New method helps determine when to shut down an inductor.

Using thermocouples to monitor the lining status of a channel induction furnace is a more exact way to help prevent runouts.

It's a melt department dilemma. How do you maximize the metal tonnage throughput on a channel induction furnace refractory lining without incurring a runout? Obviously, the longer you run a furnace without a reline, the more efficient your melt shop and the lower your refractory costs. On the other hand, molten metal runouts lose metal, force a shutdown and may cause serious equipment damage and safety hazards.

The danger lies with the inductor channel. When an inductor wears, the channel increases in size and causes the power and the current to both increase for the same applied voltage. This channel wear brings the molten metal closer to the bushing.

When an inductor builds up, the channel size decreases, causing the current and power to decrease for the same applied voltage. The concern with this channel buildup is that since the metal in the loop cannot exchange its energy with the upper hearth metal, it will superheat and melt the refractory lining. Thus, whether the channel wears or builds up, the result is the same: a certain molten metal runout.

Over the years, people have devised many methods to determine when a channel furnace inductor should be shut down before a runout occurs. These methods are reasonably good predictors of when to shut down a furnace lined with castable or wet ram refractories, but their information is insufficient when the inductor is lined with a dry vibrated refractory.

However, a new, more reliable method has been developed to make this determination. Before introducing it, a review of established methods is necessary.

Traditional Methods

Conductance Method - The oldest way to monitor the changes in an inductor is the conductance method. The conductance of an inductor is defined in Table 1.

When an inductor wears, the conductance will increase as the current and power increase. Conversely, when an inductor builds up, the conductance will decrease as the current and power decrease. When the conductance exceeds 120% of its starting level, or is less than 80% of that level, the inductor is usually shut down. These percentages will vary with operational experience.

Resistance Method - Another way to monitor the changes in an inductor is the resistance method. Table 1 contains the definition of the resistance of an inductor.

By definition, resistance is the reciprocal of conductance. When an inductor wears, the resistance decreases as the power and current increase. When the inductor builds up, and power and current decrease, resistance increases. As with conductance, when resistance exceeds 120% or is lower than 80% of the starting resistance, the inductor is shut down.

Reactance Method - The reactance of an inductor is also defined in Table 1. Reactance works like resistance: when an inductor wears, the reactance decreases as the power and current increase, and vice versa. The critical numbers for reactance are also 120% and 80%.

In a 1967 paper, J.L. Hoff made the point that the inductive reactance is affected only by the size and configuration of the channel. He called the reactive method the "shape factor," as distinguished from the conductance and resistance methods, which he termed the "metal factor." Since the shape factor is affected only by the size and configuration of the channel, Hoff argued that reactance is a better way to address changes in the channel than the metal factor, which is affected by the temperature and chemistry of the metal, in addition to channel configuration.

Figure 1 supports Hoff's premise that the reactance method gives the better analysis. Based on the resistance or conductance ratios, it would appear the inductor is experiencing serious buildup. The reactance ratio, however, contradicts this conclusion. These curves are for a 350-kW throatless inductor operating on treated ductile iron. The furnace ran for 37 months without plugging.

Frequency Method - With the introduction of the solid state power supply, an inductor could be designed to operate at a frequency other than 60 Hertz (Hz). Since the number of power factor correction capacitors in the power supply are fixed, the frequency of the operation will vary from the startup frequency, depending on changes within the inductor. When an inductor builds up, the frequency decreases; when it wears, the frequency increases. Because the frequency varies with the condition of the inductor, it becomes a tool for determining when to shut down an inductor equipped with a variable frequency power supply.

Prediction Problems

The predictions made by any one of the aforementioned methods work quite well when the inductor is wearing. However, they do not work as well with inductor buildup. The problem is that molten metal can run out through the bushing even though the prediction methods show nothing wrong.

In these cases, metal saturates the refractory lining between the loop and the bushing. The above prediction methods don't detect this saturation because the effect of the buildup masks the effect of the saturation. In other words, none of the existing methods can differentiate between simultaneous effects of buildup and saturation.

Lining saturation occurs to a depth defined by how far the metal can penetrate the lining before the leading edge freezes. With a typical 6-in. lining thickness, the saturation network should freeze within 2 in. of a copper water-cooled bushing.

However, when the bushing contains an uncooled joint that incorporates a transite T-bar for the electrical isolation, the thermal gradient in the bushing joint area is adversely altered. The front edge of the saturation now penetrates deeper and comes in contact with the T-bar. Since 99% of all bushing runouts occur at the bushing joint insulator, a better method than those based on the electrical readings is required to determine the status of the inductor.

Thermocouple Method

The thermocouple method, shown in Fig. 2, answers that need. In this method, the bushing is fitted with Type K thermocouples. They are taped to and run along the axial length of the bushing until they reach its center, where they are bent 90 degrees to protrude 1 in. into the refractory lining. Thus they can monitor the temperature condition of the lining around the bushing.

If the temperature of the copper bushing is 140F (60C), and if the temperature 1 in. from the bushing is known, the thermal gradient between the bushing and the saturation network can be calculated. When melting cast iron, final solidification takes place in the range of 2010-2066F (1098-1129C). Knowing this, you can determine how far the leading edge of the saturation network is from the bushing.

Proving the Method

Following are three case studies in which thermocouples were employed in addition to the conventional methods to predict when an inductor should be shut down.

Case No. 1 - The vertical channel furnace in this case was a melter for copper. It had a capacity of 20 tons and was powered by a 900-kW, single loop, throat-type inductor operating at 120 Hz. The inductor was lined with a mullite bonded alumina refractory grain with silicon carbide. The lining had an initial bond temperature of 1175F (634C).

Figure 3 shows the temperatures around the bushing as a function of time. Note that the temperatures all started and stayed within 100F of each other until Day 6, when the thermocouple at the 4:30 position rose sharply to reach 1164F (629C) by Day 7. The inductor was immediately shut down.

An autopsy of the inductor showed that the refractory lining had suffered numerous cracks, which had filled with molten metal and altered the thermal gradient. Yet the inductor had shown acceptable reactance and frequency readings. Although the reactance ratio and/or the frequency ratio couldn't pick up the finning, the thermocouples immediately identified the problem, saving a certain runout.

Case No. 2 - This study was done on the same vertical channel furnace for melting copper, but the power supply had been updated with additional capacitors to lower the starting frequency from 120-90 Hz. The inductor was updated to include a copper bushing with a water-cooled bushing joint cap.

After melting 4 million lb of metal, the frequency had increased from 90-115 Hz, the recommended shutdown frequency. But because the temperatures around the bushing were well below the 1100F (593C) shutdown temperature, operators decided to continue melting.

The inductor melted another 4 million lb of copper before it was shut down at a frequency of 125 Hz. The life of the inductor on this furnace is now judged by the refractory temperatures around the bushing. While the reactance and frequency ratios are still calculated, they are of secondary importance.

Case No. 3 - In this case, a vertical channel furnace was used as a holder for cupola-melted iron. The furnace had a capacity of 60 tons and was powered by a throatless, 1100 kW single loop inductor operating at 60 Hz. The inductor had a copper water-cooled bushing fitted with a water-cooled joint cap, and was lined with a spinel bonded alumina refractory grain.

Figure 4 shows how the reactance for this inductor varied with time. The reactance fell from 100% to about 85% during the first three months, before leveling off. Then, during the 13th month, reactance dropped sharply to 65%, causing concern.

Yet the temperature in the refractory about the bushing did not exhibit any problems to continued operation. Consequently, the inductor was run for another four months before being shut down at 68% reactance. An autopsy showed that, though the channel's diameter had grown from 5 in. to 6.5 in., there was still 2 in. of unsintered refractory next to the bushing.

While there are many ways to judge the status of an inductor's refractory lining, the thermocouple method appears to give the best information on which to base a decision on inductor shutdown. It is still recommended that the inductor be monitored by one or more of the other methods, since the more information available to the person making that decision, the better for the channel furnace operation.