Over-preheating scrap can cause high levels of melt slag.

A gray iron foundry discovered that excessive scrap prehaet temperatures create slag and cancel the benefits of extra BTUs.

Like many foundries, when the current resurgence of the metalcasting industry began, Grinnell Corp.'s gray iron foundry found it needed to boost meltshop output to keep pace with increased production demands.

One seemingly obvious way to do this was to increase the temperature setting of the plant's scrap preheater. Officials assumed a higher charge temperature would lead to lower kWH/ton and higher melt rates from the foundry's existing equipment - the hotter the charge, the less time and energy needed to bring it to tap temperature. As Grinnell found, however, there is a limit to how hot scrap can be without introducing other problems to the meltshop.

Foundry Operations

Grinnell's melting operations are comprised of two separate groups of furnaces that supply molding with up to 600 tons of iron per day. The oldest furnace group contains four 15-ton coreless induction melters powered by two 3700 kW powerpacks. These furnaces are fed by a 12 ft, 16-burner pre-heater. Two 4500 lb charges are conveyed to the furnaces by bucket feeders for a total charge and tap size of 4.5 tons.

The newest furnace group contains three 25-ton coreless furnaces powered by two 5400 kW transformers. In this group, an 18-ft, 16-burner preheater is used. Two 7000 lb charges are conveyed by bucket for a total charge and tap size of seven tons for this group. Both of these furnace groups feed seven high-speed green sand molding lines.

When all seven molding lines are running, however, melting is a major restriction, capable of producing only 32 tons per hour, and causing periods of downtime as molding must wait for metal. One of the easiest ways to increase output without major capital investment seemed to be decreasing melt times by increasing charge temperature and speeding the melting process.

Slag Problems

Both preheaters were set for a six-minute cycle at a 1250F (676C) exit temperature, up from 1100F (593C). At this setting, virtually all the scrap leaving the preheaters was cherry red in color.

Furnace operators, however, soon reported an increase in the amount of slag generated in the melt, and also that it was more fluid and harder to remove. Weighing the slag removed from the bath for several days confirmed this. The slag weight averaged 72 lb per ton melted, and it appeared to be more fluid than before. It was also realized that there was no dramatic increase in the melt rate of the furnaces at the higher preheat temperature.

At that point, officials decided to determine whether a reduction in scrap temperature would lead to reduced oxidation and less slag generation. One of the other aspects of the problem considered is that slag typically has a heat content about 1.8 times greater than iron. Grinnel officials therefore postulated that if reducing the preheat level led to a slag reduction, there would also be very little increase in kWH/ton, even though there would be fewer BTUs in the scrap.

Weeks of Testing

In the initial trial, both preheaters were lowered to 1150F (621C) and slag weights monitored for a week. The 15-ton furnaces showed a drop from 72 lb to 67 lb of slag per ton. The kWH/ton remained virtually constant, averaging 513.7. These results encouraged a further and more controlled look at what total effect various preheat temperatures would really have.

Over the next several weeks preheat levels were reduced and weekly data collected. Average slag weights were generated for 1100F (593C), 1000F (537C) and 950F (510C). The effects on kWH/ton, gas consumption, slag generation and alloy usage were all analyzed.

The first, and perhaps most important development was with the kWH/ton figure [ILLUSTRATION FOR FIGURE 1 OMITTED]. The 513 kWH/ton required by the 15-ton furnaces at 1150F preheat dropped to 510 at 1100F. At 1000F, that number was 499, and the last seven weeks at 950F averaged 482 kWH/ton.

These results are a clear indication that above a certain point, increased scrap preheating had actually become counterproductive. Alloy consumption and slag generation figures verified this conclusion.

The problem is best explained in the oxidation curve found in figure 2. Up to about 1200F (648C), the increase in scrap oxidation is negligible. However, between 1200 and 1500F (815C), the rise in steel scrap oxidation levels is exponential. As the amount of FeO rises, both the volume and fluidity of the slag increases. FeO is the most chemically corrosive constituent of slag and forms the lowest melting point two and three phase compounds. The low melting point slags tend to emulsify in a coreless furnace, forming very small particles that float very slowly to the surface.

This extra slag also increases furnace power consumption. Theoretically, iron only requires about 450-500 kWH/ton to reach normal tap temperatures of 2700-2800F (1482-1537C). But slag will take 700-800 kWH/ton. Thus, the large increase in slag volume from high preheating negates any extra BTUs the preheated charge contains.

Further testing at 950F helped bear out this conclusion. This temperature was maintained for several weeks while slag was weighed and average kWH/ton was monitored. With this same preheat temperature, there was a variation of 461-502 kWH/ton with slag volumes ranging from 44-62 lb per ton. This was because of so many process variables, like scrap charge makeup and molding line demand.

Alloy Consumption and Scrap

The last area of concern was alloy consumption. Silicon carbide (SIC) usage was fairly steady at all temperature settings, since it is primarily the carbon in the SiC that does the deoxidizing at iron temperatures. As should be anticipated by the SiC reaction (SiC + FeO = Si + Fe + Co), carbon usage fell as the preheat temperature was lowered [ILLUSTRATION FOR FIGURE 3 OMITTED]. Therefore, more of the carbon present as SiC was available to raise the carbon level of the bath, and less additional carbon was needed.

Another indicator was harder to prove, but lent more credence to these theories. Slag defect rejects had risen to unusually high levels at the 1250F preheat temperature. As the preheater was throttled back, slag scrap levels dropped accordingly. By the time the preheat temperature was down to 950F, plantwide scrap levels had fallen about 40%, with slag-related defects even lower.

Recommendations

Though faster melting times are the main objective of increased preheat temperatures, furnace energy consumption is hopefully reduced or made more cost effective. However, charge over-preheating wastes gas and raises energy costs for that operation.

Observations from a number of coreless induction meltshops indicate the problems caused by over-preheating are worse for ductile iron foundries. Many are paying a premium for low-sulfur/low-manganese scrap, and much of this tends to be very thin. Once in the preheater, it turns cherry red within seconds, severely oxidizing the surface. This leads to high levels of FeO in the bath, forming fayalite and other low-melting point slags that carry over into characteristic "bird shot" defects. Thin metallics must be approached very carefully in preheating operations.

While optimum cycle-times and temperature for each foundry - and even each preheater - varies, Grinnell discovered the optimum temperature to be 900-950F. A similar test program can help other coreless induction furnace meltshops discover the ideal balance necessary to benefit from increased pre-heat temperatures.