Modeling Helps Improve Performance In Industrial Heating Applications
Many high-temperature industrial heating processes suffer from poor temperature uniformity, inefficient or unknown cycle times, poor burner placement and inefficient heat transfer. In many cases, increases in furnace productivity, product quality and a reduction in specific fuel consumption can be achieved by improving burner positioning and understanding process fluid mechanics and heat transfer to a greater degree. The use of computational fluid dynamics (CFD) software, such as Fluent® (registered trademark of Fluent Inc., Lebanon, N.H.), is a valuable design and analysis tool for the heat processing engineer to accomplish these objectives. Two illustrative examples demonstrate the power, effectiveness and useful insight that CFD tools can provide in analyzing various industrial-heating applications.
Temperature uniformity, a key performance parameter of large multizone walking-beam furnaces, depends on burner performance and relative burner positions in the furnace proper. Typically, burners are positioned on side and front walls in large walking-beam furnaces. The effects of burner position and performance can be studied using CFD modeling.
One significant issue is the tendency of side burner flames to deflect from their centerline axis due to a cross flow of flue gases from previous furnace zones. The burner jet efflux is deflected as it exits the nozzle by the high-momentum flue-gas cross flow and subsequently follows a curved path downstream while undergoing a cross-sectional change. As the flow spreads laterally into an oval shape, the flue-gas cross flow shears the burner jet fluid along the lateral edges downstream in the direction of furnace flue-gas flow. At increased distances along the burner jet exit path, the shearing action folds the downstream burner efflux over itself to form a vortex pair.
Jet and cross-flow interaction is a long-standing research topic, which has generated an extensive list of publications . Previous research indicates that a cross-flow velocity of 15-20 ft/s (4.6-6.1 m/s) is sufficient to deflect the flame significantly. Further, results suggest that flame deflection could be reduced by redesigning burners to obtain increased momentum fluxes.
The analysis discussed here includes one heating zone of a large 12.5 x 48 x 138 ft (3.8 x 14.6 x 42.1 m) walking-beam furnace with a production rate of 336 metric tons per hour. The zone heat capacity is 108 MMBtu/hr (31.6 MW) in the upper section above the load, and 120 MMBtu/hr (35.2 MW) in the lower section below the load. Each zone has a total of eight burners, four on one side above the load and four on the opposite side below the load, which are staggered relative to the burners above the load. A Hauck high momentum Beta burner with converging tile was selected as the baseline burner to model the furnace zone. To increase the burner jet momentum, a natural-gas lance with an exit velocity of 500 ft/s (152 m/s) was inserted into the burner central port.
Fluent 5.4 CFD software was used to simulate both cases using the standard k-e species transport combustion and discrete ordinates radiation models. Figure 1 shows the temperature field in the cross-sectional area through a burner located above the load centerline axis. Both cases show a significant burner jet flame deflection in the direction of flue-gas cross flow. The burner with gas lance injection shows a modest improvement of jet flame penetration into the furnace, while simultaneously reducing hot spots near the burner tiles. However, even with the high-momentum gas lance, there is insufficient penetration of hot combustion products into the central region of the furnace. Therefore, CFD modeling and analysis led to the conclusion that walking-beam furnaces having large widths and high-momentum flue-gas cross flows require a frontal burner installation for optimum temperature uniformity.
Continuous furnace to heat metal strip
Continuous direct-fired furnaces are used in a wide variety of industrial process heating applications. Typically, side firing is used for both the upper section above the metal strip being heated and the lower section below the strip. The side firing arrangement is attractive from the standpoint of creating control zones in the furnace. However, there are at least two feasible burner arrangements as shown in Fig. 2:
Case I: Half the burners are placed on one side of the furnace above the metal strip and half on the opposite side of the furnace below the strip; burners are staggered above and below with respect to each other.
Case II: Burners are staggered in an alternating arrangement (left side and right side fired) in rows above and below the metal strip; burners above and below the strip are in line with respect to each other.
Both cases were modeled in a pre-fire zone using Fluent CFD. Dimensions of the continuous strip pre-fire zone were 141 in. (3,581 mm) long by 95 in. (2,413 mm) wide with a height above the strip of 27 in. (686 mm) and a height below the strip of 22 in. (559 mm). The metal strip was 71 in. (1,803 mm) wide and 0.17 in. (4.3 mm) thick. The modeled furnace zone was equipped with 8 Hauck model SVG140-HR high-velocity gas burners having a total zone heat capacity of 14 MMBtu/hr (4.1 MW).
Figures 3 and 4 show the temperature distributions for both cases in horizontal cross sections via the burner centerlines. The two cases provide significantly different flow patterns and flow exchanges between the upper and lower sections, above and below the metal strip, respectively. Case I shows significant jet flame impingement on the furnace walls compared with Case II, which shows virtually no jet flame impingement. The flow pattern in a furnace cross section through the centerline of two burners (one below and one above the strip as defined in Case II) is shown in Fig. 5. There is a large volumetric flow exchange above and below the metal strip; subsequent Fluent analysis indicated the flow exchange is 60 % higher in Case II than that in the Case I burner arrangement. These flow patterns and temperature distributions have a direct impact on the metal strip temperature uniformity as indicated in Fig. 6. The Case II burner arrangement results in a temperature uniformity across the metal strip of +/-15 F (+/-8 C) versus +/-25 F (+/-14 C) in Case I.