THE1 UNIVERSITY OF MICHIGAN INDUSTRY PROGRAM OF TEE COLLEGE OF ENGIEERING COMPUTER PREDICTION OF CUPOLA PERFORMANCE UTILIZING OXYGEN AND NATURAL GAS IN THE BLAST Rbbelrt D. Pehlkb March, 1964 IP- 663

ACKNOWLEDGEMENT The contribution of IBM 7090 time by the University of Michigan Computing Center is gratefully acknowledged.

TABLE OF CONTWENS Page ACKNOWi.DG. a el........ a J........ * 4r. F..laUIV..... ii LIST OF TABLES......**......***.**O.....*.V...* *.*...**.4.. iv LIST CF FIGES................................... v INTRODUCTI ON *............................................. 1 TBE IIEMOCtEMICAL MODELS a*TJ*a aS... s.............. 2 COMPUiTER PREDICTIONS................................. 3 XYG ERICNI. I..................... NAT AL GAS INJECTION......... 4 BLXAST EMPERA.URE..................r....J 11 SMUITTAEOUS ENRICfEMENT, INJECTION, AND BIAST PRE1EAT... 13 ADIABATIC FLAME TEMPERAT URE CONCEPT.........1........ 16 VARIATION IN OPERATING COINDITIONS......................... 19 CONCLUSION....................................... 25 BIBLIOGRAPRY.................... 26 iii

LIST OF TABLES Table Page I Influence of Oxygen Injection on Coke and Melting Rates........ a*.............a......*.. 9 II Influence of Natural Gas Injection on Coke and Melting Rates.................... 10 II: Influence of Blast Temperature on Coke and Melting Ratess................... 12 IV Effect of Operating Variables on Cupola Coke Requirement and Melting Rate as Calculated by Thermochemical Model................. 21 iv

LIST OF FIGURES Figure Page 1 Calculated Effect of Oxygen Enrichment on Coke Requirement................5....... 5 2 Calculated Effect of Oxygen Enrichment on Melting Rate.....6............................. 6 3 Calculated Effect of Natural Gas Injection on Coke Requirement.....7.................... 7 4 Calculated Effect of Natural Gas Injection on Melting Rate................................. 8 5 Calculated Effect of Simultaneous Oxygen Enrichment and Natural Gas Injection on Coke Requirement.................... 14 6 Calculated Effect of Simultaneous Oxygen Enrichment and Natural Gas Injection on Melting Rate.................................. 15 7 Calculated Adiabatic Flame Temperature in Tuyere Zone for Various Levels of Blast Requirement.................................. 18 8 Effect of H2/H20 Ratio in Stack Gas on Coke Requirement with Natural Gas Injection.......... 23 9 Effect of {2/H20 Ratio in Stack Gas on Melting Rate with Natural Gas Injection................. 24

INTRODUCTION Considerable interest has recently been generated in the use of natural gas as a cupola blast injectant. As a potential replacement for part of the scarce high-quality metallurgical coke, natural gas injection could represent a considerable economic saving in cupola operations. Plant tests have been carried out to study the effect of natural gas injection in the cupola blast, but very few results have been reported in the technical literature. In contrast, however, the use of oxygen to enrich the cupola blast has been explored and reported quite extensively.8 Independently, the use of a high temperature blast to effect coke savings and increase production rates has led to the development of the hot blast cupola. This is a particularly significant technological development in the cupola melting field, and has been studied in a number of investigations.19'20 Cupola blast injection with oxygen and/or natural gas, simultaneous with higher blast temperatures, potentially offers a means by which a considerable increase in production and a substantial savings in coke consumption can be realized with current operating equipment. A large number of studies have been implemented to evaluate the effects of blast enrichment, injection, and temperature changes on cupola performance. The difficulties in carrying out an experimental study on production-sized equipment are particularly troublesome in view of the fact that a considerable variation in most of the operating conditions is experienced throughout a trial period. It is possible to overcome some of this difficulty and provide a more

-2consistent analysis of the experimental data when mass and energy balances on the experimental period are utilized in conjunction with the plant test results. To date, however, very few complete complete and detailed plant data have been published which permit accurate mass and energy balances to be made. In an effort to stimulate interest among foundrymen in ultimately providing for a full study of any operating variable on the cupola furnace, and to execute a preliminary prediction of the coke savings and production increases possible with oxygen enrichment and natural gas injection, a computer study of these variables has been carried out utilizing a previously developed Thermochemical Model.21 THE THERMOCHEMICAL MODEL The Thermochemical Model utilizes a mass and energy balance for the cupola melting furnace and a kinetic relationship based principally on the wind rate in order to provide a mathematical description of furnace operation. The Model utilizes prior furnace performance data to establish a base period of operation. Thermochemical equations employing a mass and energy balance are then applied to predict coke consumption and melting rate changes for different operating conditions. This Model has been previously discussed in some detail and its derivation outlined.21 The predictions made by this Thermochemical Model are dependent upon having the general operating conditions of the cupola furnace remain fairly constant. Both physical and chemical limitations

-3exist which control cupola operations, and it is these conditions which must be satisfied for a meaningful prediction in terms of attainable operating conditions. Consequently, the Thermochemical Model must be used with discretion and with a careful consideration of the feasible operating conditions for a specific cupola operation. This approach should provide a reasonable basis for predicting the conditions under which optimum cupola performance can be realized. COMPUTER PREDICTIONS The use of the Thermochemical Model in predicting cupola performance has been discussed in some detail; and coke requirement and melting rate changes with oxygen enrichment, natural gas injection, and variable blast temperatures have been reviewed.l In the present discussion, however, these preliminary calculations have been considerably extended and the mutual effects of these three operating variables have been considered. In executing these calculations, the reference period data are taken from the report of Sub-committee TS 522 Qualitatively a change in operating condition should provide a similar effect regardless of the size of operation, but quantitative agreement may not be precise. The predictions recorded below should therefore be considered as pertaining only to the small cupola operation from which the reference data were taken; a similar result however, would be expected on larger operating units. Prior to discussing the combined effect of oxygen enrichment, natural gas injection, and blast temperature, the individual contribution of each variable will be considered.

-4OXYGEN ENRICHMENT Plant and laboratory tests of cupola operation with oxygen enrichment (see Table I) unanimously concur that oxygen enrichment will reduce coke rate and provide a higher melting rate. Calculations of predicted cupola performance with oxygen-enriched blasts were carried out using the Thermochemical Model. The results of these calculations are presented in Figures 1 and 2. Figure 1 demonstrates a marked decrease in coke rate with increasing oxygen enrichment, as indicated by the relative postions of the respective curves of 0%, 2%, and 4% oxygen enrichment. In Figure 2 the positions of the oxygen enrichment curves indicate that melting rate will increase with increasing oxygen enrichment. NATURAL GAS INJECTION The predicted coke and melting rates of 0%, 2O and 4% natural gas injection were evaluated for various blast temperatures. The results of these calculations are presented in Figures 3 and 4. In Figure 3 the coke rate is shown to decrease with increasing natural gas injection, as indicated by the respective positions of the curves. In Figure 4, the curves have the same relative positions that they held in Figure 3. This indicates that the production rate will slightly decrease, as shown by the proximity of the curves, with increasing levels of natural gas injection. The reported plant trials (Table II) on natural gas injection indicated in two cases 3'24 that the use of natural gas (amount undisclosed) would markedly decrease the coke rate. In a more completely reported

-5-5 -10 i\ z zo -3 w -05 P% 0w n- -20 - z w -25 z tX -30 -35 -40 0 200 400 600 800 1000 BLAST TEMPERATURE,0F Figure 1. Calculated Effect of Oxygen Enrichment on Coke Requirement.

70 60 50 o'al -loo 0 40 w z 0-0 30 z 20 0080 a. -10 I BLAST TEMPERATURE,OF Figure 2. Calculated Effect of Oxygen Enrichment on Melting Rate.

-70 I I -5 -10 IZ w w -15 o 0 o -20 z -25. — z w -30 w -35 -40 I 0 200 400 600 800 1000 BLAST TEMPERATURE, OF Figure 3. Calculated Effect of Natural Gas Injection on Coke Requirement.

-850 40 w z " -20 - M20ltn2 % NATURAL GAS z oD 010 I10 Iz o 00 a-I0 I 0 200 400 600 800 1000 BLAST TEMPERATURE, OF Figure 4. Calculated Effect of Natural Gas Injection on Melting Rate.

-9TABLE I INFLJUENCE OF OXYGEN INJECTION ON COKE AND MELTTING RATES Change in Change in Reference Coke Rate Melting Rate % ~2 Injection 1 0 4%0 2%o 1 0 20 to 25p 2%** -3 0 2%* -22 0 2%** 1 -- 25% 2.5-3.0 2 -1.3% 25% 1.4-4.5 3 0 9.0 4.5 4,5 0 19.4 8.6 4, 5 ~ 40.2 13.8 4,5 -11.5 to 12.5% nil 1.0 6 -8.3 to 15.4% +16.7% 1.7 nil +20.Oo 2.8 7,8 -4% +100o 5% * Direct oxygen enrichment Oxygen injected in cupola well

-10TABLE II INFLUENCE OF NATURAL GAS INJECTION ON COKE AND MELTING RATES Change in Change in Reference Coke Rate Melting Rate g Natural Gas 23 -50% -- 24 -50o 25 -27.5 5* 4.7 -46.5 16* 4.52 -46.5 18.2* 4.34 -40 to 50** 15 to 20** 4.5** * Corrected to reference blast rate ** Average values

-11series of tests, the coke rate was observed to decrease on the order of 40-5X%, whereas the melting rate increased by 15-20%, at the 4 1/2% natural gas level. The results predicted by the Thermochemical Model are much more conservative than these reported plant results and are based on unchanged operating conditions. There is, however, some indication that improved combustion efficiency and top gas temperatures may be achieved with changes in blast conditions. The conclusion that can be reached from the computer prediction combined with plant tests is that a coke saving can be realized with natural gas injection, but that the improvement in melting rate is less than the corresponding improvement in coke rate; moreover, in the limiting case there may even be a slight decrease in the melting rate. BLAST TEMPERATURE The curves of Figures 1 and 3 indicate that the coke rate sharply decreases with increasing blast temperature whereas in Figures 2 and 4 the melting rate is shown to increase markedly with increasing blast temperature. These results are in good agreement with numerous plant investigations on the hot-blast cupola. Table III summarizes the results of various plant and laboratory investigations of the influence of blast temperature on coke and melting rates. These studies are almost unanimous in their agreement that increasing blast temperature decreases coke rate and increases melting rate. A statistical average of the results reported in Table III reveals that a blast temperature on the order of 500CC (9300F) should decrease the coke rate by 25% and increase the melting rate by approximately 300.

-12TABLE III INFLUENCE OF BLAST TEMPERATURE ON COKE AND MELTING RATES Change in Change in Hot Blast Reference Coke Rate Melting Rate Temperature 1 0 +13.5% 5200C 1 33% 0 5200C 9 -30%o 0 450-5000C 9 o +30 to 50%o 45o-5000C 10 -8% -- 600~C 11 -35 to 50 -- 12 -- +4% 5000C 13 -22 to 30%o -- 5000~C 14 55 +50% 6o00 15 -- 15 to 20o 16 -33% 33% --- 17 -27% * 500-550~C 17 -28% * 430~C 18 nil 20 400~C 18 -14.4 17 4700~C 18 -19.o 22 470~C 18 -27.0 25 550~C 18 -28.7 33 45o0C 18 -33.0 20 45o0C 18 -6.o nil 500~C 18 -25.5 50 3900~C 18 -14.2 17 390~C 18 -28.7 49 500~C 18 -29.5 21 500-6oo0c 19 +33% -12.5 4oo-500~C 19 -33% +33 500~C 19 -23 nil 5000C 19 -12 nil 520 C 19 -20 +12.5 520~C 19 -29 nil 525~C 20 -28 +25% 320~C * Increased corresponding to decrease in coke rate.

A comparison of these average results with the zero injection level curves of Figures 1 and 2 indicates that there is a good correspondence between the computer prediction and the average results obtained in practice. SIMULTANEOUS EJRICHMENT, INJECTION, AND BLAST PREHEAT The simultaneous effects of oxygen enrichment, natural gas injection) and blast preheat are not directly calculable by summation of the individual effects of each blast modification. However, this technique of totalling the individual effects will provide a reasonable estimate of the combined influence of various injections on cupola coke requirement and melting rate. Calculations were made of the simultaneous effect of oxygen enrichment and natural gas injection; in Figures 5 and 6 the results at the 2% and 4% levels of each constituent are presented as a function of blast temperature. The simultaneous employment of oxygen and natural gas offers considerable opportunity for increasing the melting rate and decreasing the coke requirement. A further advantage in the control of the cupola melting process could be realized through injection rate, which would provide a control technique to which the process would rapidly respond. Of even greater importance may be the necessity for simultaneously using oxygen and natural gas to obtain optimum performance with practicable operating conditions as discussed in the following section.

H- -10 z O 0dL L\0 ~ W.1.1 -20 - w ~ i 0 ~z ~~ adarGa e I0z -25 0 0r -30 4. -35 -40 I I 0 200 400 600 800 I 000 BLAST TEMPERATURE, OF Figure 5. Calculated Effect of Simultaneous Oxygen Enrichment and Natural Gas Injection on Coke Requirement.

-1570 60,0 o40 300 w w I0 a W0 2 10 - I 0 200 400 600 800 1000 BLAST TEMPERATURE,) Figure 6. Calculated Effect of Simultaneous Oxygen Enrichment and Natural Gas InJection on Melting Rate.

ADIABATIC FLAME TEMPERATURE CONCEPT The operating conditions attainable with oxygen and natural gas injection are not well established. However, one of the physical limitations which is placed on the coke-fueled shaft furnace (such as the cupola or blast furnace) is the necessity for a specific temperature distribution to exist in the melting zone just above the tuyeres. As oxygen or natural gas is injected into the blast, the temperature distribution in the melting zone will change. To enable the furnace to accommodate these changes in blast composition and maintain the desired temperature distribution above the tuyere zone, the concept of adiabatic flame temperature can be adopted. The adiabatic flame temperature is assumed to relate to the combustion of the blast with the downwardly moving coke in the zone just in front of and above the tuyeres. If the furnace has been satisfactorily operating with a given burden and blast composition and temperature, the adiabatic flame temperature can be computed for these conditions, assuming an incoming temperature for the downwardly moving coke and neglecting temperature changes of the metal and slag components as they move through this zone. The combustion products at the flame temperatures encountered in normal plant operations can be assumed to be carbon monoxide and hydrogen along with the unchanged nitrogen. If a suitable flame temperature for a particular furnace has been determined, it is also possible to compute the adjustments in blast composition and temperature to maintain this flame temperature. Several possiblilties exist. For example, the blast could be enriched

-17with oxygen and a compensation made for the increased flame temperature by adding moisture or natural gas. Alternatively, the blast temperature could be increased, which would increase the flame temperature, and a compensation made by adding moisture or natural gas. This concept has been very useful as a guide in estimating the ranges of potential blast conditions for the iron blast furnace. 2627 Futhermore, cupola refractory life could be extended by controlling the temperature distribution at the tuyeres, and in addition some of the initial difficulties encountered with injection in the cupola could be better controlled.7 The injection of cold natural gas at the tuyeres decreases the temperature in the melting zone because of the decomposition of the gas (approximately 9~0 methane) to carbon and hydrogen. Since the ensuing combustion in the presence of hot coke does not proceed beyond carbon monoxide and hydrogen, the overall result is that less energy is developed within the melting zone with injection. Consequently, natural gas injection should be accompanied by oxygen enrichment or an increased blast temperature to maintain a given melting zone temperature. Figure 7 presents the effect of natural gas injection and oxygen enrichment on the adiabatic flame temperature at the tuyeres. These calculations are based on a constant moisture level in the entering air of 5.5 grains/ft3 and a coke temperature of 2800A. The curves should prove useful for comparative purposes, although the actual accuracy

-183800, 5.5 GRAINS MOISTURE PER FT. AIR COKE AT 2800 OF 3600 3400 001 Q 2800 2600 2400 2200 0 300 600 900 1200 BLAST TEMPERATURE, F Figure 7. Calculated Adiabatic Flame Temperature in Tuyere Zone for Various Levels of Blast Enrichment.

-19of the calculated flame temperatures is limited by the accuracy of the assumptions on which they are based. It is reasonable, however, to assume that blast conditions which produce an adiabatic flame temperature in the same range as those which are already achieved with good operating conditions would provide a smooth cupola practice. Recognizing that the foregoing analysis is presented only as a useful guide, another possibility might be considered upon which satisfactory operating conditions could be estimated. This prediction of operating conditions could be based on the oxidation level, or oxygen potential, of the blast to provide a consistent metalloid recovery to the iron. Oxidation loss at higher levels of oxygen enrichment may represent a limit to the extent of economical oxygen utilization in 28 cupola melting. Consequently, an analysis of the oxygen potential of the blast could possibly be based on a mass balance and a dynamic model for combustion at the tuyeres. This model might be developed to provide a basis for predicting levels of blast injection which are acceptable in terms of temperature and chemistry of the slag, metal, and gas phases in the tuyere zone. VARIATION IN OPERATING CONDITIONS The predicted results when utilizing oxygen and natural gas in the blast are based on an assumed set of operating conditions. The normal range of conditions realized with a present cupola operation may be markedly changed when the blast conditions are altered. Changes in metal temperature, metal temperature metal compositions CO/CO2 ratio metalloid recovery, and top gas temperature can result when blast modifications

-20. L1 4817-20)22p29-53 are made4'8'70''9 Furthermore, the technique employed in making a blast modification is very important in controlling changes in operating conditions. For example, marked differences in metal temperature and metal carbon and silicon contents were noted when 2% oxygen was injected into the cupola well, as compared to enriching the blast by 2%. In order to provide a means for considering these changes in predicting cupola performance, computer calculations were carried out over a range of operating conditions in order to determine the influence of each operating variable on the coke requirement and melting rate. Table IV summarizes the results of these calculations in the form of linear coefficients, which may be used to estimate the changes in coke and melting rates brought about by assumed changes in operating variables. Hence, if the cupola operator has a reasonable estimate of changes which may result in operating variables with a given set of blast modifications, the predicted cupola performance can be modified to account for these accompanying changes in furnace operation. The coefficients of Table IV were found to be nearly linear over a wide range of operating conditions. Calculations involving simultaneous changes in several variables showed the effects to be additive. Consequently, the coefficients can be utilized to make a reasonable estimate of corrections for changes in operating variables over the entire range of conditions encountered in standard commercial operations. One other variable which becomes particularly significant as the level of natural gas injection is increased is the H2/H20 ratio in the stack gases. In normal operation, this change in stack gas composition is of secondary influence, important only as the moisture level in the

-21TABLE IV EFFECT OF OPERATING VARIABLES ON CUPOLA COKE REQUIREMENT AND MELTING RATE AS CALCULATED BY THERMOCHEMICAL MODEL CHANGE IN VARIABLE PERCENT CHANGE IN PERCENT CHANGE IN COKE REQUIREMENT* MELTING RATE* Decrease in slag volume of 10 lb/NT Metal -1.1 +1.0 Increase in Steel scrap in charge (substituted for pig iron) of 100 lb/NT Metal +4.1 -1.8 Decrease in blast moisture from 5.5 to 4.5 gr/ft -1.0 +0.5 Decrease of 0.1 in CO.C02 ratio in Stack Gas (1.17 to 1.07) -3.2 +1.8 Increase of 1% in carbon content of coke -1.4 +0.2 Decrease in Stack Gas Temperature of 100~F -4.3 +4.5 Increase in slag temperature of 1000F +.3 -0.3 Increase in metal temperature of 1000F +3.2 -3.2 Increase of 0.2% in carbon content of metal +2.4 -0.7 Increase of 0.2% in silicon content of metal +1.8 -1.0 Increase of 0.1% in manganese content of metal +0.5 -0.4 Increase of 0.05% in phosphorous content of metal +1.0 -0.7 * Based on reference period of operation Coke rate - 265 lb/NT Ton Metal Melting rate - 6.425 Net Tone Metal/hr. Slag Volume - 136 lb/Net Ton Metal

blast changes. In operations with natural gas injection, however, the amount of hydrogen in the stack gases as moisture or as hydrogen gas is increased. Figures 8 and 9 present the influence of variations in the stack gas H2/H20 ratio on cupola coke requirement and melting rate. At lower H2/H20 ratios, the coke rate is found to decrease whereas the melting rate increases. A lower ratio indicates a higher H20 content in the stack gas and hence an increase in the thermal energy released in the furnace.

-230l,,,,, CONSTANT CO/CO2 RATIO IN STACK GAS CONSTANT BLAST VOLUME RATE 2% NATURAL GAS 4% NATURAL GAS -10 - z 2 1 -15 BLAST TEMPERATURE 0 0 Figure 8. Effect of H2/1f20 Ratio in Stack Gas on Coke Requirement with Natural Gas Injection.

30 CONSTANT CO/C02 RATIO IN STACK GAS CONSTANT BLAST VOLUME RATE 25- 2% NATURAL GAS L.4% NATURAL GAS OC z 20' BLAST TEMPERATURE Z |-_500 F I".....3.....500 F 0 10 VanI1 5 2. 0 2.5 300 1.0 1.5 2.0 - 2.5 3.0 3.5 4.0 H2/HzO2 RATIO IN STACK GAS Figure 9. Effect of H2/1H20 Ratio in Stack Gas on Melting Rate with Natural Gas Injection.

CONCLUSION The use of the Thermochemical Model has been shown to provide a means for testing plant data for consistency and general accuracy, and is a particularly powerful means for searching out general areas in which optimization of cupola operations should be studied using plant trials. The use of adiabatic flame temperature of the blast as a guide in choosing realistic operating conditions has been discussed. The combined utilization of this concept with a thermochemical model of the cupola furnace presents a new horizon for exploration in the continued development of the cupola process. The thermochemical study of cupola operation utilizing oxygen enrichment, natural gas injection, and blast preheat has shown that oxygen enrichment and blast preheat both provide substantial coke savings and production increases. The use of natural gas in the blast is an effective means for saving coke, but alone provides a slight decrease, if anything, in metal production rate. However, the combined effects of enrichment, injection, and increased blast temperature offer a multitude of operating conditions under which the maximum economic potential of the cupola can be realized.

BIBLIOGRAPHY 1. H. J. Leyshon and R. B. Coates, "Acid-Lined Cupola Operation with Cold Blast, Cold Blast with Oxygen and Hot-Blast," B.C.I.R.A. Journal, 10, 1962, p. 28. 2. Francois Danis, "Oxygen in the Cupola," Fonderie, 145, 1958, p. 79. 3. F. Morawe, "Enriching the Blast of the Cupola Furnace with Oxygen," Giesserei, 17, 1930, p. 132 and 155. 4. W. C. Wick, "Cupola Operations Improved with Oxygen-Enriched Blast," American Foundryman, 13, 1948, p. 64. 5. W. C. Wick, "Oxygen-Enriched Cupola Blasts," Trans AFS, 56, 1948, p. 246. 6. F. J. Webbere, "Auxiliary Oxygen Applied in 72 Inch Production Cupolas," American Foundryman, 17,(June, 1950,)p. 40. 7. A. K. Higgins, Discussion 1 to Ref. 25, Trans. AFS, 56, (1948), p. 256. 8. A. K. Higgins, "Oxygen Enrichment of the Cupola Blast," Iron Age, 161, 1948, p. 72. 9. Fritz Schulte, "Recent Development of Cupola Design with Special Reference to Hot-Blast," Foundry Trade Journal 92, (1952) p. 405. 10. R. J. Sarjant, "Fuel and Metal," Proceedings, Institute of British Foundrymen, 45, 1952, A-24. 11. F. I. Smirnov, "Producing Super Heated Iron in a Magnesite-Lined Cupola Run on a Hot-Blast," Liteinoe Proizvodstve, Oct, 1956, p. 10. 12. M. K. Sin, "Effect of Blast Preheating and Coke Consumption on the Operation of the Cupola," Liteinoe Proizvodstvo, Jan, 1962, p. 32. 13. A. Upmalis, "Hot-Blast Foundry Cupolas," Tekn. Tidskr., 83, 1953, p. 407. 14. M. Bader, "The Hot-Blast Cupola," Foundry, 74, no. 12, 1946, p. 104 and 252. 15. "Hot-Blast Package Reduces Melting Costs," Canadian Metals, 20, April, 1957, p. 46.

-2716. B. Thyberg, "Operating Experience with Hot-Blast Cupolas," Tekn. Tidskr, 84, February, 1954, p. 77. 17. "First Report of Sub-committee TS 43: Cupola Developments," Proceedings of the Institute of British Foundrymen, 47, 1954, p.A38. 18. "Second Report of Sub-committee TS 43: Cupola Developments," British Foundryman, 50, 1957, p. 116. 19. "Third Report of Sub-committee TS 43: Cupola Developments, British Foundryman, 53, 1960, p. 93. 20. F. K. Vial, "Hot-Blast Applied to the Cupola," Iron Age, 120, 1927, p. 1072. 21. R. D. Pehlke, "Thermochemical Model for Computer Prediction of Cupola Performance," Modern Castings, 44, (1963) p. 580. 22. "Report of Sub-committee TS 52: Developments in the Melting of Metals for Foundries," British Foundryman, 54, 1961, p. 105. 23. S. I. Tsukerman and Yu. G. Rozenberg, "Cupola Fired with Coke and Natural Gas," Liteinoe Proizvodstvo, July, 1959, p. 28. 24. C. Stefanescu, T. Stefleu, and S. Teodoresdeu, "The Application of Natural Methane in the Treatment of Cast Iron in the Cupola Furnace," Met. Constructia Masini, 11, 1959, p. 672. 25. N. A. Voronova, Yu. N. Khil'hlein, 0. A. Mogilevtsev and V. N. Danilets, "Use of Natural Gas in Large Cupolas," Russian Casting Production, 1962, p. 492. 26. J. A. Cordier, "Injection of Different Materials in the Blast," Blast Furnace, Coke Oven, and Raw Materials Proceedings AIME, 19, 1960, p. 238. 27. H. N. Lander, H. W. Meyer, and F. D. Delve, "Prediction of Blast Furnace Performance from Operating and Thermal Data," Blast Furnace, Coke Oven, and Raw Materials Proceedings AIME, 19, 1960, p. 219. 28. J. M. Crockett, Discussion 2 to Ref. 25, Trans. AFS, 56, (1948), p. 257. 29. "Hot-Blast for Small Foundry Cupolas," Mechanical World and Engineering Record, 18, 1958, p. 362. 30. E. L. W. Loebbecke, "Development of Hot-Blast Cupola Melting Technique in Europe," Foundry Trade Journal, 101, 1956, p. 697.

-2831. J. V. Harding and J. A. Charles, "The Use of Oxygen in Cupolas, " The British Foundryman, 54, 1961, p. 365. 32. E. C. Evans, "Oxygen Enrichment in the Cupola," B.C.I.R.A. Journal, 3, 1949, p. 109. 33. S. C. Clow, "Twenty Years Progress in Cupola Melting," Foundry, 91, 1963, p. 44.