Furnace design and operating practices, including furnace size and shape, and burner design and location, strongly influence the performance of aluminum smelters.

Fig. 1 Technology justification based on production increase.
The aluminum industry experimented with oxygen-based combustion as early as 40 years ago. These trials were, however, largely unsuccessful due to the nascent state of the art. Today, the industry is re-evaluating the concept because of growing production demand and advancements in oxygen-enhanced combustion technologies.[1] Although sufficient evidence has been accumulated within the industry to prove the technical feasibility of oxygen-based aluminum smelting, the key remaining issue is whether the conversion to oxygen is economically justifiable. Our experience shows that, in the majority of cases, oxygen-based combustion provides substantial financial benefits to aluminum makers.

Side-Well Reverberatory Furnaces
Reverberatory furnaces are widely used for melting aluminum in both primary and secondary facilities. A typical side-well reverberatory furnace consists of a number of burners firing inside the hearth, against the furnace hot wall or door. A charging well and a pump well, when present, are attached to the furnace hot wall on the outside of the furnace. Both wells are connected to each other and with the furnace hearth by arches, which permit aluminum circulation between the furnace chambers.

During the charging period of the furnace cycle, molten aluminum circulates through the charge well where it melts the scrap, then circulates back into the hearth for reheating. Charging continues until the molten bath reaches the desired level. The subsequent steps in the smelting process are degassing and alloying. These are often accomplished in the same furnace during the holding period and can consume from 25% to 50% of the total furnace cycle time. The molten material is then tapped into a casting line, crucible, ladle, or into a holding furnace for final processing.

The efficiency and productivity of a side-well furnace are dependent on the ability of molten metal to receive thermal energy released within the combustion chamber. Several important factors influence this process. First, a layer of dross forms at the molten bath surface inside the furnace hearth. This layer acts as a thermal insulator, restricting heat transfer between the combustion space and the melt. It should be noted, however, that a thin layer of dross is essential to prevent excessive aluminum oxidation at the gas-melt interface. A second factor is that the rate of heat transfer to the charge material increases with increasing molten metal circulation rate. Hence, furnace efficiency is generally improved by the addition of a liquid metal pump. Finally, flame characteristics have a great influence on the rate of heat transfer to the melt. High momentum air-fuel burners with relatively transparent (non-luminous) flames deliver less energy to the melt per unit time when compared to highly luminous, higher temperature oxygen enriched flames.[2]

Direct-Charged Reverberatory Furnaces
The direct-charged reverberatory furnace[3,4] differs from the side-well furnace by the absence of an outside charging well. Metal is charged directly into the furnace hearth where it is exposed to the open flames. The metal recovery of such a system is usually lower than a side-well furnace, but these furnaces are nevertheless widely used for clean scrap because of their relatively compact size and high melting efficiency.

After material has been charged into the furnace, the door is closed, the burners are turned on to their full fire setting and the melting process begins. As the solid melts, the bath height increases while the solid pile height decreases. Heat is transferred into the solid metal from both the furnace combustion space and from the molten phase. Once melting of the charged pile is complete, the burner firing rate is lowered and, if necessary, dross can be skimmed out from the furnace hearth through the door. The charging procedure then starts again, and is repeated until the molten metal reaches its required level. Refractory life in a direct-charged furnace is often an issue because of the greater potential for mechanical damage and higher thermal stresses.

Economics of Oxygen-Based Combustion in Aluminum Reverb Furnaces
Radiation heat transfer is dominant in oxygen-based combustion due to higher flame temperatures and emissivity. This leads to more rapid load heating and, hence, increased furnace productivity. Productivity increases of the order of 30 to 40 percent have been achieved.[5] Moreover, removal of nitrogen from the combustion space can lead to fuel efficiency improvements in excess of 40 percent. These factors provide economic incentives toward the use of oxygen that must be weighed against the added cost. An economic model was developed to quantify these trade-offs, thereby providing the basis for an informed decision concerning the conversion of an air-fuel furnace to oxygen-based combustion technology.

Calculations were carried out for a typical side-well aluminum reverb furnace having a holding capacity of 120,000 lbs and 90 percent yield. Two oxygen-based combustion technologies were evaluated relative to air-fuel (AF); air-oxy-fuel (AOF) with 35 percent (by volume) oxygen in the oxidizer, and 100 percent oxy-fuel (OF). Key economic and operating parameters used for the three cases are provided in Table 1. Sample economic results are also given in this table for the case of 30 percent productivity improvement using oxygen-based combustion technologies.

Calculations assume a fixed cost for air-fuel operation equal to $4,800 per day; this cost remains constant for the oxygen-based technologies. Additional assumptions include an oxygen price of $0.28 per 100 scf; natural gas price equal to $4.00 per thousand cubic feet; continuous furnace operation with 75 percent of the operating time used for scrap melting and 25 percent used for holding, degassing and tapping. The profit of air-fuel operation is taken as $0.01 per pound of aluminum produced. Variable costs such as salt, disposal, electricity, etc. per pound of scrap processed are assumed to be constant for all three technologies.

To illustrate the calculation, the profit per pound of aluminum produced for the AOF technology ($0.0204) is equal to the profit per pound for AF ($0.01) plus the difference between the total daily cost of AF and AOF per pound produced [5,856/(0.9x120,000) - 6,153/(0.9x156,000)]. The monthly benefit ($53,525) is equal to the product of the aluminum production rate and profit for AOF (0.9 x 30 x 0.0204 x 156,000) minus the corresponding term for AF (0.9 x 30 x 0.01 x 120,000). It is important to note that while both oxygen-based technologies provide substantial economic benefit relative to air-fuel operation, the monthly benefit of AOF is approximately $14,000 higher than for OF. This illustrates that, at equal production levels, the reduced oxygen cost of the AOF technology far outweighs the slightly higher fuel efficiency of the 100 percent oxy-fuel operation.

The same calculations were carried out over a range of productivity improvement (relative to AF) from 5 to 50 percent and are summarized in Figure 1. Results show that AOF technology provides a significant economic advantage over conventional air-fuel operation (approximately $10,000/month) even with a production increase as low as 5 percent. The economic benefit of AOF approaches $75,000 per month at the higher end of the productivity range, while the advantage of AOF versus OF is consistently in the $15,000/month range. The actual productivity increase will naturally depend on furnace design, refractory limits and operating practices. Our experience has shown that, in retrofit applications, productivity increases from 20 to 40 percent are readily achievable. Moreover, we have found that a relatively low level of oxygen enrichment (35%) provides a productivity boost comparable to that attained using 100% oxy-fuel.

Direct-Charged Furnace Economics: Importance of Furnace Yield
The economics of oxygen usage is similar for direct-charged reverb furnaces except that furnace yield (i.e., material recovery) is more sensitive to the combustion technology used than in side-well furnaces. Oxidation rates increase with surface metal temperature and oxygen concentration near the charge. The direct flame impingement that occurs during the melting period has the potential to result in significant material loss if the combustion system is not properly designed. A head-to-head comparison of AF, AOF and OF burner technologies in a direct-charged furnace showed differences in yield of the order of 2 percent, with AOF delivering the highest yield, followed by AF, then OF. To put this into economic perspective for a furnace producing 156,000 lbs aluminum per day, assuming scrap is purchased at $0.50 per pound, a one percent improvement in furnace yield translates into a profit increase of almost $25,000/month. This additional benefit further strengthens the economic justification for AOF technology.


The conversion of air-fuel based aluminum smelters to oxygen-based combustion systems has a strong track record of increasing furnace productivity. Economic calculations indicate that these productivity improvements translate into substantial economic benefits to aluminum manufacturers. Emissions of nitrogen (NOx), CO2 and particulates are also substantially reduced with oxygen-based technologies, introducing additional economic and environmental benefits. Air-oxy-fuel technologies can be more profitable than 100% oxy-fuel systems, are easier to install and provide added operating flexibility. IH