Evolution of high powered single loop inductors.

Continual refinements to the channel induction furnace have led to an effective alternative source for melting casting irons.

Throughout the 1800s, the cupola presented the only feasible method for melting cast iron. While the channel furnace was invented in 1890, it wasn't until 1917 that James Wyatt developed a workable single loop inductor. In 1942, the twin loop inductor was introduced. In 1945, the single loop inductor was improved by going to a circular geometry. In 1961, the twin loop inductor was also improved through geometry changes.

Despite these improvements, the cupola continued to be the main melter for cast iron well in the U.S. into the 1960s. However, in the late '40s, the situation in Europe was quite different. Most European foundries were destroyed during the World War II, and were being rebuilt using large capacity, high powered, line frequency, coreless induction furnaces as the prime melters.

It wasn't until the late '50s that these line frequency coreless furnaces were introduced to American foundries by an American furnace manufacturer under a European licensing agreement. Because not many of these line frequency coreless furnaces were sold, the American furnace manufacturer began to develop a magnetic tripling transformer to power a medium frequency coreless induction furnace.

Likewise, in the mid-'60s, impending air pollution regulations required significant capital expenditures for foundries equipped with cupolas. Rather than investing in air pollution equipment, many foundries replaced their cupolas with line frequency coreless induction furnaces. Since there was more experience with line frequency coreless melting in Europe than here in the U.S., the sale of European furnaces prevailed, which accounts for the large number of these furnaces in this country.

Because several line frequency furnaces were normally required to do the job of one large cupola, the coreless induction furnace didn't find favor with all high production foundries. But because the cupola tended to produce cold iron with varying chemistries, the American furnace manufacturers introduced the drum-type, channel holding furnace. The purpose of this unit was to superheat cold iron and to homogenize off chemistry iron.

In the early '70s, the furnace manufacturers again became concerned that with the even more stringent air pollution requirements on cupolas, the need for channel holding furnaces would diminish. Consequently, furnace manufacturers introduced the vertical channel furnace to gain a portion of the coreless furnace market. The operational intent with a vertical channel furnace was to melt during off-peak shifts to take advantage of lower power rates, and to pour metal during the peak shifts. Thus, the operation of the vertical channel furnace had to be such that sufficient metal would be available during the day to meet pouring requirements.

During that period, inductors were relatively low powered, as there was plenty of time to melt and refill the furnace during off-peak hours. Although its energy consumption per ton of iron melted was more favorable than that of the coreless furnace, these low powered, vertical channel furnaces were unacceptable to many high production foundries.

The '80s saw the medium frequency coreless furnace get its second wind. Introduction of a silicon controlled rectifier made possible the efficient conversion of line frequency power to medium frequency power. Thus, the high powered, medium frequency coreless furnace became the preferred choice in new melting equipment. Since these furnaces are extremely fast batch melters, a holding furnace is often required to take full advantage of this melting concept.

Further developments with the vertical channel furnace also took on a new direction in the late '80s. Furnaces were fitted with higher powered inductors, which gave them the ability to compete with the coreless furnace. Power ratings reached 2000 kW for a single loop inductor and 3000 kW for a twin loop inductor.

Recent Developments

It is not the intent of this article to rate one type of inductor over another, but to report on progress made during the past six years with yet even higher powered single loop inductors.

For many years, it was believed that the temperature rise in a single loop inductor was too high and that these inductors would probably never be successful for high powered applications. As power levels on a single loop inductor were increased well beyond 2000 kW, no evidence of channel overheating was found.

In order to resolve this disparity, a review of how a channel furnace inductor operates is required. As shown in Fig. 1, the inductor is an electrical transformer wherein the coil is the primary winding and the molten metal loop is the secondary winding. The manner in which the transformer action of an inductor works is that when voltage is applied to the primary coil, magnetizing current results and produces magnetic flux in the core. That portion of the flux which is mutual with the loop causes a voltage to be induced in the loop. Since the loop is a short circuited winding, a current of a high magnitude results. It is this current that heats the metal in the loop.

Since the loop current surrounds the magnetic field in the core, a force is exerted on that portion of the molten iron that carries the current, causing it to be pushed radically outward as shown in Fig. 2. Note that the current is not uniformly distributed in the channel. Rather, the current is maximum on the inside of the loop and falls off exponentially in the radial direction. The depth at which the current is 36.8% of its surface value is defined as the depth of penetration, which for molten iron is about 3 in.

The depth of the channel in the early single loop inductors was in the range of 2 in. Since this depth was less than the penetration depth of the current in molten iron, all of the iron in the channel was being heated by the induced current and all of the iron was being pushed radially outward by the electromagnetic forces.

Because the energy that is imparted to the metal in the channel in a single loop inductor is exchanged with the upper bath iron by thermosiphon means, there has to be some place for the heated iron to go without being redirected by the electromagnetic forces. When this flow path is not provided, there is a minimal exchange of heated loop iron with colder bath iron resulting in high channel temperature. This was the problem with the early Wyatt furnace and where the myth about overheating in a single loop inductor began.

When the depth of the loop is at or above 1.5 times the penetration depth, the thermosiphon system operates and lower channel temperatures result. Hence, the depth of the channel is as important - or maybe even more important - than the width of the channel. Our preference for the channel depth is two times the penetration depth or 6 in.

Channel Temperatures

Quantitative temperature measurement of the molten iron in the channel is not an easy task. What is required is an alternative method for determining the temperature in the channel. The method that was used to evaluate these temperature excursions is qualitative and is based on the following technical premise. Because the resistivity of molten iron varies directly with temperature and furnace resistance is directly proportional to the resistivity, it follows that temperature changes in the channel can be evaluated by monitoring changes in the channel resistance. In other words, an increase in resistance means an increase in temperature, while a decrease in resistance means a decrease in the temperature.

Figure 3 illustrates how the channel resistance varies with the power level. The drop in resistance suggests that the temperature in the channel drops with increasing power. At first glance, this does not appear to make sense. However, as more power is applied, more metal is moved through the channel, bringing colder iron from the main bath into the channel, which lowers the temperature in the channel.

Figure 4 shows how the channel resistance varied during a melt cycle with a furnace operating at about 3400 kW. During the first four minutes of the melt cycle, the resistance decreased. For this to occur, the temperature of the iron in the channel had to decrease. Such was the case as the addition of cold charge to the furnace lowered the bath temperature, which in turn lowered the channel temperature. Then, during the latter portion of the melt cycle, the resistance increased back to near the starting resistance. For this to occur, the temperature of the iron in the channel had to increase. Again, this was the case as the addition of energy to the bath heats it back to temperature. Since the resistance at any point during the melt cycle never exceeded the starting resistance, it can be concluded that the temperature in the channel never exceeded the starting temperature.

Experience over the past four years with five 30-ton vertical channel furnaces, each fitted with a 3000 kW, single loop inductor shows that these inductors are capable of melting ductile base iron to 2650F (1454C) at a rate of 6.5 tons per hour with an energy consumption of about 462 kWh/ton. Inductor life is six to eight months, with an average throughput of 22,000 tons. One inductor ran almost a year, having melted 34,000 tons. When these throughputs are compared to an industry average of about 7500 tons per inductor, it may be concluded that properly designed high powered single loop inductors present an advantage over conventionally designed inductors.

It should be noted, that it is not the inductor that shuts down a furnace, but a usually worn-out upper case lining. The high alumina brick lining wears from slag attack while the dry vibrated magnesia refractory in the inductor shows no wear and virtually no buildup over the entire life of an inductor. At present, the limit to inductor life is not known because it is the upper case lining that normally fails first.

Increasing the channel depth or implementing the use of dry vibrated refractories were not the only changes that enabled these high powered inductors to achieve the aforementioned throughputs.

Channel Geometry

By conventional design, the loop is positioned as close to the bushing as possible to maximize the furnace power factor in order to keep the initial equipment cost down. The typical 3 in. refractory thickness between the bushing and the loop was considered inadequate for high powered inductors. A minimum refractory thickness of 6 in. was chosen. However, when the loop is moved away from the core, both the power factor and the electrical efficiency drop. Since additional power factor correction capacitors are a one-time purchase and since the drop in electrical efficiency is minimal, the trade-off was made in favor of a thicker refractory lining.

Channel Forms

Conventional channel forms for dry vibrated refractories were steel weldments with either a square or rectangular cross-section. Even though the corners were rounded, in many cases they caused quadrential cracking of the inductor refractory which led to premature inductor failures. This problem was eliminated by using a round cast steel channel form.

Bushing Joint Cap

Dry vibrated refractory linings are readily saturated (this will be discussed in greater depth next month) by molten iron, which can lead to premature inductor failure with runouts at the bushing joint insulator. While refractories have long been considered the major problem in many channel furnace failures, it is estimated that 95% of all inductor runouts occur at the bushing joint insulator. Often the culprit is the design of the bushing. Test work revealed that design changes could significantly improve the reliability of the channel furnace inductor.

Because of the thermal discontinuity created by the bushing joint insulator, the thermal gradient is such that the saturation network can come in contact with, destroy and runout through the joint insulator. In the development work mentioned above, joint runouts were eliminated by replacing the original stainless steel bushings with copper, water cooled bushings fitted with a water cooled, bushing joint cap. Since the water cooled cap eliminates the thermal discontinuity in the joint area, joint runouts have virtually been eliminated.

While saturation of the lining still occurs, the saturation network is held at about 1.75 in. from the bushing for the entire life of the inductor. If 1.75 in. of unsaturated refractory remains with a 6 in. lining, consider what depth of unsaturated refractory exists with lesser lining thicknesses. It was for this reason that the lining was thickened.

Operational Considerations

When a high powered inductor is installed on a channel furnace, the furnace can now outmelt its own capacity in off-peak hours. Consequently, off-peak melting is no longer of any value. Likewise, as the power level is increased, the ferrostatic head must be increased to keep the molten metal from pinching in the channel by electromagnetic means. Thus, the high powered channel furnace must be operated on a back charge to a heel basis with the metal level fluctuating between full and one charge down.

Since the low powered vertical channel furnaces were poured down to minimum heel, they were equipped with a conventional lip-type pour spout. Now that the metal level can be maintained near the full level, a teapot spout can be used. This has a number of advantages. First, no slag pours from the furnace into the transfer or treatment ladle. Second, the high melt height keeps the slag from being drawn into and building up in the throat. Third, the furnace can be operated with a slag cover that needs to be removed only once a shift. The slag cover helps reduce the amount of heat loss when the lid is open for charging.

Power Upgrades

The original 3000 kW inductors have since been modified. These inductors now operate over 3500 kW. A 4500 kW, single loop inductor has been built and is presently in service at reduced power to prevent overheating of the main 4200 kva air cooled transformer. However, this inductor has been operated for short periods of time at power levels over 4600 kW. Presently, delivery of a 7000 kva, 12.47 kv, water cooled transformer is pending. Not only will the inductor operate at the designated power, but it will also allow for the determination of the maximum power rating for this inductor.