The ALUREC Process
The ALUREC Process
Henrik Gripenberg and Hans Grab, AGA AB Gerhard Flesch, MAN GHH Michael Mllerthann, Hoogovens Aluminium Httenwerk GmbH
Abstract To solve the environmental and financial problems of the traditional Rotary Salt Furnace process, a new saltfree method for aluminum dross and scrap melting, ALUREC, has been developed by AGA and its partners, Hoogovens Aluminium and MAN GHH. AGA and Hoogovens Aluminium developed the main process concept and jointly carried out trials on pilot plant sea le. Owing to excellent results, the partners decided to jointly erect a commercial scale test plant. The order for the design and erection of the furnace went to MAN GHH. The three companies designed tl1e test plant and began operation in the spring of 1994. The results have been excellent and the process is expected to be a technical and commercial success. lntroduction Dross is an undesirable by-product of all processes involving molten aluminum. The oxide layer formed on the surface of the molten metal has to be removed in order not to influence the quality of the final product. Depending on process conditions, a dross amount of 2-10% of the produced aluminum will be formed. The dross consists of oxides of aluminum and alloying elements, aluminum nitride, entrapped metal and a small amount of fluxes. The metallic content depends on process procedures and is normally 30-60%, although higher and lower contents may occur. The dross represents a huge aluminum source to the industry, in the magnitude of one million metric tons a year in Europe and the Americas. The industry has adapted several methods to minimize further metal oxidation in the hot dross after skimming and to recover the metallic aluminum. Examples of techniques to reduce oxidation are forced cooling in a water cooled rotary drum or dross cooling under a protective argon shield. Mechanical separation systems can be used to recover coarse metal particles in the dross and to upgrade the aluminum concentration by separating fines with low metal content. However, the dross treatment loop ends up in a process for melting of the entrapped metallic aluminum and separation of the liquid metal from the oxides, referred to as dross melting. This paper deals with this process step and the solution of related by using the recently introduced ALUREC process. Dross Melting-Process Considerations The Rotary Salt Furnace The most widely used dross melting process is the Rotary Salt Furnace (RSF) process. These aces may vary in shape and layout. A schematic view of a usual layout is in figure 1. The furnace consists of a horizontal, refractory-lined cylinder. The cylinder is rotated during the process. The energy is supplied by an air-fuel flame
entering at one end. The exhaust gases leave through a duct at the opposite end of the furnace. The air-fuel flame transfers energy directly to the charge through radiation and convection and indirectly through conduction from the refractory wall during rotation. The fuel consumption is about 1000 kWh/t (1600 Btu/lb) of dross.
Figure l. The Rotary Salt Furnace (RSF) Stirring and mixing during rotation of the furnace (1). The amount of flux added depends upon the oxide content of the dross. To give an example about 500 kg flux would be added to 1000 kg of dross containing 50% aluminum metal. At the end of the RSF process, the separated molten aluminum is at the bottom of the furnace and the liquid non-metallic product (NMP) mix of salts and oxides floats on top. The aluminum and the NMP are removed from the furnace through separate tap holes. A major drawback with RSF is the non-metallic by-product. lt is referred to as salt cake, a mixture of oxides, aluminum nitride and metal and salts. More than one ton of salt cake is produced for every ton of dross. The salt cake is a growing environmental problem. In the past this residue was dumped, as is still the case in many countries. However, due to environmental concerns, the costs for disposal have increased dramatically in many countries, and in some countries disposal is even prohibited. Processes based on dissolving the salt and re-crystallizing to recover the salt and o rides exist. However, these involve high energy inputs and investments. Other drawbacks are that salt vapors emitted from the furnace and that the salt requires energy for melting and is costly to purchase, transport and store.
To solve the environmental and financial problems of the RSF process, new salt-free methods have been suggested and developed.
Electric energy-based process The process solutions based on electric energy supply are characterized by operation without flux in a rotary and tiltable converter shaped furnace. Furthermore the NMP is solid at the end of the process. The metal is tapped through a tap hole and the NMP is discharged by tilting the furnace and by mechanical scraping of the furnace walls. A process using a plasma torch for the transfer of electrical energy was developed by Alcan (2). Two commercial plants exist. The first plant was constructed in 1989 (3). A method for supplying the electrical energy to the dross by means of an arc between graphite electrodes has been tested in laboratory scale by LTEE, Hydro-Quebec. The method is claimed to be more efficient than the plasma process, and the arc system is said to require less specialized personnel for operation and maintenance as well as less capital (3). The process will be tested in a plant due startup in 1995. Drawbacks for both electric energy-based processes are the significant investments in the electric energy supply systems and the need of specialized personnel for operating these systems.
The selection of energy supply system The use of electricity as the source of energy was regarded as too complicated and costly at an early stage. Instead, efforts were focused on fuel based energy, and oxygen-fuel burner technology was selected. The background fort the selection of oxy-fuel will be discussed further with reference to table 1, where some important characteristics of fuel flames are shown and compared with electricity generated plasma flames. The oxy-fuel burner and plasma flame temperatures are higher than that of an air fuel flame. This means that the energy transfer by radiation is much higher compared to the air burner flame since radiation is proportional to the fourth power of the temperature.
The exhaust gas from the oxy-fuel burner does not contain nitrogen. Air-fuel burners and plasma burners (air or nitrogen based), on the other hand, produce exhaust gases with high concentrations of nitrogen. The presence of nitrogen in the flame provides a potential for the formation of aluminum nitrides that would contaminate the rest product from the process and could affect the aluminum yield. The presence of nitrogen in the air plasma also promotes the formation of NOx. Compared to an air burner, the stoichiometry of the furnace can be much better controlled. The volume of exhaust gas from an oxy-fuel burners is small, only about 25% of that of an air-fuel burner. This results in a smaller flame, and direct contact with the charge can be avoided. There will be less convection and less mass transfer between the oxy-fuel flame and the charge. Furthermore, the thermal efficiency of an oxyfuel heated furnace is much higher due to the small exhaust gas volume. Thos reduces the energy consumption compared to the use of air-fuel energy. Considering the small exhaust gas volume, the exhaust gas composition and the possibilities of accurately controlling this composition with oxy-fuel, it was concluded that the aluminum oxidation in the process would be under control. While the RSF process uses a flux, ALUREC uses atmosphere control to prevent oxidation. The reduced exhaust gas volume in combination with the increased flame temperature results in a more energy efficient process. Compared with air-fuel, the energy consumption will be lower with oxy-fuel. The emissions of CO2 an SO2 will be reduced, since they are proportional to the fuel consumption. In addition, the formaation of NOx is reduced because nitrogen is excluded from the combustion process. Characteristic for oxy-fuel is the possibility of applying high power in relation to the size and cost of the equipment.
From the above considerations, we came to the conclusion that an air-fuel burner could not fulfill the demands. It was concluded that oxy-fuel technology would be the best solution owing to the following main characteristics of oxy-fuel: low mass transfer from the flame to the charge high heat transfer by radiation high thern1al efficiency Iow fuel consumption cost efficient low emissions Furnace configuration A rotary, tiltable converter type furnace is used for the process, figure 3. The oxy-fuel burner and the exhaust gas port are located on the same side. This design offers a high energy efficiency and very good possibilities to control the furnace atmosphere. The oxy-fuel burner heats the converter refractory wall in a short time to about 1000 deg C (1800 deg F). Through furnace rotation, the heat is transferred to the charge by conduction, and the heat is further distributed within the charge by mixing. Heat is also transferred through direct radiation from the flame to the charge. As the entrapped metal becomes molten, it agglomerates as the oxide film surrounding the metal droplets is broken by the mechanical mixing caused by the rotation of the converter. The metal is collected at the bottom of the converter, and the solid NMP floats on top. The metal is tapped separately from the NMP directly into melting or holding furnaces or is cast into sows or Tingots, for example. As no flux is used, ALUREC operates with a dry solid (slag) NMP that cannot be removed through a tap hole as in the RSF. The furnace is equipped with a tilting mechanism, and the NMP is discharged through the converter mouth. Depending on dross composition, about 500 kg of NMP is produced for each metric ton of dross. This is about half the amount of rest product compared with the RSF process. The NMP from the ALUREC process will be re-used in other industries. The ALUREC process will be a closed system without the need for dumping of solid wastes.
ALUREC Pilot Plant Test At the beginning of 1992, AGA and Hoogovens Aluminium carried out pilot plant tests. The work was done at the Foundation for Metallurgical Research, MEFOS, in Lulea, Sweden, where AGA is one of the principal funders. The universal converter at the foundation was used. This converter has a useful volume of 1 m3 (35 ft3).
The drosses were supplied from Hoogovens Aluminium Huttenwerk GmbH in Voerde, Germany. The dross has been raked from the melting furnaces into an insulated bucket and had been treated in an AROS rotary cooler system. The dross has a bulk density of 100-1200 kg/m3 (69-75 lb/ft3). The converter allowed about 1.2 metric tons (2600 lb) of dross to be charged. Test were carried out according to the process design described a hove. At the end of the campaign, stable and reproducible results were obtained. lt was concluded that aluminum metal could be extracted from the dross without the use of flux salts. From eight comparable tests, the metal recovery, defined as metal cast/dross charged, was between 45-60% with a fuel consumption in the range of 200-400 k Wh/t (310-620 Btu/lb) of dross.