Designs and operation of melting furnaces differ markedly.

Designs and Operation of Melting Furnaces Differ Markedly

Melting furnaces date to the invention of the cupola by John Wilkinson of England in 1794. But while the cupola still accounts for over 60% of iron production, it is just one of many melting options available. In less than 200 hundred years, incredible advances in melting technology and metallurgy have been spurred by the call for lighter, stronger, and less expensive castings.

The Cupola

As shown in Fig. 1, the cupola is essentially a vertical steel shaft with a refractory-lined well supported on steel legs. The shaft may either be fully or partially refractory lined. Combustion air is introduced through a windbox and augmented by blasts of air through nozzles called tuyeres.

Traditionally, the refractory-lined "well" where the melt settles was built over a sand base. An initial charge of coke is ignited in the well, and the temperature increased by blowing air directly into the cupola through the tuyeres.

The melt charge, consisting of metal, limestone and coke, is introduced through a charging door located above the air inlets. An initial charge of coke is ignited in the well, and the temperature increased by increasing the amount of air forced through the tuyeres. A charge might consist of 2000 lb of scrap, 80 lb of limestone and 250 lb of coke. Ten thousand pound scrap charges are not uncommon, however, with limestone and coke additions increasing proportionately.

Limestone acts as a flux which combines with the coke ash, rust, dirt and sand returns forming a slag. The slag is withdrawn through an opening appropriately called a slag hole, located at the top of the well, just below the tuyeres and opposite the tap spout. Fluxing aids in reducing sulfur and lowers the melting point of the slag-forming deposits, making them easier to remove from the iron.

As a new addition of coke is consumed and the metal begins to melt, successive charges are added. Fully operational, cupolas exhibit an efficient countercurrent flow with the hot gases preheating the charge as they percolate up through successive charge layers and out the stack. Molten metal and slag are superheated as they settle to the bottom of the cupola well.

Melting can be controlled by either increasing the air flow and preheating the air, or by decreasing the flow and increasing the coke additions. Cupola capacity is directly related to its inside diameter at the top of the coke bed. But the melting rate for a given diameter is a function of the ratio of coke to iron, after the coke bed is established. The number and size of tuyeres as well as the air pressure determines the rate of burning.

Design Refinements--Traditional cupolas have been lined with fireclay brick or block. But because the refractory lining is gradually dissolved by the hot mixture of slag and flux, daily repair of the cupola lining was the rule. Swing-out bottom doors, covered with the sand bed, made daily repair practical by "dropping the bottom" after the last melt is tapped. Some foundries employed pairs of cupolas in order to maintain daily melting operation.

Water-cooled cupolas were developed to lessen the downtime associated with refractory maintenance. Two different types have been used: the enclosed jacket or tank design, and the externally cooled shell. First patented around 1900, water-cooled cupolas did not gain general acceptance until the 1950s. This was a result of the increase in high-volume, two-shift production, which exceeded the corrosive resistance of cupolas that were not water-cooled. Eventually, by banking the cupola fire on weekends, it became possible to operate these new cupolas continuously for two to four weeks.

Unlined, water-cooled cupolas suffer from a great deal of heat loss--as high as 20%. With the cost of coke more than doubling in the '70s to over $100/ton, many foundries retrofitted their cupolas with refractory, lowering the loss to an average 4-5%.

The use of two levels of tuyeres, which produce a divided blast, have been shown to decrease coke consumption by about 10% and increase the melt rate 11-23%. Preheating the air blast has also been shown to increase the melt rates by 25%, and decrease coke consumption by about 10 lb per ton of metal.

Tapping--In the past cupolas were tapped intermittently. Today, most cupolas are tapped continuously. Slag may be either skimmed in a specially designed trough or in a separate vessel.

Comparing the two procedures, it is worth noting that in exchange for increased efficiency, a degree of process control is lost using continuous tapping. A high percentage of cupola charges contain scrap. The differing metallic constituents of the charges are "averaged" when mixed in the cupola well, which is lessened when the cupola is continuously tapped. Mixing can occur in a receiving vessel (i.e. a transfer laddle), but only if it is sufficiently large.

Emission Controls--Cupola emission controls have been in place for many years, but the past two decades have witnessed rapid evolution in design. Both wet scrubbers, baghouses and electrostatic precipitators have been widely used, permitting cupolas to meet the most stringent air pollution standards. A number of different wet scrubbers are available, but the basic concept is one of bringing the exhaust gas particles in contact with cleaning liquids; the wetted particles then precipitate as sludge which is piped to settling ponds.

Baghouses contain interconnected fabric bags which have pores of various sizes at random locations. What might appear to be a relatively simple process is in fact complex, dependent upon the velocity of the air stream and the ability of the bags to maintain an electrostatic charge. Baghouses are an excellent means of particle entrapment and are widely used in U.S. and European foundries.

The first practical use of electrostatic precipitators dates to 1906. An electrostatic field is created and as the dust-laden gas passes through the field, the particles become negatively charged. They are then deposited on the oppositely charged collecting electrode. Electrostatic precipitators have been used on cupolas in Europe, and recent improvements may lead to wider acceptance in the U.S.

Computer Control--Continuous cupola charging affords an excellent means of balancing the different scrap additions. Computers can be used to monitor and control scrap additions which frequently differ in weight and composition.

Software has been available for several years which can determine the least cost charge-mix based on a given weight and consistency of scrap. Adjustments in coke additions may also be necessary to compensate for changes in atmospheric humidity.

Computers are now being used to implement statistical process control based on the chemical analyses of the melt, with adjustments based on the control limits of selected elements.

Reverberatory Furnaces

The reverberatory or air furnace is a modified open-hearth design. The reverb furnace is chiefly used for batch melting and holding. The furnaces are rectangular, with a shallow, saucer-shaped hearth and a low, arched roof. The design stems from the use of coal or coke in early furnaces which necessitated locating the firebox at the opposite end from the flue. The slanted roof was heated by bouncing or "reverberating" the flame off the ceiling. In modern reverb furnaces, the flame is used not to heat the refractory, but the melt itself.

There are basically two standard types of reverberatory furnaces: wet or dry hearth. Wet hearth furnaces use a combination of convection and radiation to directly heat the top of the metal bath. A dry hearth furnace typically contains two chambers: a sloped hearth in the melting chamber which drains into a dip-out well located in a secondary holding chamber. Burners are located in the roof of the two chambers which are separated by a partial wall. Tilting reverberatory furnaces eliminate many of the problems associated with tapping.

Dry hearth furnaces have the advantage that the temperature of the molten metal does not rise above the pouring temperature, thereby avoiding possible gas entrapment. The dry hearth must be cleaned regularly, however, to avoid the buildup of accumulated melt residue.

Some of the newer furnaces incorporate a hinged roof, which provide easy access to the hearth for cleaning. Scrap preheaters have also been used to increase fuel efficiency.

Crucible Melting

Crucibles were the earliest melting vessels. Modern crucibles are used in both smelting and melting--most often for nonferrous metals. They are either metallic--steel or ductile iron--or nonmetallic--a mixture of clay-graphite or clay-silicon-carbide. Selection of the crucible material usually depends upon the type of metal to be melted. Coated steel crucibles have been used successfully by aluminum foundries. Uncoated steel will be degraded.

Crucible furnaces may be either stationary or movable, lift-out, push-out or tilting. Lift- and push-out crucibles offer flexibility: different crucibles can be used for melting different alloys, preventing contamination. Metal is removed with hand ladles, or the crucible itself may be removed and used to pour the melt. Maximum crucible size is usually 300 lb (140 kg) of aluminum or copper.

Operation--The crucible is placed atop a refractory block inside a firebrick-lined steel shell having a movable cover. Crucibles can be heated either by means of a flame or electrically by mid-frequency induction coil, either lift-coil (push-out) or fixed-coil furnace. Multiple burners help to avoid hot (or cold) spots, thereby prolonging crucible life, increasing fuel efficiency and leading to faster melting and lower noise.

Care must be taken during the initial charge. Not only is there the danger that refractory crucibles may be damaged from sharp-edged materials, but the crucible wall may be ruptured from the expansion of wedged scrap.

The scrap itself should be free of adhering sand. Once a bath is established, any metal ingots that are added should be preheated to avoid condensation moisture and the possibility of an explosion.

Melting efficiency is maintained if a heel of metal equal to approximately 25% of crucible capacity is left in the crucible following each pour. This efficiency results from the heat transferred between the continuous interface of the molten metal and the crucible wall.

Particularly in nonferrous melting, once all of the charge has reached the molten state, very little additional increase in temperature is required to reach pouring temperature. Foundries rely on operator judgement to avoid overheating. "Boiling" of the melt frequently leads to gas entrapment and lowered metal quality. It can also result in increased crucible degradation.

Induction Furnaces

Induction furnaces are of two types: coreless and channel (sometimes called "core"). In both, an electromagnetic field is created as a current passes through electrical coils which are wound around a refractory lining. The alternating magnetic field induces a high alternating current in the surface of the melt which generates heat. The heat quickly distributes itself by conduction across the charge and melts the metal.

The advantages of coreless induction melters include their ability to: quickly attain high superheat temperatures; melt a variety of alloys; satisfy both volume and alloy flexibility; melt a wide variety of charge shapes; and provide precise control of melt temperatures and analysis. Induction melters are increasingly used in aluminum foundries precisely because the stirring immediately immerses the charge--particularly light gage, recycled scrap--minimizing oxidation and metal loss.

Coreless Furnaces--In a typical coreless induction furnace (Fig. 2), water-cooled, electrical coils surround the cylindrical refractory melting vessel. The furnace body can be made of either high-strength insulation boards or interconnected aluminum panels. In larger furnaces with a structural steel frame, the primary coil is backed by yokes which surround the entire refractory-lined melting hearth. The yokes serve to pick-up "stray" flux generated by the coil. They are not required in "box" or nonmagnetic furnace construction. Line frequency coreless furnaces (60-120 Hz), may require a starting block or a molten heel of metal during start-up, while medium to high frequency furnaces (120-10,000 Hz) melt from a cold charge.

Once the melt is established, the molten metal moves in currents determined by its response to the secondary magnetic field established in the charge. The efficiency of converting electricity to heat energy ranges between 60-85% depending upon the reactance of the charge and the lining wear or slag build-up on the walls of the crucible. Each can account for as much as a 30% reduction in circuit reactance.

The movement at the center of the melt is toward the center and then perpendicular to the field, while the molten metal reverses direction at both extremes of the coil winding. This creates a double torroidal or "doughnut" circulation pattern and results in a metal surface which is dome shaped. The metal becomes less "domed" with the increasing height of the metal above the coil.

Increased stirring occurs as the power increases, but decreases with increased frequency. The vigorous stirring imparted by the four-quadrant flow, typical of line frequency coreless induction melting, is shown in Fig. 3. This stirring results in uniform temperature gradients and homogeneous melt chemistry.

In coreless lift-coil furnaces the furnace shell containing the induction coil can be lifted off the melt crucible. In a push-out furnace the crucible is pushed up from a stationary coil. They are most commonly used for ferrous or nonferrous melts in the range of ten to several hundred pounds of metal where alloy segregation is important.

The molten metal can be poured from coreless furnaces in a variety of ways depending on the type of alloy and molding configuration. Nonferrous melters are designed so that the melt crucible can be used as the pouring ladle. Melt capacities are limited to about 500 lb. Larger-sized, nonferrous, as well as all ferrous alloys are tilt-poured from the furnace using either a hoist or hydraulic cylinders.

Coreless furnaces are either poured completely empty and batch-charged, or tapped and charged while maintaining a molten heel between pours. In the tap-and-charge method, 10 to 30% of the furnace is tapped, and an equal weight of charge added into the molten heel.

Batch operation, which has really only become feasible with the advent of solid-state, medium-frequency furnaces, is actually a more electrically efficient operating method because the cold charge responds to the magnetic flux better than molten metal. High powered systems with relatively small capacity furnaces can be used and scrap can be melted instead of relying on starting blocks. Although the charge can be dried in the furnace, it is important to recognize that wet materials should never mix with molten metal.

Despite batch operating efficiencies, partial tapping affords greater control of the melt chemistry, and because no holding furnace is required, it may be more economical, particularly for foundries pouring small batches at precise intervals. In a modern medium frequency melting furnace, either practice can be employed. Also, scrap can now be used for starting instead of relying on starting blocks.

For safety reasons, power is usually turned off during charging, slagging, sampling and tapping. Average utilization varies between 75-85% on-line operation. Prior to the development of the solid-state systems, erosion of the refractory wall resulted in increased electrical demand due to a change in coupling between the coil and melt in the lower frequency systems. Various controls were needed to compensate for this wear and the resultant overheating. The ability of solid-state systems to control the frequency output eliminates this situation.

Furnace manufacturers now offer computer controls which can provide continuous monitoring of different parameters, including: melting sequences for various alloys, programmed temperature control, furnace charging, lining curing cycles, ground fault detection, and electrical consumption. Software is available which can be used to determine alloying adjustments based on charge data and casting specifications.

Channel Furnaces--In channel furnaces a U-shaped channel at or near the bottom of the furnace wraps around an iron core wound with an air-or water-cooled copper induction coil. As shown in Fig. 4, the lower half of the furnace containing the channel and inducer is surrounded by refractory and (usually) enclosed within a steel box. Metal passing through the channel becomes superheated, and this heat is transferred by convection to the remainder of the melt in the upper unit.

The primary coil is backed by shunts around the refractory-lined melting hearth. Lower frequency furnaces (60-275 cps) require a molten heel of metal or a "starter" block to initiate the melting cycle, while medium and higher frequency power-trucking furnaces melt from a cold charge.

Because of this limited heat transfer, channel furnaces have seldom been used for large volume melting and alloying except in the off-peak mode. Instead, they have primarily been used extensively for superheating, duplexing and holding large volumes of both ferrous and nonferrous metals. They offer the lowest operating cost when melting and storing metal during low-cost, off-peak hours, with pouring done during the day shift.

The widespread use of channel furnaces in aluminum foundries has recently given way to the use of coreless and resistance furnaces because of fewer problems with clogging, although they are still widely used in the production of copper alloys.

Twin-coil channel furnaces, which require a molten heel of metal, have a pouring capacity of 200 to 5000 lb of aluminum. Until recently, high-powered, 60 cycle coreless furnaces had been an industry mainstay. New solid-state medium frequency furnaces, operating in the range of 70-5000 Hz, offer increased power density for the same size furnace, without decreased stirring. Greater efficiency is achieved with these furnaces because heat loss increases with increasing furnace area, and by varying or tuning the frequency to the charge, full configuration/density power can be maintained.

Electric Arc Furnaces 

Arc furnaces were first introduced at the beginning of this century. One of the most commonly used types of furnaces in the production of all grades of iron or steel, they can be cycled as needed without leaving a heel for startup.

Direct- or indirect-arc furnaces are commonly used to melt cast iron and steel. Both types of furnaces use low voltage, high amperage current. In the first, an arc is formed between the carbon electrode(s) and the melt. Both AC and DC (single electrode) units are in use. A new DC single electrode unit, shown diagramatically in Fig. 5, is said to offer significant reductions in electrode consumption.

Indirect melting occurs when the arc is struck between two electrodes independent of the melt. Arc furnaces are generally not used to produce low-temperature alloys because of pyrophoricity and control problems.

Design--Arc furnaces are refractory-lined, shallow, cylindrical vessels with a movable, domed refractory-lined roof. The choice of refractory will determine the nature of the slag and is therefore specific to particular metals. For example, if a silica or alumina acid refractory is used, the charge materials must be low in phosphorus and sulfur, because neither refractory will react with and remove these two elements. While this type of scrap is more costly and requires selectivity, the tradeoff frequently yields increased production, and flexibility to produce small quantities of different steels.

Modern "clamshell" arc furnaces are charged by removing the roof which swings up and to one side. The shell itself is usually made of rolled and welded carbon steel sections. The steel pot is lined with either an acid or basic refractory depending on the melt.

The electrodes are raised and lowered through the roof either manually or automatically, and are positioned so that they form as small an arc-circle as possible, in order to intensify the heat. Roofs are arched in order to minimize thermal degradation and to provide structural strength. They are increasingly made of cast iron except for the areas around the electrode ports, exhaust and oxygen lance holes. Water cooling has been used around the electrode ports and the roof ring, and is now beginning to be used to cool the furnace sidewall above the slag line.

Operation--Successful furnace operation depends on careful control of the electrodes. During the initial melting stage, they are lowered, independently to a predetermined distance above the charge, and then kept submerged in the scrap to reduce air flare damage during the meltdown. In the case of direct-arc melting, the arcs heat the molten bath as well as the solid charge. Oxy-Lp burners are sometimes used to decrease the melting time. As the size of the bath increases, the electrodes are raised. It is critical that the electrodes be kept parallel to the sidewalls. If not, hot spots may develop which can lead to excessive, nonuniform refractory wear.

Power is decreased once most of the charge becomes liquid. This is to protect the sidewall refractory from erosion by the arcs. Alloying and refining can be performed at this point under the shorter arcs.

Electric arc furnaces are designed for tilt pouring. The slag door is located conventionally, opposite the tap spout. Slagging is performed by back tilting the furnace over the slag pit.

A number of techniques are employed to help reduce electrical consumption, including: the use of burners during meltdown; water-cooled panels on the sidewalls to permit greater use of high power; and bottom tapping which permits faster pouring and less heat loss.

Vacuum Processes

A number of vacuum melting and remelting process are in use. Vacuum Induction Melting (VIM) is used to produce high purity, high strength castings ranging from several kilograms to 33 ton ingots or electrodes. VI melters are used in both primary and secondary refining to produce powdered metal, Directionally Solidified (DS), equiaxed and Single Crystal (SC) cast parts or wrought products. Many of these are aluminum or aluminum alloy castings used in aerospace applications.

Unlike electric arc furnaces in which the refractory lining--acid or basic--significantly affects the melt, dephosphorizing or desulfurizing is possible with the VIM process only using additions. This is kept to a minimum because any slag will solidify on the crucible wall, which is susceptible to slag erosion, and/or buildup.

VIM is however, very successful at reducing certain volatile trace elements to acceptable levels, producing castings with substantially increased strength and life expectancy. The high vapor pressures of such elements as arsenic, selenium, bismuth and copper permit their distillation and removal.

Secondly, VIM is a very effective method of degassing. This is because the solubility of nitrogen and hydrogen is directly proportional to their partial pressures. Reduction of hydrogen to levels as low as 1 ppm is common. But the removal of nitrogen can be complicated by the presence of nitride-forming metals such as chromium, vanadium, aluminum and titanium. In their absence, VIM can reduce nitrogen to as low as 20 ppm.

Deoxidation using VIM occurs from the formation of carbon monoxide in a two-stage reaction. The reaction is partially dependent upon the hydrostatic pressure of the liquid metal, which can be reduced by agitation--either argon purging or electromagnetic. Deoxidation, and the accompanying decarburization, is particularly important in the production of low-carbon, high-chromium steels.

The Process--Vacuum induction furnaces have two chambers--one for molding, the other for melting. Instead of modifying equiaxed furnaces for the production of DS or SC castings, more dedicate, single purpose furnaces are being purchased. Melting takes place in a crucible positioned inside a vacuum-tight chamber. Different configurations are used: the mold chamber may be located beneath or next to the melt chamber. Both tilt and bottom pouring are used.

VIMs must be charged and tapped in a precise, sequential manner in order to maintain vacuum. This sequential operation lends itself to automated process control. Two-color pyrometers are used to control the melt temperature, and pouring can be programmed based on pouring or mold temperature, mold position or pouring height.

Precise process control is also critical because, as was mentioned, an active slag would react with the crucible wall. But, in the absence of an active slag, impurities collect on the crucible wall. Melt purity is therefore partially related to the length of time the melt comes in contact with the crucible wall, which acts as a ceramic macro filter.

Efforts to improve the purity of VIM melts have led to more stable crucible refractories. Inert gases can also be used to rinse the melt. Tundish and launder systems, with elaborate weirs and dams, filters and deslaggers are used during the pour to remove oxide contaminants. [Figure 1, 2, 3, 4 and 5 Omitted]