There are a number of conventional solutions to increasing output and product flow rate in a standard air/fuel combustion furnace. Solutions include increasing the length of the furnace and the number of burners, improving flue gas heat recovery to improve preheating, installation of alternating regenerative burners to preheat combustion air and charging hot billets directly from a previous production operation. Major features of all of these are high investment and length of downtime required to refurbish or refit the furnace. In addition, preheating air can result in unwanted high NOx emissions and an increase in scale (oxidation) on the heated billets.

The Pure-Oxygen Solution

The use of pure oxygen instead of air for combustion has both efficiency and product quality advantages including:

  • Improved thermal homogeneity of heated billets (including thick section billets) resulting in less deformation and a more uniform temperature throughout the billet Reduced surface oxidation (scaling) despite the oxidizing nature of the process
  • Improved efficiency of an existing heat recovery system (by not heating nitrogen in the air)
  • At least as good as energy efficiency as the highly preheated air solution, together with minimized NOx emissions
  • Decreased use of natural gas per ton (although outweighed by an increased use of oxygen)
  • Low investment due to unchanged furnace configuration, limited adaptation of work to the burners and minimal downtime

An oxyfuel flame has an enhanced capability to transfer heat energy by radiation so direct energy transfer to load and cells is much improved. With the smaller volume of flue gases, induced losses are automatically smaller. Table 1 shows the consumption of fuel and combustion constituent and the generation of fumes in a system at 900?C (1650?F), which transfers an effective 18 MJ (5 kWh) of power to the load.

A properly located oxyfuel flame contributes to more uniform heating of a furnace resulting in fewer hot and cold spots (e.g. high vault temperatures and lows in charging zones). This speeds up heating of large section steel product, reduces delays, decreases temperature differentials along or through a product, and, thus, reduces distortion and scaling. Product quality also can be better maintained due to the more homogenous condition of the product when undergoing rolling.

While the efficiency of using high-temperature preheated air for combustion is similar, high-temperature combustion in the presence of nitrogen leads to the risk of high NOx levels and requires an investment in flue-gas scrubbers to meet emission regulations. By comparison, the use of pure oxygen minimizes nitrogen in the combustion process.

Fig 1 Modeling an existing furnace heating configuration predicts the improvement in furnace and product temperature uniformity through the addition of oxyfuel burners. This model demonstrates the closeness of fit between the predicted and actual temperatures of the furnace wall and the surface temperature of a billet in a trial run.

Modeling Predicts Quality and Efficiency

Air Liquide's advanced modeling techniques enable the evaluation of an existing plant configuration to determine how it will improve through the addition of oxyfuel burners. Figure 1 demonstrates the closeness of fit between the predicted and actual temperatures of the furnace wall and the surface temperature of a billet in a trial run.

Three-dimensional modeling techniques enable fine tuning of the process, such as influencing burner location and control of the flame size and temperature. Modeling also assists predicting furnace temperature throughout the furnace, thus allowing optimization of temperature to improve product heating control.

Fig 2 Air combustion furnace at a power of 29 MW produces a temperature differential across the billet (from top surface to bottom surface) of 50¿C (90¿F); ideal billet temperature is indicated in the middle (green) range.

The difference in temperature homogeneity of a billet achieved in air combustion and in oxyfuel-assisted combustion of the same overall power is demonstrated in Fig. 2 and 3. The overall temperature differential is halved in the oxyfuel process.

Fig 3 Air combustion (26 MW) plus oxyfuel combustion (3 MW) produces a more homogenous temperature through billet than air combustion alone with a temperature differential of 25¿C (45¿F); ideal billet temperature is indicated in the middle (green) range.

Improved Efficiency: Case Study

LME Trith Saint L?r (Valenciennes, France), Air Liquide and the French Agency for Energy and Environment Management (ADEME) collaborated in a project to demonstrate the improvement in a walking-beam billet reheating furnace by switching to oxyfuel-assisted combustion. The 31 m (102 ft) long natural gas-fired furnace comprises a charging zone, a long non-fired recovery zone, a preheating zone (10.5 MW), a main firing zone (15.5 MW) and a soaking zone (4 MW). A typical exhaust zone temperature of 700?C (1290?F) enables preheating combustion air to 300?C (570?F), while the target reheat temperature is 1100?C (2010?F). Temperature regulation of each heating zone provides the major furnace process control.

Two options preselected during preliminary studies were global enrichment of combustion air to a maximum of 24% oxygen through existing burners and the addition of a new oxyfiring zone, using oxyfuel burners.

A model of the furnace (global energy balances) is shown in Table 2, which shows the expected output gains using oxygen solutions. With oxyfuel burners, the long recovery zone in the LME furnace means that exit temperature and the small increase in fumes volume will not materially affect the energy recovery performance. Because an increase in productivity was one of LME's primary objectives, four oxyfuel burners were installed on the reheating furnace in the heat recovery zone immediately after the existing air combustion zone.

The oxycombustion zone can be regulated by firing two or four burners, as well as through regulating the gas flow to the burners. This allows the oxycombustion to be used when required to help the throughput of thick products, which otherwise would cause bottlenecks and restrict output.

The efficiency of recovery changed very little as indicated by an increase in exhaust temperature of only 20?C (35?F); the energy from oxycombustion is transferred to the product. Visual observation in the discharge zone shows less scale on the billets, despite the oxidizing nature of the process. Reasons for this are the low temperature (900?C, or 1650?F) in the oxyfiring zone and a shorter time for the product in the high temperature zones. A side benefit is that scale formed under primary oxidizing conditions is more porous and easily removed.

NOx emissions are reduced by up to 15 %. This is achieved both through empirical optimization of the oxygen/fuel gas mixture and by matching the number of oxyfuel burners under combustion to the total power output of the furnace.

Productivity Increases

LME increased output by 20% (92 to 110 t/hr) at the same average cost per metric ton, which allowed the company to pay off the capital cost through these savings. Operating cost savings of the reheated, rolled and finished product are shown in Table 3 with the previous air combustion process costs given as 100. A saving of 12% per metric ton of finished steel was achieved, which is a significant cost advantage for a steel producer.

For more information: Daniel Levert is corporate foundry market manager, Air Liquide, 75 Quai d'Orsay, Paris 75007; tel: ++ 33 (0) 1 40 62 53 05; fax: ++ 33 (0) 1 40 62 54 36; e-mail: Mike Grant is senior metals applications engineer, Air Liquide America Corp., 5230 South East Ave., Countryside, IL 60525; tel: 708-579-7859; fax: 708-579-7858: e-mail:

This article is derived from a paper presented at AISE September 2000 by O. Delabroy and G. Le Gouefflec (Air Liquide) and C. Lebrun, A. Barbotin and R. Cervi (LME). Proceedings available from AISE.