20/1

Energy Saving Potential of Melting in Medium-Frequency Coreless Induction Furnaces F Donsbach and D Trauzeddel. Otto Junker GmbH, Simmerath, GERMANY.

Abstract The available energy saving approaches in induction melting operations are twofold:

• Reduction of the furnace system's electrical losses through improved design

• Use of optimized operating and control regimes The first part of this paper describes in quantitative terms, based on a review of investigations and test results, how the surplus consumption of power can be reduced markedly through proper feedstock selection, correct charging techniques and an optimum adjustment of furnace parameters. Furnace design has been improved with the aim of addressing the largest loss factor, i.e., electrical losses in the induction coil, in an effort to cut these losses substantially. It was demonstrated that the new design yielded energy savings in the range of 5 to 10 %. Key words induction melting; cast iron; energy saving; operating regimes; furnace design

20/2

Introduction Optimised medium-frequency technology reduces thermal and electrical losses to a minimum. Exact weighing of the charge materials, a correct calculation and input of the appropriate amount of power based on the use of a melting processor, and a precise computer control of all equipment provide excellent conditions for an energy-saving melting operation. However, the full benefits of this technology can only be achieved through an appropriate furnace operating and control regime, safe and reliable operating practices, and optimum equipment design. The overall efficiency of the furnace system can thus be pushed to over 75 %. As a result, the energy input needed to melt cast iron at up to 1,500 °C would be brought down to a mere 490 - 520 kWh/t at an enthalpy of 390 kWh/t. In practice it has been found that the average power consumption of existing foundries in real-life cast iron melting operations is significantly higher. From UK foundries, for instance, a value of 718 kWh/t is reported [1], whereas French industry statistics show an even higher level of 855 kWh/t [2]. A large energy-saving potential can thus be identified here. Substantial cuts in power consumption are achievable through the use of advanced medium-frequency melting furnaces, but also via modified furnace operating and control regimes. On an existing furnace system, up to 20 % energy can be saved by adopting improved operating and control modes. This is a cost-cutting potential not to be neglected, specifically in times of ever increasing energy prices. Influence of the operating and control regime 1. Charge materials and make-up An accurate calculation of the necessary charge make-up, based on material analyses, and a precise weight determination and metering of charge materials and alloying additives (including correction for set/actual value deviations) are basic prerequisites for minimising melting times and power needs. The use of clean and dry charge materials will definitely pay off, given that the formation of slag due to sand adhering to uncleaned returns will consume just as much specific energy as is necessary to melt the iron, viz., about 500 kWh/t. With a realistic amount of 25 kg of sand per tonne of iron this adds up to 12.5 kWh/t. Beyond that, of course, the quantity of slag is increased as well. An even more decisive factor is rusty charge material. Its inferior electromagnetic coupling properties impair the transfer of melting energy and result in much higher melting times. The energy consumption and heat cycles for clean and highly corroded steel scrap, respectively, have been determined in comparative trials [1]. It emerged that rusty steel scrap

20/3

took 2 - 3 times as long to melt and required a 40 - 60 % higher power input. Even assuming that these values reflect an extreme case, the negative effect of rusty charge material is quite severe. In addition, there are higher melting losses and greater slag volumes. Hence the use of rusty charge material should be avoided wherever possible. The level of electromagnetic coupling achieved and hence, the power consumption of the charge, is a function, not least significantly, of the charge packing density. The heat cycle and energy consumption of the charge will thus vary with the packing density. The nature of this correlation has been examined with charges of different packing density in a high-power medium-frequency furnace operating under production conditions. The system employed for these trials had a capacity of 10 tonnes and a power rating of 8,000 kW at 250 Hz. The empty furnace was filled once with a charge of the specified composition, comprising pig iron, scrap castings, returns, steel scrap and additives. No further charge material was added as the metal was heated to 1,380 °C. The power consumption was measured throughout this period. Different dimensions of the returns and steel scrap fractions made for packing densities in the 2 - 2.7 t/m³ range. It is evident from the results that a drop in packing density from 2.5 to 2.0 t/m³ caused a 25 kWh increase in power consumption (Fig. 1). Despite the additional cost and effort, it is therefore advisable to crush all too bulky returns to achieve a higher packing density. This will also facilitate furnace charging and reduce the danger of material bridging in the furnace. The example of a U.S. foundry demonstrates that this practice can save money despite the costs caused by additional crushing operation [3]. At the same time, a quick and continuous charging workflow is important when it comes to saving operating time and cost. A high filling level should be maintained at all times. Mobile shaker chutes and a bin accommodating the full charge are prerequisite to meeting this requirement. An extractor hood closely covering the chute will minimize radiant heat loss while ensuring that the furnace fumes will be reliably captured. 2. Chip melting As foundries extend their vertical integration and machine their own castings, they increasingly find themselves with large amounts of chips on their hands – and what would make more sense than to try and use these chips in their own melting operation. Coreless induction furnaces, unlike other melting processes, are highly suitable for melting down machine tool chips. Since grey cast iron is normally machined without coolants, these chips are dry and clean and can therefore be melted down without any pre-treatment.

20/4

However, it should be noted here that the electrical contact between metal chips, despite their good packing density, is notoriously poor as a result of the small contact surface and surface oxidation. This is why the furnace should always be operated with a heel (> 40 %) when chips are melted. If the furnace is operated without heel, the power consumption for melting chips should be anticipated to be 50 kWh/t higher than for lumpy material. An increase in melting time must also be expected. Chips can be charged on the liquid heel either continuously ('trickle' method) or in one batch up to the top of the active coil. Filling the furnace all the way, without overloading it, will save 2 - 3 % energy and reduce the melting loss. On the other hand, there is the risk of bridge formation in the charge. Where chips make up only a portion of the charge, the solid or lumpy fraction should be placed in the furnace first, and the chips should be charged into the liquid heel once it has formed. 3. Carburising Another factor reported [4, 5] to affect power consumption is the method of adding carburising agents. Power consumption will clearly be higher if carburising agents are added into the molten metal bath after melting down rather than along with the solid charge material at the beginning. In-house experience indicates that this practice will consume about 1 to 2 kWh more per kg of carburising agent. This means that with a realistic input of about 2 % of carburising agents, an additional consumption of max. 40 kWh per tonne of iron is to be expected. An average of 70 kWh per tonne of iron for carburisation, as quoted in part of the literature, appears to be unreasonable. If the carburising agent is introduced into the furnace with the other charge material, this should be done in controlled proportions so that the carbon content of the melt will not rise unnecessarily. An excessive increase in carbon concentration would cause premature crucible wear. It is also advisable to avoid the use of too fine-grained, low-grade carburising agents which tend to adhere to the crucible wall. Local erosion effects would be the inevitable result. Furthermore, the input of silicon carriers should not take place until after carburisation is completed because increasing Si content in the melt decreases carbon solubility and also increases silicon losses. It should be noted in this context that the enthalpy levels of returns and synthetic cast iron differ markedly. When melting synthetic cast iron (steel scrap, carburising agent, silicon carrier), the power consumption should be expected to be 8 - 15 % higher than with a charge of homogeneous returns [6]. 4. Melting furnace operating regime In theory, the most favourable operating regime would be one involving the maximum available electric power and hence, high power densities. The overall efficiency of a melting plant, as can be seen from the following equation

20/5

=

power rated

power holding 1 electricaltotal - η η

is also determined, quite decisively, by the ratio between holding power and rated power. For a furnace of the same dimensions, the consumption of electrical energy will decrease with increasing rated power. This has been conclusively confirmed by systematic trials: as the heat cycle is reduced and thermal losses diminish, the electrical power consumption is reduced. From the calculated power diagram of a 12-tonne-furnace (refer to Fig. 2), it is evident that the electric power consumption increases exponentially with decreasing power density since the percentage of energy required to make up for steady-state thermal losses will become disproportionately high when the power density is very low. A comparison between a 6,000 kW melting operation and one with 3,000 kW reveals (cf. Fig. 3) a substantial power consumption difference of 20 kWh/t. This advantage can be utilized by changing from a mains-frequency to a medium-frequency system, since the maximum power input for a mains-frequency furnace of this size is around 3,000 kW. The use of medium-frequency technology makes it possible to operate without heel and to melt down small-sized charge material. Thanks to better electromagnetic coupling of the solid charge material (although this applies only to cast iron melting) the energy consumption in batch operation is 8 % less because a much higher coil efficiency is achieved up to the Curie point (Fig. 3). Since mains-frequency systems must be started up with a liquid heel, this energy saving can only be realized by shifting to medium-frequency technology. The higher power density and superior coil efficiency in batch operation add up to an energy saving of 12 - 15 %, always provided that the switch is made from mains-frequency to medium-frequency melting. The amount of heat stored in a coreless furnace (i.e., the energy required to heat up the cold furnace to its fully heat stored condition) is normally higher by a factor of 3 - 5 than the holding energy required over a similar period. Thus, the heat stored in an 8-tonne-furnace is 800 kWh, i.e., melting a charge in a still-cold furnace would require 100 kWh/t more energy than is needed to melt the same charge in a unit already storing full heat (Fig. 4). Since it will take only a quarter of this energy (25 kWh/t) to keep the melt at holding temperature for an hour, it makes sense not to let the unit cool off but to keep it at temperature with a heel of molten metal during interruptions or breaks of less than four hours. It should also be considered that the service life of the refractory furnace lining can be maximized by keeping the furnace continuously at operating

20/6

temperatures, or at least by not perpetually switching it off and on. As a general rule, extended holding periods will no longer affect the metallurgical quality today thanks to the advanced treatment and inoculating technologies available. Energy is wasted also by operating the furnace with its lid open for longer than necessary. The low heat losses achieved by design, e.g., of only 140 kW for an 8-tonne-furnace, would thus rise to 400 kW which means an additional consumption of 4 kW per minute of open lid time. Over 20 minutes this would add up to as much as 80 kWh per charge, equivalent to an additional consumption of 10 kWh/t. Moreover, energy will also be 'sucked off' the furnace unnecessarily if the exhaust system is run at full capacity even at times when no, or only little, flue gas is produced. Under unfavourable circumstances this may increase the power consumption by as much as 3 %, corresponding to 15 kWh per tonne of iron. Another issue is superheating of the iron, if one considers that a 50 K temperature rise will consume about 20 kWh per tonne. The melting processor allows the final temperature to be maintained to an accuracy of 5 K, eliminating any unnecessary input of superheating energy. 5. Refractory lining The wall thickness of the ceramic furnace lining, which in cast iron melting systems will almost invariably be quartzite, always constitutes a compromise between good thermal insulation, adequate mechanical protection of the coil, and good electromagnetic coupling between the coil and the charge. Reducing the thickness of this lining will increase the coil efficiency and power consumption while causing higher thermal losses through the thinner crucible wall. However, since coil losses exceed the thermal losses across the crucible wall by nearly a factor of 10 in terms of magnitude, coil losses remain the dominant influence here. Studies have shown [1] a substantial reduction in power consumption with decreasing thickness of the refractory lining, as shown in Table 1. With increasing furnace operating time and thus increasing refractory erosion the power consumption will decrease by nearly 10 % over the first three weeks. Assuming that a lining having an original thickness of 125 mm loses 30 mm of that thickness within the first three weeks, the coil efficiency would rise by 3 % only according to our calculations (Fig. 5). It follows that this fact alone cannot explain the above-mentioned energy savings, which are presumably augmented by the increased power input and the resulting shorter melting time. It might therefore make sense to consider eliminating excessively high "safety margins" on the thickness of the refractory lining with the aid of advanced crucible monitoring equipment such as the OCP optical coil protection system [7].

20/7

Further reduction of electrical losses Continuous progress in the design of inductive melting equipment has yielded a considerable increase in power densities and output, apart from widening the range of technical applications potentials significantly. Much work has also been done to cut thermal and electrical losses through optimized furnace design and improved frequency converter technology. Superior efficiencies of more than 75 % in cast iron melting are the reward of these efforts and now define the state of the art. Further reductions in power losses have been attained through the new energy saving concept developed over several years of intense R&D and tested for its reliability and practical viability in extensive trials. The engineers focused their efforts on addressing the largest loss factor, i.e., electrical losses in the induction coil, in order to cut this source of inefficiency by a significant margin. Extensive calculations and numerous model trials were necessary to progress the initial concepts into a valid solution, viz., a special coil design combined with advanced frequency converter technology (Fig. 6). The new system was integrated into a 1.5-tonne coreless furnace in a stainless steel foundry and subjected to gruelling tests in day-to-day production operations over several months. The substantial energy savings achieved, in conjunction with high dependability and performance levels, attest to the successful achievement of our design targets and prove the system's industrial viability. The reduction in energy consumption can be put at 5 - 10 %, depending on specific operating conditions. Conclusions The actual average energy consumption in cast iron melting operations, at over 700 kWh/t, holds a significant saving potential if one considers the fact that figures of 490 - 520 kWh/t are now achievable in practical operation. The factors contributing to the excess consumption are exemplified in Table 2.

20/8

References 1. Efficient melting in coreless induction furnaces, Good practice guide

No 50: ETSU, Harwell, Didcot, Oxfordshire, 2000. 2. Jolivot R, Fonderie Fondeur d`aujourd`hui, No 229, Nov 2003,

pp36-39. 3. Foundry Management &Technology 131 (2003) No 11, pp14-16. 4. Smith L and Bullard H W, The Foundryman 88 (1995) No 7, pp246-253. 5. Brockmeier K-H, Induktives Schmelzen, Brown, Boveri & Cie,

Aktengesellschaft Mannheim; Essen: Giradetverlag 1966 6. Duca W J, Trans. Amer. Foundrym. Soc. 81 (1973) pp108/109 7. Donsbach F, Schmitz W and Hoff H, Giesserei 90 (2003) No 8, pp52-54 Tables Table 1

Table 2

20/9

Figures

Fig. 1 Influence of charge packing density on electric power

consumption

Fig. 2 Power diagram of a 12-tonne-furnace plant

Fig. 3 Influence of operating mode on coil efficiency

20/10

Fig. 4 Stored heat and holding power as a function of furnace size

Fig. 5 Influence of lining thickness on coil efficiency Fig. 6 The new system