images of each other; in fact both burners induce a clockwise swirling motion of the air and a counter-clockwise swirling motion of the gas. The resulting flames have irregular shapes with parts of the flame coming close to the tube-bank walls. If the burners had been fired in a free quiescent atmosphere, the flames would have been symmetrical about the burners’ centerline. Furthermore, if the burners had been fired within a symmetrical enclosure (ex. cylindrical furnace), the flames would still have been symmetrical although shaped a little differently. When fired within enclosures with irregular geometries however, the developing internal recirculation patterns become warped and contribute to distort the flame envelope. Figure 3 shows the vector field representing the flow patterns responsible for distorting the flame of figure 2a. The vectors colored in red show the envelope of the luminous flame, which is being stretched out towards the left wall by secondary currents that are not part of the central clockwise swirling motion of the flow leaving the burner.




Fig 3: UV - Velocity vectors at z=2.5m



Figs 4: Luminous flame contour- optimized burner configuration
 

A luminous flame contour represents the part of the flame emitting light that is visible to the naked eye. This luminosity of the flame is mostly due to unburned carbon particles that have been heated to the point at which they become incandescent. Natural gas flames, however, emit very little light in the visible spectrum (the flames are light blue), when burning in fuel-lean to stoichiometric proportions and so the absence of a visible contour does not imply the absence of a flame. In fact, the highest flame temperatures are reached for near-stoichiometric mixtures and so it is also useful to examine temperature profiles in order to ensure that no hot spots exist on the furnace walls. The temperature contours of figure 7 show that the optimized flame is indeed well centered and does not impinge on any of the furnace walls.

In summary, CFD methods, applied using STAR-CD, have allowed us to observe and quantify the details of industrial combustion phenomena that are almost impossible to measure. We have shown how CFD tools can enable us to better understand the underlying physics and so develop more efficient combustion technology.

 

 

Fig 1: Burner geometry





Figs 2a and 2b: Luminous flame contour - 2a standard configuration - 2b mirror image configuration


While it would be extremely difficult to control these secondary recirculation patterns, CFD does allow us to perform an in-depth analysis of the flowfield and of the factors affecting it. The STAR-CD simulations enable us to determine what is causing the flame to become distorted. It is thus possible to identify which spuds are responsible for the fuel burning close to the left wall and to identify which modifications to the burner’s standard configuration will eliminate or reduce the distortion. Figure 4 shows the luminous flame contour for the optimized burner configuration where the injection angle of some of the high-velocity gas spuds has been changed by rotating the pipes. This flame, although it still has a non-symmetrical profile, is now well contained in the center of the furnace. Another view of the luminous flame contour for the optimized burner configuration is shown in figure 5.

Figure 6 gives us an idea of the complexity of the flowfield inside the furnace. A strong swirling flow can be seen developing in an annular pattern in the center of the furnace. Recirculation patterns are present close to the floor and roof of the furnace, while no recirculation patterns can be seen developing in the horizontal plane at the elevation of the burner’s centerline. Had the enclosure been cylindrical, the recirculation pattern would have been toroidally shaped thus surrounding the flame. The shape of a flame developing in a non-symmetrical enclosure is clearly strongly influenced by the shear forces created by the non-symmetrical recirculation patterns surrounding it


Fig 5: Luminous flame contour- optimized burner configuration

Fig 6: Velocity vector field - optimized burner configuration

Fig 7: Temperature Contours - optimized burner configuration
 




 
 
 
OpenCube Drop Down Menu (www.opencube.com)