Document v17N9ZYjwa7GzKkEZKXV67MgE

PLAINTIFF'S EXHIBIT Thermal Conductivity of Castable Refractories in Relation to Bulk Density RICHARD W. WALLACE and G. HUGH CRISS Haribson-Walker Refractories Co. Garber Research Center, Pittsburgh, Pa. Thermal conductivity measurements of castable refrac tories of widely different compositions and densities confirm that conductivity is related to bulk density. General equa tions have been derived by digital computer which allow - \ prediction of thermal conductivity from the bulk density of the dried castables. efractory castables, mostly based on grain aggre gates of alumina-silica compositions, have developed wide spread use in a great variety of industrial furnaces. With few exceptions the bonding constituents are calcium aluminate cements; these are available in several composition ranges. With various methods of installation, industry has at its disposal versatile materials that can be used for repairs or incorporated in original designs. Reliable data on physical properties are necessary for ap plications under exacting conditions, but the situation is com plicated by the diversity of both composition and installation methods. In the present study of thermal conductivity of castable refractories it was possible to relate this property closely to bulk density of the dried product. The thermal conductivity of castables is a function of product thermal history and bulk density. This has previously been discussed in publications by Hansen and Livovich,1 Ruh and Renkey,2 Wygant and Crowley,1 Bohling and Stanford4 and Modem Refractory Practice.1 We have here extended and studied in greater de tail the generalization of product density as related to thermal conductivity. Presented at the Sixty-Ninth Annual Meeting, the American Ceramic Society, X'ew York, X. Y., May 3, 1967 (Refractories Division, No. 32-R-67). Received May 15. 1967; revised copy received August 14, 1967. 176 Description of Materials Twenty-five commercially available castables were selected for thermal conductivity tests: twenty-one alumina-silica including high alumina and insulating castables, and four chrome based castables. Properties of these mixes are listed in Tables I, II and III. Of the chrome castables, mix Y was chemically bonded and mixes V, W and X were hydraulically bonded with calcium aluminate cements. Specimens were cast in 9 by 4'/i by 2`/j in. molds as described in ASTM designation C 268-65T. After curing 24 hr at 75F in covered molds, the test specimens were removed and dried 18 hr at 230F. Bulk density was then determined prior to testing in the thermal conductivity equipment. Test Equipment Ma^surments were made on the ASTM C 201-47 thermal conductivity equipment. This apparatus is a guarded calorimeter-furnace unit which is generally accepted to yield data with 5% precision. Heat conducted from the furnace through the test specimen is quantitatively determined by a 3 by 3 in. test calorimeter. Test Procedure For refractory castables it is now generally agreed that tests on dried unfired specimens after heat treatment in the ASTM apparatus represent the proper method of obtaining thermal conductivity data for use in engineering design. Bv this method the castables are heated from one face only Ceramic Bulletin Symbol; A Max service temp. F: Approx AI?Oj content, `/r' Bulk density, pcf after di ying at 230F: Thermal con ductivity, k ( Btu hr-1 ft F " in.) HM)0 5) 29 op 173 280 337 478 559 635 804 859 1098 K 0 85 0 86 0 93 0 96 1 00 1 01 1 06 1 05 1 14 Table I. B 2000 -10 Insulating Castable Properties and Test Results C 2200 32 D >200 37 / 2000 53 / Extra strength type 2300 34 34 52 57 52 84 op 249 370 458 657 1020 1374 1063 802 698 536 264 K 1.61 1 44 1.32 1 37 1.38 1 48 l 31 l 23 1 21 1 17 1 01 F 399 801 1280 1102 915 610 304 K 2. SO 2.21 1 80 1.72 l 64 1 56 1 50 F 400 783 1255 953 706 460 K l 68 1 78 1 72 1.61 1.50 t 42 F 275 435 531 868 1131 1441 1922 1503 1141 845 595 335 K 2.72 2. 10 1.73 1.90 1.73 1 92 2 19 1.92 1.77 1.61 1.51 1 37 F 395 868 1400 1198 933 607 290 K 3.24 2.90 2.67 2.63 2.54 2 49 2 30 G 2800 51 87 F 426 916 1449 1140 830 536 236 K 4.17 3.42 3.79 3.59 3.37 3.23 2.98 // 3300 90 93 F 510 1216 1778 1414 1048 686 310 K 6.63 4.90 5.07 5.09 5.40 5.90 6.97 resulting in the establishment of a temperature gradient through the castable as is the case in most industrial applica tions. Figure 1, mostly from Ruh and Renkev,2 shows an example of different results obtained from the various methods of heat treatment. It is clear from these curves that a significant difference in thermal conductivity occurs if the specimens are preheated to temperatures high enough to form ceramic bonding. In the work presented here, thermal conductivity measurements were taken both on heating and cooling after reaching equi librium at a number of temperatures. The hot faces of the test specimens were heated to a maximum of 100F below the recommended maximum service temperature of the castable or to 2500F, whichever was lower. Test Results Results of these measurements are listed in Tables I, II and III. For their further study the data were plotted on large scale graphs. From these graphs thermal conductivity values were taken at 500, 1000, 1500 and 2000F along the cooling curve and at 300 and 000F along the initial heating curve. These data for 20 alumina-silica castables were then analyzed 0 200 400 600 800 1000 1200 "C Fig. t. Thermal conductivity of extra-strength type flrecioy castable after various heat treatments. Table II. Alumina-Silica and High Alumina-Castable Properties and Test Results Symbol: / Type of castable: Max service temp, F: Approx Al*0* content, %: Bulk density, pcf after drying at 230F: High alumina 3100 62 140 F ` 458 1137 1735 1358 990 626 283 K 10 8 7 56 8.16 7.58 8.00 8 02 8.00 J Extra high alumina 3200 94 159 F 437 1173 1744 1360 980 622 269 K 21.8 9.73 9 74 10 0 10.9 11.7 13. 1 K Extra high alumina abrasion resistant 3200 93 160 F 563 1159 1814 1435 1034 637 274 K 13 3 11.1 10 4 10 8 11 5 13 0 13.3 L Fireclay 1000 34 126 op 279 399 557 465 364 261 K 8.08 7.63 7.40 7.37 ,.^7.03 7.24 M Fireclay 2500 40 123 op 489 1063 1646 1360 1002 642 292 K 5.89 5.15 5 77 5.70 5.54 5 50 5.13 N Fireclay extra strength type 2400 40 129 F 300 912 1784 1426 1056 68t 306 K 7.12 5.50 6.25 6.15 6.01 5.91 5.50 Symbol: Type of castable: Max service temp, 3F: Approx. AltOi content, %: Bulk density, pcf after drying at 230 F: Thermal conductivity, k (Btu hr"* ft "*F in.) P Fireclay extra strength Low iron 2600 46 127 F 492 1064 1719 1362 1010 645 268 K 7.85 6.11 6 86 6.80 6.66 6.61 6.46 Q Fireclay 2700 44 125 op 527 1183 1754 1388 1015 647 286 K 6 44 6 25 6.93 6.72 6 44 6.38 6 33 R Fireclay abrasion resistant 2800 55 142 F 473 1192 1794 1405 1021 636 276 K 11.8 9.3 10 0 10.1 10.1 10.3 10. 1 5 Fireclay 2800 46 129 op 524 1167 1768 1391 1019 650 293 1761 K 8.86 6 18 7.02 7.00 6 70 6.70 6.57 7 16 T Fireclay 2800 44 129 op 263 564 1012 1370 1747 1384 1025 651 287 K 10.95 8.34 6.12 6.18 6.78 6 64 6 43 6 38 6.18 O Fireclay extra strength coarse 2400 41 140 F 495 998 1663 1435 1049 660 293 K 9 42 8.53 8.03 8.13 8.04 8.03 7.95 U Fireclay coarse 2800 44 135 F 513 1163 1749 1394 1029 655 298 K 9.54 7.32 7.92 7.87 7.60 7.53 7.75 Vol. 47, No. 2 (1968) 177 Symbol: Type of castable: Max service temp, F: Bulk density, pcf after drying at 230F: Thermal conductivity, k (Btu hr-' ft-*F-> in.) Table III. Chrome Castable Properties and Test Results V Chrome high strength 2600 W Chrome 2600 X Chrome abrasion resistant 2600 166 170 176 F 555 1125 1819 1458 1092 699 310 K 7.49 6.30 6.66 6.82 6.90 6.87 6.56 F 464 1163 1789 1436 1066 687 304 K 8.84 6.53 7.06 7.10 7.14 7.22 6.87 F 497 1215 1870 1.500 1092 689 298 K 11.4 9.70 9.85 10.2 10.7 11 4 11.1 Y Chrome chemically lx mded Sim 174 "F 548 1202 1800 1448 1063 664 286 K 9 41 8 97 9.06 9.22 9.50 9.89 9 81 by a least squares regression routine on a digital computer* to obtain the relation of dried bulk density to thermal con ductivity. Based on the degree of fit of the higher order equations generated, it was seen that the third order equations adequately generalized the data. These equations are shown in Table IV. In Fig. 2 the equations from the measurements made on cooling are plotted. The four lower curves represent a generalization of all such measurements made at this labora tory on alumina-silica and high alumina castables with the single exception of mix H, a 3300F insulating castable con taining electrically fused bubble alumina (see Table II). This accounts for a higher conductivity than that of ordinary alumina-silica castables of equal bulk density. The other alumina-silica mixes tested, however, showed exceptionally good agreement with the computed curve. The average coefficient of variation of these curves was 6.4%. It is perhaps interesting to not-' that all of the alumina-silica curves passed through a common point at 137 pcf. This has a dual significance; namely, this is the density at which the slopes of the conductivity-temperature curves change from positive to negative and also, that an alumina-silica castable with this density would be predicted to have a thermal conduc tivity that is constant with temperature. The chrome curves shown in Fig. 2 will be discussed later. In instances where castables are to be used at temperatures so far below their recommended service temperatures that appreciable hydraulic bonding remains, the data and equations obtained from measurements made on cooling, and shown in Fig. 2, do not apply. For this situation the data obtained on heating were considered. The thermal conductivity values at 300 and 600F from the heating curves were analyzed by computer. Equations which resulted from this analysis are shown in Table IV and plotted in Fig. 3 along with the 500F cooling curve from Fig. 2. It is apparent that with very low service temperatures, the remaining hydraulic bond causes appreciable elevation of the conductivity-density function. In addition, due to the variable nature of the hydraulic bond we are not able to predict conductivity based on bulk density with as great an accuracy as we could had the refractory been heated to a higher temperature. Mixes V, W, X and Y were chrome based castables and consisted of hydraulically bonded as well as chemically bonded materials. Although it is clear that additional data are called for to properly analyze a conductivity-density relation it was decided to see if a reasonable curve would result from computer analysis. For this purpose the data taken at various * Honeywell 120 digital computer, Honeywell Inc. Fig. 2. Thermal conductivity vs. bulk density of dried castobles for various mean temperatures. (Data was derived from cooling curves.) 178 Fig. 3. Thermal conductivity vs. bulk density of dried castables for various mean temperatures with heating curve data compared with 500F cooling curve. Ceramic Bulletin Table IV. Equations for Deriving Thermal Conductivity from Bulk Density of Dried Alumina-Silica and Chrome Castables Genera! equations: For alumina-silica castables For chronic castables K = Ad*3 1*F* Bd* + Cd + D, Third order equation K = A'd* + B'd + C', Second order equation where K = Thermal conductivity, Btu/ft2 hr F/in. d = Bulk density, pcf A, A', B, b', c, a. and D = Constants Us at various temperatures On cooling 500F 1000F 1500F 2000F On initial heating 300 F 600 F On cooling 0 to 2000 F On initial heating 0 to 1000F A X 10-* +5.569 + 1.858 -0.859 +0.171 +21.62 + 11.91 A' X io- +5.201 +2.109 B X lO"4 5 -7.572 + 0.831 + 7.642 + 2.891 -45.23 -23.83 B' -1.385 -0.3843 c X 10-* +5.870 +o.;9 -5.015 +0.496 +33.73 + 18.72 C' +92.911 + 12.927 D -0.280 +0.955 +2.455 + 1.242 -5.720 -2.805 Cocllicicnt of variation 7.5 7.1 5.7 5.4 25.0 12.7 6.7 14.1 temperatures were averaged and a single analysis performed for both initial heating and cooling measurements. From a comparison of coefficients of variation, the second order formu las obtained from this analysis appear to be reasonable. These two formulas are shown in Table IV and the formula from the data taken on cooling plotted in Fig. 2. Summary Thermal conductivity measurements on 21 alumina-silica and four chrome castables have been presented. These data were generalized into formulas that can be used to predict thermal conductivity from the bulk density of dried castables. These formulas apply within the density range tested, namely, from 29 to 160 pcf for the alumina-silica castables and from 166 to 176 pcf for the chrome castables. Bulk densities on dried rather than burned castables were used in this study for two reasons: (1) these values will be easier to obtain in the field and (2) exact duplication of the thermal history experienced by the specimens measured in this study would be difficult. Other monolithic materials such as ramming and plastic mixes would not necessarily be expected to show agreement with the curves for castables presented in this paper. Al though some measurements have been made on these mate rials, no relation has yet been established except that the data for plastic' refractories seem to fall somewhat below those for castables. Acknowledgment The authors express appreciation to their associate L. Fraas for his assistance in making the measurements presented in this paper. References 1 W. C. Hansen and A. F. Livovich, "Thermal Conductivity of Refractory Insulating Concrete," J. Am. Ceram. Soc., 36 [11] 356-62 (1953). 1 Edwin Ruh and A. L. Renkey, "Thermal Conductivity of Refractory Castables," ibid., 46 [2] 89-92 (1963). 1 J. F. Wygant and M. S. Crowley, "Effects of High-Conduc tivity Gases on the Thermal Conductivity of Insulating Refrac tory Concrete," ibid., 41 [5| 183-88 (1958). 4 W. C. Bohling and J. C. Stanford, "Relationship of Thermal Conductivity Versus Bulk Density of Alumina-Silica Castables"; presented at the 65th Annual Meeting, The American Ceramic Society, Pittsburgh, Pa., May 1, 1963 (Refractories Division, No. 26-R-63); for abstract see Am. Ceram. Soc. Bull., 42 [4] 243 (1963). 5 Modern Refractory Practice, 4th ed.; pp. 323-24. Harbison- Walker Refractories Company, Pittsburgh, Pa., 1961. -fr Vol. 47, No. 2 (1968) The Authors Richard W. Wallace is re search physicist at the HarbisonWalker Refractories Co., Garber Research Center, Pittsburgh. His photograph and biographical sketch appeared on page 697 of the July 1967 Ceramic Bulletin. G. Hugh Criss is a research engineer at the Garber Research Center. He received his B.Sc. degree in geology from Ohio State University in 1960. Mr. Criss joined Harbison-Walker's Technical Sales Department upon graduation and in 1962 trans ferred to the research staff. His efforts have been devoted to the development of specialty refrac tory products. Babcock &Wilcox Refractories Division Product Information July 31, 1973 Page 1 of 2 B&W DENSE CASTABLES AVERAGE PROPERTIES 3 14 - 1 Attached are average properties released by the B&W Technical Development Laboratory covering B&W Dense Castables. This property sheet should be used in conjunction with Product Information Sheet 3 14 - 4 which covers the Factors Effecting the Use of Castable Refractories.3 3 14 - 1 Supersedes 3 14-1 dated August 30, 1972 Firebrick Insulating Firebrick * Refractory Castables Plastics Ramming Mixes Mortars Ceramic Fiber * Special Oxide Refractories Mineral Wool dated August 30,1972. Supersedes 3 1 4 - 1 i i*! n ,8Sgg a Sgg ,, ,g ,8, |8fSig | i|| > ;S 3,3 5 3 2 2 gg s ,,*gg ' MH...... it' MM...... i' HMM I I I I" I I M P- r O m i id e I d I ^ e ei z i ni.^w r1J1| . 8i xx 2 i l * , ! ISR9* * i * i 81851888 nil ; iii i ss o ^ a o O 5 |, ??S 11*11 Si SS * SSS11 1 1 * * ^1 f<f *71**1 * a s S " S 5 1 * 28 S"o s 5S*8" 5 ! i,ii.iis,|.. 3.???,??.'r ,,3 3=-5| 1 2 8 * "'ll 5 I' ; 1 s-1si 11 1 si 1 s 11 * * Po 1 M^(Mj> 1 M 1 ^ 1111 a ? Ou O M M <n 1 Oau art avaraga rauilta o f ta*t* conduced undar standard procadurai w ith east aOnpias axcapt as otharwisa notad and are subject to variation. Results should not be used for specification purposes. 1 S iii,S1111 i.iiilg.i.... M9'??'? 111*3 ^ | 'O NO ON 6N Op I 1 O* Oq %e jsagg 1 | .|| .|li mm f(nm ppM ^^ 8a 8 ^_ 1 I <7^ I 1 g d viol I n 8 -- ? ?is|i s i.is.ggs.siii !,!!,!!,|ggg ,99,9??,?:???,, 8s3ss15, , 3g ll: PI IH '"1 * 11,9, iii, ,lg I. > ,|, ,| 11 ill I..i?ii?..i9? 1 ggSSSdSii Sg ! ??lg| * !,,,,,l,M|g g,,,| 88 9 9*1 0p-M I I l g M I I^I I^I M 97 I g 50 o-M 6 o I l Oo 08 - aa ttm ' 1 ill illS 1 HI 1 |'II'gg1 ?5g' e 'm?m? 'Hippp mmm ' 1 3g b p md ps md rd II ! Ix ** ebojg.g ? i,iS,gii,gi|, I ill .ill ill! ,99,999,999. 8 ggsssss, , sg *** 5 ?Sjg| 5 18918*8 1 88 1 * s115***53 * 1 9oi^r^p*i^^>ar *97*$1 *i 3 t 9 6 m o 1 mp * 1 9doo u oc 111 ^ co tu Z0 Hi 2 0s 1g >< H r- dh" M M i IBsS I I I I I ............. ,99,999,,,., 8 , ,, 2- 3 j 3 . , ~ :| gss i g I,Si,IS....... 1,11 8 8O MM 00 91 S<* 1*1*11 ..................................... 9^ 9^ *M: 9O 9O .* ,> 90 eO *'*5 S*88 5 g.SS.SS...... gg.BB.SS...... eS.^,99..........:-33 3 , , 58 I a- 1t gsiggli & <,I5 _ f 3 Illt B* II 1m 1 |V I i ail k^zi t *is 2 115 1511 IP 1111 p " Co s* |0j$0 irm i ini ..ss- 5i " iHHnm; 1 sL 5 5 |p | CURING REFRACTORY CASTABLES IT ISN'T THE HEAT. IT'S THE HUMIDITY J. F. WYGANT AND M. S. CROWLEY RESEARCH AND DEVELOPMENT DEPARTMENT AMERICAN OIL COMPANY WHITING, INDIANA CURING REFRACTORY CASTABLES: IT ISN'T THE HEAT. IT'S THE HUMIDITY J. F. Wygant and M. S. Crowley Research and Development Department, American Oil Company, Whiting, Indiana ABSTRACT Curing methods for refractory concretes placed at low water content were studied. Curing was done (1) in a damp room; (2) by water spraying; (3) with saturated steam; and (4) by the application of impermeable surface membranes. Internal temperature rise caused by hydration of aluminous cements does not impair strength. The only requirement for satisfactory curing is the retention of moisture within the concrete. Properly selected organic resin or asphalt membrane coatings are economical and superior in performance to water-spray curing, provided they have low water-vapor permeability, adhere to moist refractory concrete, rapidly form films and dry, and are free from deleterious substances. Curing with membrane coatings has been used success fully for eight years in large-scale refinery operations. CURING REFRACTORY CASTABLES: IT ISN'T THE HEAT. IT'S THE HUMIDITY J. F. Wygant and M. S. Crowley Research and Development Department, American Oil Company, Whiting, Indiana INTRODUCTION The traditional method of curing troweled or pneumatically-placed refractory castables is frequent spraying with water from about 6 hours after placement to about 24 hours after placement. It is the only method generally recommended by refractories manufacturers. This procedure is costly and sometimes unreliable. Laborers must be employed for several consecutive shifts to do the spraying. Scaffolding cannot be removed, and vessels or furnaces cannot be closed up, until the spraying ends. In many locations the castable is inaccessible for repeated spraying. A less costly and more reliable curing method was desired. The exclusive use of water spraying has been justified on two bases. One is the need to keep the exposed surface of the castable damp, so that moisture will be available continuously for the hydration of the cement. This need for moisture has been confirmed (1). The other basis rests on the beliefs that temperatures much above 90F during hydration cause low strengths, and that the cooling effect of the water spray prevents temperature rise in the castable during hydration. The former belief was reinforced by the work of Hansen and Livovi-eh (2) , in which castable (1) Venable, C.R., "Erosion Resistance of Ceramic Materials for Refinery Applications", Bull. Amer. Ceram. Soc., 2. (?) 363-368 (1959). (2) Hansen, W. C., and Livovich, A. F., "Factors Influencing the Physical Properties of Refractory Concretes", Bull. Amer. Ceram. Soc., 34. (9) 298-304 (1955). 2 refractory materials were preheated, mixed, and cured at temperatures from 50F to 110F. They found reduced strengths after curing at 100F and 110F, and interpreted them to show that curing temperatures above 90F caused loss of strength. Unfortunately, their water: cement ratio, held constant up to 90F, was increased at 100F and again at 110F to retain constant slump. The consequent reduction of strength could have been caused by the increased water: cement ratio rather than by the increased curing temperatures. However, in field installations. Cook et al. (3) observed no detrimental effect when castable temperatures were allowed to rise due to retained heat of hydration of the cement. Most types of aluminous cements have high heats of hydration. It appeared reasonable, in view of Venable's work (1), that strengths would be reduced if the retained heat of hydration caused substantial drying of a castable. The evidence was not convincing that elevated temperature by itself was harmful, if drying could be prevented. This investigation had two purposes: to determine whether temperatures caused by heat of hydration were harmful when drying was prevented; and, if not, to develop simplified curing methods. EXPERIMENTAL The' temperature rise, due to the heat of hydration, was measured by placing thermocouples in a freshly-molded castable. Slab specimens were (3) a. Cook, M.D., Cook, C.P., and King, D.F., "Pneumatic Placement of Refractory Castables", Paper No. 9-R-62, Amer. Ceram. Soc. 64th Annual Mtg., New York, May 1, 1962. b. Cook, M.D., "Pneumatic Placement of Refractory Castables, Part 2", Paper No. 4-R-62F, Amer. Ceram. Soc. Refr. Div. Fall Mtg., Bedford Springs, Pa., October 5, 1962. 3 formed by tamping the castable in molds at a stiff plastic consistency. Two slabs were made, 12 in. by 12 in. by 3 in., representative in thickness of many refinery applications. The castable was a high-strength commercial product containing about 25 wt.7. aluminous cement and 75 wt.7. calcined flint clay. Its density after heating at 1000F was 128 to 130 lb/cu ft. The molds were insulated on five sides with 2 in. of polystyrene foam, leaving a 12 in. by 12 in. face exposed. One slab was cured in a damp room, at about 96% relative humidity and 80F. The other was cured in the open at 507. relative humidity and 75F, the exposed face being coated with a 1/16 in. layer of asphalt mastic. Another slab of the same material, 18 in. by 18 in. by 8 in., was insulated similarly and cured under a 1/16 in. layer of asphalt mastic. The mastic was a high-viscosity dispersion of asphalt and mineral pigments in a hydrocarbon solvent, sold as a vapor barrier coating. To obtain a yet larger temperature rise, an 18 in. cube was molded of a 60 lb/cu ft insulating castable. This castable contained about 50 wt.% pure calcium aluminate cement and 50 wt.7. perlite. The exceptionally high heat of hydration of the cement, together with the low thermal conductivity and low volumetric heat capacity of the castable, was expected to cause very high temperatures during curing. The mold was insulated on all sides with 2 in. of polystyrene foam. Thermocouples were inserted as the castable was placed at plastic consistency. Compressive strengths were determined on 2 in. cubes cut from the large cube. -4- To confirm the results of these tests, and to evaluate steam curing and other types of organic membranes, thirteen additional series of tests were made. Each series included three to eight castable slabs made on a single day. The castable used in each series was mixed and split to yield uniform samples. Nominally equivalent high-strength castables from four manufacturers were used. Day-to-day variations in ambient tem peratures, tap water temperatures, and humidity were unavoidable, and there was a practical limit on the number of slabs to be handled simul taneously. The series arrangement provided identical ambient conditions for all the specimens in a given series. Hater spray curing, simulating conventional practice, was done automatically in a cabinet equipped with spray heads, timers, and solenoid valves. Slabs, in their molds, were placed vertically in the cabinet within 30 min. after forming, and the door was closed to cause the re lative humidity to build up. Spraying began 6 hours after forming. A 30 sec. spray period followed at 30 min. intervals for 18 hours. Damp room curing, the usual standard laboratory methods, was done in a room at about 96% relative humidity and 80f. Specimens were placed in the damp room shortly after forming and left there for 24 hours. Steam curing was accomplished by building a hollow cube of 12 in. by 12 in. concrete panels. The specimens'"in their molds constituted one or two sides, with their exposed faces inward. Corners were sealed, and exhaust 5 steam at 5 to 15 psi was introduced, beginning 2 to 4 hours after forming and continuing for the remainder of the 24 hours after forming. The ex haust steam was ordinarily saturated, but up to about 10F of superheat occurred occasionally. Some specimens were cured in open laboratory air, at about 75F and 507. relative humidity, with no measurable air movement. For membrane curing, organic films were applied to the exposed surfaces of some specimens. Molded slabs were cured in this way in the laboratory for 1 day, then were removed from the molds for drying. The same asphalt mastic used for the earlier temperature tests was applied to a number of slabs by troweling, about 1/16 in. thick, 30 to 60 min. after the concrete was placed. Other membranes also were used. Molten paraffin was brushed on one slab. (Paraffin solutions proved impractical.) Four commercial con crete curing compounds, solutions of resins in hydrocarbon solvents, were applied by spraying. Two of the resin compounds contained red dyes, but no pigment. The other two contained white titania pigments. Two water emul sions of resins were applied by spraying. These were a styrene co-polymer and an alkyd resin emulsion, stabilized and pigmented in commercial coating products. Brief experiments, but no strength tests, were made with a water emulsion and a ketone dispersion of asphalt. 6 Several tests were made with membranes of a special asphalt cut back, containing appropriate additives and fast drying solvents, developed for this use in cooperation with a vendor. In most of the tests, only 12 in. by 12 in. by 3 in. slabs were formed, either by tamping a plastic mix or by use of a Cement Gun Co. Size 00 cement gun. Tamping mixes were placed at a uniform stiff ball-in-hand consistency. Gunning was done so as to obtain the temporary water sheen recently described as indicating the optimum water content (4). After curing, these slabs were cut into two 12 in. by 6 in. by 3 in. specimens, each of which was broken as a simple beam to determine modulus of rupture. To measure compressive strength, 2 in. cubes usually were cut from the broken halves and crushed. In a few cases 2 in. cubes were molded separately to determine compressive strength. In all tests, strengths were determined after 24 hours of curing, 1 to 3 days of air drying, and 24 hours at 220F. In some tests, duplicate slabs were heated to 1000F, after drying at 220F, before testing. RESULTS Effects of Temperature Table I shows the temperatures developed in 3 in. and 8 in. slabs of dense refractory concrete when the concrete was not cooled by spraying with water. By analogy with the 3 in. slabs, the internal temperature in (4) Wygant, J. F., "Effects of Variables in Pneumatic Gun Operation on Refractory Castables", Paper No. 3-R-62F, Amer. Ceramic Society Refr. Div. Fall Meeting, Bedford Springs, Pennsylvania, October 5, 1962. 7 the 8 in. slab probably was about 165F. The similarity between center and bottom face temperatures in the 3 in. slabs is evidence that heat loss frdm the insulated sides of the molds was insignificant. Each slab therefore was approximately the equivalent of a semi-infinite slab, twice as thick, with both surfaces exposed or with the back surface in contact with an uninsulated vessel shell. The temperature was slightly higher in the surfacesealed 3 in. slab than in the unsealed, probably because of the insulating effect of the asphalt mastic. In Table II, Series 2, 6, and 7, direct comparisons can be made between 3 in. slabs cured by water spraying and slabs that were not water sprayed. The strengths with damp room curing generally about the same as those with water spray curing. Asphalt mastic membranes gave result!, similar to water spraying in Series 2 and 6; in Series 7 asphalt membranes were superior. The small cooling effect of water spraying at 30 min. intervals did not lead to increased strengths. In fact, temperature:: in the membrane-cured slabs must have approached 120F, from retained heat of hydration, but did not adversely affect strengths. Under most plant conditions, gunned or trowelled castables would have to be about 6 ii . thick to reach this temperature during curing. Effects of even higher temperature are shown in Table III. The temperature of the large insulating castable cube reached the atmospheric boiling point of water, and 18 hours after molding the cube was still steaming 8 vigorously. Because of the high absorption of the perlite, the initial water content was high and only a small part of the water was driven off. In spite of the high curing temperature, strength was considerably higher than usual. Results with the 18 in. cube are confirmed by those with steam curing in Series 7, 8 and 9, Table II. The castable temperatures during steam curing must have approached 200F. In these series, strengths of steam-cured slabs ranged from about 207. to 807. higher than with water spray or damp room curing. The variability probably was due to varying vaportightness of molds. Under favorable conditions, strength increased with curing at steam temperatures. Elevated temperatures, even though achieved by external heating therefore appear harmless. They may even be helpful with respect to strength, if evaporation of significant amounts of water is prevented. In plant practice, insulating castable sections 18 in. thick (and probably thinner) can reach steam temperatures. The usual high-strength dense castables, in thicknesses of 8 to 16 in., would ordinarily reach temperatures not exceeding 165F. New Curing Methods Strengths of eleven series of slabs cured by various methods are summarized in Table II. Direct comparisons~&re possible within each series, but different series cannot be compared directly because of differences in 9 materials and placement conditions. However, either damp room or water spray curing can be used as an internal standard, against which other methods in a series can be compared. Damp room and water spray curing gave similar results in Series 2, 6, and 7. In these three series, the averages of damp room and spray curing results were used as standards for comparison. In each of the other series, only one of these two methods was included, and it was used for comparison. Table IV shows the percentile strength effects of other methods of curing compared with strengths from damp room and/or spray curing. Open air curing caused reductions of strength in most specimens. But steam curing caused increased strengths, except in Series 1, 2, and 3. The variable effects of steam curing in the first three curing cycles were traced to 5 to 10F superheat in the steam. The use of unsaturated steam results in less than 1007. relative humidity, and moisture evaporates rapidly from the wet refractory. Asphalt mastic and paraffin wax membranes, with two exceptions, yielded strengths superior to, or statistically indistinguishable from, those resulting from damp room curing. These membranes contain thin spots and defects, permeable to water vapor, unless they are applied very carefully. Nevertheless, these tests showed the potentially excellent performance of impermeable membranes. The ketone dispersion of asphalt (not shown in Table IV) showed no advantage over the asphalt mastic except the absence of low flash point solvents. Coatings of asphalt emulsion (not shown), and of alkyd and styrene 10 emulsions, are not sufficiently impermeable to water vapor. The high water vapor permeability of similarly formulated water-base paints is well known in paint technology. The resin solution curing compounds gave uniformly good results. This was also true with the asphalt cutback curing compound, as shown in Tables II and IV, Series 11, and in Table V. The average results of all membrane curing tests were superior to those of damp room and/or spray curing. The results with resin solutions and the asphalt cutback in Series 10 and 11 average 217,, higher than with spray curing. The tests in Tables II and IV were completed in 1955; those in Table V in 1956. Hundreds of additional test slabs, made and cured with membranes since then, confirm the results shown here. A considerable num ber of 2 in. cubes and 2 in by 4 in. cylinders also were made during the same period, and cured by the use of membranes under the semi-adiabatic con ditions recently described by Treffner and Williams (5). The temperatures reached were like those in Table I, and the strengths were normal. PRACTICAL APPLICATIONS Several petroleum refineries have been using membrane curing, with selected resin solution and asphalt cutback membranes, since early 1955. It has been eminently successful in curing refractory castable vessel linings and filler sections up to 12 in. thick. Hairline surface cracking, often observed (5) Treffner, W. S., and Williams, R. M., "Neat Evolution Tests with Calcium Aluminate Binders and Castables", Paper No. 5-R-62F, Amer. Ceram. Soc. Refr. Div. Fall Meeting, Bedford Springs, Pa., October 5, 1962. 11 after water spray curing, usually is absent. Cured surfaces are notably hard. Lining service has been good. Substantial savings have been made by the elimination of two or three shifts of water spraying labor. Earlier start-ups are permitted by removal of scaffolding as soon as the membrane is applied. No special precautions are necessary in the initial heating cycle, and the membranes are not deleterious to refractories, metals, or catalysts when burned off. The first requirement of a curing membrane is that it be nearly impermeable to water and water vapor. This can be judged according to ASTM Des. C309-58, "Liquid Membrane-Forming Compounds for Curing Concrete" and ASTM Des. C156-55T, "Water Retention Efficiency------- ", except that 24 hour water retention efficiency is a better index than 72 hour efficiency for the curing of aluminous cements. Curing membranes preferably are applied within 30 min. after installation of the refractory. White pigmented compounds usually are used onLdark castables made with iron-containing cements. Black asphaltic compounds are preferred when the castables contain white- low-iron cements. The contrasting membrane colors permit easy inspection for complete coverage. Hand sprayers of the garden tank type or conventional power paint sprayers are most convenient, but membranes can be applied with a soft full paint brush 12 Membrane curing compounds should be selected with caution. Some drying oils, surfactants, inorganic salts, and other compounds cause soft surfaces or interfere with the setting of aluminous cements. Curing membranes often must: be sufficiently viscous or thixotropic to form continuous films, preferably 0.010 to 0.015 in. thick, on vertical and overhead surfaces. The membrane materials must adhere to moist refractory surfaces. They should dry quickly--in 30 minutes, for example--to a tack-free condition, so that inadvertent contact with workers or tools will not cause damage. They should have good storage life. The solvents are volatile hydrocarbons, and the safety precautions usual with paints are necessary. Severely toxic or dangerously combustible solvents should be avoided. Membrane compounds should be reasonable in cost. Suitable commercial materials range in cost from about $0.80 per gallon for the special asphalt cutback product to $2.25 per gallon for some white-pigmented resin solutions, or $0.01 to $0.02/sq ft of treated surface at about 100 sq ft/ gallon coverage. This is insignificant compared to the labor cost of spray curing for three 8-hour shifts. The assured quality of curing and the early availability of the equipment for startup are additional advantages. CONCLUSIONS Moderately elevated curing temperatures, per se, do not cause reduced strengths in hydraulically-setting refractory castables. Either - retained hsat of hydration or external heating with saturated steam may be - 13 - somewhat beneficial to strength, if escape of moisture is prevented. Curing with steam is risky, because slight superheating of the steam can cause drying and loss of strength. The essential condition for good curing is maintenance of an adequate moisture content. In dense castables, placed with low water contents, it is necessary to prevent evaporation almost entirely. Curing by application of organic membranes, impermeable to water vapor, yields uniformly good results when the membrane curing compounds are carefully selected and applied. This practice has been proven by eight years of plant use. TABLE II EFFECTS OF CURING METHODS ON STRENGTHS _OF REFRACTORY CONCRETES Series Refractory Forming Method Curing Method Compressive: strength. psi, after heating at 220 F 1000 F Modulus of: Ruptur' psi.after hea ting 220 F 1000 F 1 A Plastic tamped Damp room 3320 Steam 3390 Asphalt Mastic 3630 670 550 530 2 A Plastic tamped Damp room 4070 3730 790 430 Water spray 4070 4000 530 410 Steam 3660 3000 670 270 Asphalt Mastic 4140 3290 710 430 3 A Dry gun (d) Damp room 8360 Steam 10120 Asphalt Mastic 7800 955 770 945 4 B Plastic tamped Damp room 2770 3060 Open air 2320 2540 Asphalt Mastic 3030 3270 Paraffin wax 3060 2680 7 5 B Plastic tamped Damp room 2830 Asphalt Mastic 3530 6B 7B 8C Dry gun (d) Damp room Open air Water spray Asphalt Mastic Plastic tamped Damp room Spray (1) (2) Steam (1) (2) Asphalt Mastic Plastic tamped Damp room Steam Asphalt Mastic 7630 7090 7170 7180 1960 1860 1800 3100 1920 2790 2920 4070 3730 1750 1850 1760 3260 2030 2470 610 540 670 580 430 490 400 610 360 480 660 920 610 210 150 100 270 120 270 9 D Plastic tamped Damp room 3890 Steam 4700 Asphalt Mastic 4200 800 920 630 TABLE III NEAR-ADIABATIC CURING OF INSULATING REFRACTORY CONCRETE Cube Size: Material: Forms: Consistency: Thermocouples: Temperatures: Compressive strengths, 220 F: 18 in. x 18 in. x 18 in. 60 lb/cu ft (100 F), perlite-IFB*-aluminous cement 3/4 in. wood with 2 in. foamed polystyrene insulation on all sides Spading Center and top center 210-230 F at center, reached 11-1/2 hr after pouring,190 F at center, 167 F at top center, 19 hr after pouring Blank: 2 in. cubes, damp room cured in brass molds: Operator A (3 cubes) - 380 psi Operator B (3 cubes) - 349 psi 2 in. cubes cut from 18 in. cube: adjacent to center (3 cubes) - 503 psi 505 psi 400 psi center of side (2 cubes) - 535 psi 523 psi *IFB = insulating fire brick TABLE IV PERCENTILE EFFECTS OF CURING METHODS ON STRENGTHS Series Curing Method Compressive Strength, psi, after heating at: 220 F 1000 F Modulus of Rupture, ; after heating at; 220 F 1000 F 1 Steam Asphalt Mastic +2 +9 -12 -21 2 Steam Asphalt Mastic -10 -22 + 2 -15 0 -36 +1 0 3 Steam Asphalt Mastic +20 -7 -19 -1 4 Open air Asphalt Mastic Paraffin Wax -16 -17 +9 +7 +10 -12 5 Asphalt Mastic +24 6 Open air Asphalt Mastic -4 -4 -15 9 7 Steam Asphalt Mastic +27 +49 +47 +39 0 +1 +16 +60 8 Steam Asphalt Mastic +36 +28 +39 -8 9 Steam Asphalt Mastic +21 +8 +15 -21 10 Open air Steam Resin Solution A Resin Solution B Styrene Emulsion Alkyd Emulsion -24 +79 +32 +40 +12 +7 +38 +45 -13 -22 11 Resin Solution C Light Coat Heavy Coat Resin Solution D Light Coat Heavy Coat Asphalt Cutback 0 -2 -1 +20 +35 -5 -5 -3 _ +13 +16 +10 +13 +7 +13 +19 +12 +20 +4 +29 +65 Average Effects Open Air Steam Membranes (a) -15 -17 +25 +13 +16 + 4 -2 +4 +6 -10 +27 (a) Excluding emulsions in Series 10. TABLE II (Cont'd.) Refrac- Series tory Forming Method Curing Method Compressive strength, psi, after heating at 220F 1000F 10 B Plastic tamped Water spray 2690 Open air 2040 Steam 4160 Resin Solution A(a) (1)vertical 3560 (2)horizontal 5590 Resin solution B(a) vertical 3750 Styrene Emulsion -- Alkyd Emulsion -- 11 f > B Dry gun (d) Water spray 5100 5300 Resin Solution C(a) (1)Light coat (b) 5080 (2)Heavy coat 5010 (c) Resin Solution 4980 5050 D(a) (1)Light coat (b) 5040 5160 (2)Heavy coat (c) 6120 5970 Asphalt Cutback (c) 6880 6160 Modulus of Rupture, psi, after heating at 220F 1000F 550 690 590 760 830 800 480 430 740 485 812 542 839 581 795 505 839 625 878 800 (a) Resin Solutions A and B were unpigmented commercial concrete curing compounds containing fugitive red dyes. C and D were commercial white-pigmented concrete curing compounds. (b) Approximately 300 sq.ft./gal. coverage. (c) Approximately 100 sq.ft./gal. coverage. (d) Each value is the average of two slabs, and of four compression cubes and two moduli of rupture per slab. TABLE V COMPARISON OF ASPHALT CUTBACK AND RESINOUR CURING MEMBRANES ON PNEUMATICALLY-APPLIED REFRACTORY CONCRETE (a)_______ Type of Cure Series 12 - (1 slab each) Resin Solution D Special Asphalt Cutback Series 12 - (4 slabs each)(b) Resin Solution D Special Asphalt Cutback Series 14 - (4 slabs)(b) Special Asphalt Cutback Compressive Strength, psi, after heating at 220 F 1000 F Modulus of Rupture, psi, after heating at 220 F 1000 F 5000 5210 4400 3890 937 837 682 607 6610 6450 1026 1098 7780 1183 (a) Applied at approx. 100 sq ft/gal coverage, or about 0.015 in. wet film thickness. Material was Refractory B of Table III. (b) Gunned with new design water ring in nozzle. r Reprinted from Brick & Clay Record Babcock & Wilcox Unique kiln design can save up to 75% fuel costs IFB construction saves time and money. Clay pipe fired two-high. Turn-over time up 60%. By: Lawrence j. midriff, jr., Production Mgr., Pomona Pipe Products Co. and cordon thomas, District Sales Representa tive, Refractories Div., The Babcock & Wilcox Co. Fig. 2. The sliding door has the same construction as the kiln. In the past year, Pomona Pipe Prod ucts put in what is believed to be the largest periodic kiln in the country to fire structural clay sewer pipe, and they are about to start con struction on a second insulating fire brick kiln of equal size. In addition to its size and insulat ing firebrick construction, the kiln is unusual because it has a top-fired premixed, fuel burner system, and a flue system with individual damper controls. Because of these systems, Pomona has realized substantial sav ings in construction costs and con struction time. Furthermore, the kiln has proven much more economical to operate than the thick-wall units, and has provided significant savings in ware recovery. The thin-wail kiln cost exactly half as much as a thick-wall kiln required to produce equal ware. The thin-wall unit costs a total of $40,000 for all material, labor, and equipment. At Pomona a conven tional thick-wall, heavy firebrick kiln, with equal ID, would have cost about the same. But the thin-wall unit has twice the capacity of the thick-wall unit because it can be set with ware two-high, whereas heavy duty firebrick kilns cannot, as will be explained later. Two of them, of course, would cost twice as much as one insulating firebrick kiln. Also, it would take one month longer to put up a single thick-wall kiln and four months longer to build two such kilns. From an operational point of view, this thin-wall kiln, when com pared with heavy firebrick units, saves more than 75% in fuel in some cases, and can be loaded in 60% less time. In addition to these econo mies, better control and efficiency of the thin-wall insulating firebrick kiln reduces loss of ware by 10 to 15%, and reduces each firing cycle by 10 hours. Basically, three main factors account for these advantages: 1. The insulating firebrick used in construction. 2. The burner system. 3. The flue control system. Pomona's large diameter pipe was being produced in 30 heavy fire brick kilns. After a thorough investi gation of the performance of other existing thin-wall kilns, there were enough facts to show the advantages of constructing a single thin-wall insu lating firebrick kiln. For example, the proven slower heat transmission and lower heat storage of insulating firebrick meant faster heat-up and faster cooling. It could also be shown that periodic kilns of larger size kilns produce* greater quantities of ware per dollar of capital invest ment. After careful consideration, Po mona gave the go-ahead for con struction of the 40-ft. ID insulating firebrick periodic kiln with walls 17 ft. high as shown in Figure 3. The walls were lined with headers of 9in. insulating firebrick recommended for service to 2000F, backed up by a 1-in. layer of 1900F block insula tion and a 14 -in. plate of carbon steel for structural support. Although '. /. Dickson "Kiln Designing advances using characteristics of Insulating Refrac tories' American Ceramic Society Bulle tin. 46 (12) 1178-1181 (1967). the cost of insulating firebrick is higher than that of firebrick, the total insulating firebrick cost was lower because fewer brick were re quired than for two thick-wall heavy firebrick units with the same ID. The thin-wall kiln used 50,000 insu lating firebrick compared with 150,000 high duty brick that two thick-wall units would have re quired. At $220 per M, insulating firebrick cost $ 11,000. High duty firebrick, at only $150 to $160 per M, would have cost between $22,500 and $24,000. The insulating firebrick kiln has a sliding door as shown in Figure 2 of the same thinwall construction as the kiln itself. This simplifies operation, compared with the old way of bricking up and taking down the door opening with each firing. The crown of the thin-wall kiln has a 2'/2-in. rise, one layer of insu lating firebrick is coated on its top side with refractory paint to reduce gas permeability through the porous insulating firebrick. This is covered with 2 in. of 2000F insulating castable, and a layer of coarse wo ven glass cloth, painted with an as bestos impregnated bitumastic com pound. In addition to savings in construc tion costs and time, insulating fire brick walls help reduce operating fuel costs because heat storage and heat loss are lower than that of dense fire brick as illustrated in Figure 4. The heat storage of the thin-wall kiln is less than 9% that of the conventional heavy wall construction. This is a major factor in reducing fuel con sumption with insulating firebrick. Also, a firebrick kiln loses addi tional heat through cracks in the lining, a constant problem with thick-wall construction. The thinwall kiln's steel jacket prevents gas flow either in or out and thereby prevents heat loss through cracks. Burner system precision set Pomona's second major advance in kiln construction in the clay pipe industry was the installation of the burner system. This system shown in Figure 6 fires the kiln through 16 ports which are located 14 ft. 6 in. above grade, instead of the conven tional ports 1 -ft. above grade. The bottom ports in the thin-wall kiln are bricked up, having been installed be fore it was known that top-firing would work. Pomona uses a wide-range burner premixing system to maintain a proper 10 to 1 air-gas ratio through out the burning cycle. This is a high-pressure premix burner system which Pomona modified so that the kiln lights off on four burners, then goes to eight burners, and finally to all 16. THERMAL CHARACTERISTICS Of THIN WAU VS. CONVENTKXAi. CONSTRUCTION STOCK INSUlATfON HEAT STORAGE ~ 6860 8TU PER SO FT. HEAT STORAGE 76,400 BTU PER SQ. FT. hii;. 4. This chart shows some of the advantages of the wall ItB constructions. e By top-firing, Pomona eliminated bag walls, which are necessary for bottom-firing kilns. Bag walls take up valuable setting space, add to construction costs, and cause many maintenance problems. Top-firing also eliminated the need for fire box arches and the maintenance prob lems that go with them. Figure 1 shows an interior view including burner locations and setting efficien cy. Combined with the kiln's unusual flue system, high pressure top-firing produces a complete change in the concept of heating periodic kilns. The pressure within the thin-wall kiln has proven to be sufficient for exhausting the kiln without either a stack or an exhaust fan. This elimi nates the need for the exhaust stacks required for most other periodic units. Such stacks are usually either 40 to 50 ft. units that create a draft to draw heat through the kiln, or 17 ft. units with induced-draft fans. Not sure that top-firing would create suf ficient positive pressure, Pomona had a 17 ft. exhaust stack construct ed and illustrated in Figure 5 when the kiln was built, but the induceddraft fan has never been installed. The new thin-wall kiln now being constructed does not have a stack, only a gas collection exit placed 6 ft. above grade to protect personnel. Flue system provides efficiencies The flue system in the thin-wall kiln is a key factor in the efficiency of the heating system. It is a system with several control points and. along with the kiln's burner system, size, and insulating firebrick con struction provide efficiencies not pre viously approached by any periodic kiln. The two main benefits of the flue system are: 1. It controls the pressure differential between the skew line at the top and the bottom of the kiln. 2. It exerts greater control over the amount and distribution of heat in the kiln. When the insulating firebrick kiln is at 2000F . the crown is 40 de- J A Fig. 5. The stack and service walkway are shown in this view from the back of the kiln. . . Quick turnover & high ( recovery are kiln features M grees cooler than the side walls 12ft. above grade. The crown of a thick-wall kiln, by comparison, is 400'F hotter than the walls 12-ft. above grade during its early burning cycle, and 75F to 100F hotter when the kiln reaches maturing tem perature of 2000F. The new kiln also shows a re markable vertical temperature uni formity in the ware measured by chromel-alumel thermocouples. At the end of the bum, the temperature is 2000F at the top of the ware and 1980F at the bottom or a difference of only 20. In a thickwall heavy firebrick kiln the differ ence between top and bottom is ap proximately 70=F with a setting only half as high. Controls in the flues permit close regulation of the amount and direc tion of heat going through the sys tem. Each damper can be completely pened, completely closed, or ad- isted to any intermediate point. With these controls and this flue system, the kiln is never significantly hotter at the skew line than at the bottom of the ware. Firing ware two-high is successful As a result of these tight controls over heating, Pomona can set sewer pipe two-high in the kiln and there by double the kiln's capacity. In creasing the capacity results in greater fuel economy. Since the con trols reduce dimensional variation, a greater percentage of saleable ware can be recovered. At one time it was common at Pomona to fire ware two-high in thick-wall heavy firebrick kilns. But as specifications for ware became tighter, this practice became ex tremely impractical. When the pipe was set two-high, it was extremely difficult to obtain correct shrinkage on the bottom pipe without, over burning the top pipe. This caused high losses of saleable ware. To in crease its recovery rate, Pomona turned to firing ware only one-high. But in the new thin-wall insulat ing firebrick kiln, because of the close control of heating, ware can be successfully fired two-high. Pipes from 12 to 18 in. ID cannot be set two-high because the dry strength of their thin-cross section is not sufficient to support the top load; but Pomona fires 5 ft. pipes with ID's from 21 to 36 in. two-high to obtain the increased efficiencies such setting makes possible. Firing two-high in the thin-wall, Pomona recovers 85 to 90% of saleable ware. This compares with a 60% recovery firing two-high in a thick-wall kiln of equal ID. The in sulating firebrick kiln, while firing two-high, reduces ware losses 10 to 15% below the heavy firebrick unit with only one-high. This is accom plished while contending with the inherent higher ware load pressures associated with higher settings. Loading the kiln for two-high set ting is much more economical than for one-high setting. Most loading time, in either case, is spent placing and leveling the setting rings. It takes very little more time to set the ware in two layers. Pomona cuts loading time 60% by setting pipe two-high. Firing cycle 6 'h days The firing cycle in this thin-wall insulating firebrick periodic kiln runs about six and a half days from light to light. First, the ware is set two-high. After light-off, kiln temperature is raised to 2000F during an 85 hour period using the top-fired burner system. During the heat-up period the flue system can control differ ences in temperature, from one side of the kiln to the other or from top to bottom. The dampers are ad justed to regulate heat flow and to equalize temperatures throughout the kiln. Fuel economies realized Perhaps the greatest economies of the thin-wall insulating firebrick kiln are in savings on fuel. Because fuel efficiency increases as tonnage in creases, setting ware two-high in creases kiln tonnage and, therefore, fuel efficiency. A larger kiln, with even greater tonnage, would yield greater efficiency. Other ways to ac celerate fuel efficiency as exempli fied by the Pomona kiln are reduc tion of heat loss through insulating firebrick walls, and elimination of heat loss caused by wall cracking. In controlled tests, Pomona deter mined the fuel consumption rates of an insulating firebrick kiln and a heavy firebrick kiln, each with a 40 ft. ID and firing to 2000F, as charted in Figure 7. The insulating firebrick kiln, loaded with 148 tons of 30 in. clay pipe set two-high, consumed 3730 cu. ft. of gas per ton of ware. The heavy firebrick kiln, loaded with 72 tons of 30 in. ware set one-high, used 12,000 cu. ft. of gas per ton of ware, more than three times as much fuel per ton. The best fuel efficiency was ob tained with 24 in. pipe because the greatest tonnage of that size pipe could be loaded into the kiln. With the 24 in. pipe set two-high, the insulating firebrick kiln held 174 tons of ware and consumed 3000 cu. ft. of gas per ton. In the heavy firebrick kiln, 84 tons of 24 in. pipe set one-high consumed 12,800 cu. ft. of gas per ton, more than four times the rate of the insulating firebrick kiln. Of course, the greater capacity of the insulating firebrick kiln is a sig nificant factor in achieving these fuel savings, but it is not the only factor. Even when firing equal tonnage, the overall efficiency of the insulating firebrick kiln results in lower fuel comsumption. Pomona fired its 12 in. and 18 in. pipe one-high in each kiln. Firing 70 tons of 12 in. pipe, the insulating firebrick kiln consumed 4300 cu. ft. of gas per ton of ware. The thickwall kiln used up almost twice that amount--8000 cu. ft. per ton. Firing 80 tons of 18 in. pipe re f quired 3750 cu. ft. of gas per ton in the insulating firebrick kiln and 9000 cu. ft., more than twice as much, in the thick-wall kiln. After reaching maturing tempera ture, the kiln is cooled to ambient temperature. Because temperature variation between the top and bot tom of the ware varies by no more than 20F, it is not necessary to have a soaking period at the end of the bum to equalize temperature, as is required for the heavier kiln. The 155 hours for the complete cycle is 10 hours less than with a thick-wall kiln. This time is gained in reaching the maturing temperature. Summary With the thin-wall kiln at Pomona the following results have been real ized: 1. Lower capital expenditure per ton of ware. 2. Lower gas consumption per ton of ware. ( 3. Lower labor cost per ton of ware/ Based on the success of the thin- wall insulating firebrick kiln during its first year of operation, Pomona has begun construction on a similar kiln with the same dimensions and burner and flue system. In fact, Po mona plans to construct thin-wall insulating firebrick kilns whenever any of its existing periodics require replacement. 1 30-INCH PIPE THIN-WALL KILN THICK-WALL KILN FUEL CONSUMPTION PER TON Of WARE 74-INCH PIPE THICK-WALL KILN 1 7-INCH PIPE THIN-WALL KILN thick-wail KILN 18-INCH PIPE THIN-WALL KILN THICK-WALL KILN 12,800 ONE HIGH SETTING TWO HIGH SETTING CU. FT. GAS PER TON 70 TONS 8,000 80 rTxONS Fig. 7. Typical fuel consumption figures charted here show the kiln's economy. K-1620 Service Temperatures, F Exposed Backup 1600 2000 Density Lb/9 in. str. Lb/cu ft 1.60 27.3 Fusion Point, F 2700 ^ Thermal Conductivity, / Btu/sq ft in. thickness, hr F 500 F 1000 F 1500 F 2000 F 2400 F Hot Load Strength % Deformation 10 psi, 11/2 hr at 1600 F 2000 F 2200 F 12.5 psi, IV2 hr at 2730 F 10 psi, IV2 hr at 2640 F Reheat Shrinkage at Temp, F 0.91 1.19 1.49 -- -- 0 0 -- -- -- 1550 0 Cold Crushing Strength, psi 90 Modulus of Rupture, psi 90 Chemical Analyses % Alumina, Al203 Silica, Si02 Iron Oxide, Fe203 Titanium Oxide, Ti02 Calcium Oxide, CaO Magnesium Oxide, MgO Alkalies, as Na20 38.7 44.1 0.5 1.5 14.9 0.1 0.2 1 Coefficient of Linear Reversible ' Thermal Expansion, in./in./F 3.0x10-* K-20 K-23 2000 2000 2300 2300 1.70 29.1 2750 1.85 31.6 2750 0.97 1.25 1.55 -- -- 1.07 1.32 1.59 1.91 -- -- 0 -- -- -- 1950 0 no 110 -- -- 0.1 -- -- , 2250 0 145 150 39.6 43.0 0.7 1.1 15.0 0.1 0.3 40.1 43.3 0.6 1.1 14.4 0.1 0.4 3.0x10* 3.0x10* K-26 K-28 2600 2600 2800 2800 2.60 44.5 3100 2.70 46.2 3190 1.70 2.10 2.71 3.43 -- 1.82 2.16 2.64 3.20 -- -- -- 0.3 -- -- 2550 0.2 195 200 -- -- 0.3 -- -- 2750 0.6 190 200 40.1 54.6'^ 2.4 1.2 1.5 0.1 0.4 46.1 50.8 1.0 1.4 0.3 0.1 0.3 2.9x10* 2.9x10* K-30 2900 2900 3.06 52.3 3190 2.04 2.36 2.96 3.80 -- -- -- 0.2 -- -- 2800 0.5 295 280 45.3 51.4 0.9 1.4 0.5 0.1 0.4 2.9x10* K-3000 Insulpure Insalcor 3000 3000 2800 2800 3250 3250 3.25 55.5 3350 2.9 49.5 3475 4.60 78.7 3350 2.13 2.37 2.83 3.44 4.00 2.03 2.07 2.26 2.56 2.86 6.03 5.87 6.30 7.30 8.40 -- -- 0.5 -- 0.5 2950 0.6 275 230 -- -- -- -- 0.5 2750 0.3 220 200 -- -- -- 0.1 -- 3250 +0.4 900 350 64.8 33.2 0.6 0.7 0.3 Trace 0.4 94.5 0.3 -- -- 5.2 -- -- 76.6 21.5 0.4 0.6 0.1 0.1 0.3 2.9x10* 5.2x10* 3.8x10* DaU are average results of control tests and are subject to normal variation. Results should not be taken as maximum or minimum requirements for specifications. For more information contact your nearest BiW sales office or The Babcock & Wilcox Company, Refractories Division, Augusta, Ge^-^ia Alhambra, Cal. 91803............... 813 So. Fremont Ave.. .283-6625 Atlanta. Ga. 30303 ........................1315 Candler Bldg... 522-5387 Augusta, Ga. 30903..............................1288 Merry St.. .738-4574 Buffalo, N.Y. 14225..........................132 Cayuga Road. .633-2800 Chicago, III. 60603..........................29 So. LaSalle St...236-6546 Clarksville, Ind. 47130................... 813 Eastern Bled.. .282-6691 Cleveland, Ohio 44113.........1367 Illuminating Bldg.. .781-4381 Dallas (Richardson), Tex. 75080.. 777 S.Central Expwy. .231-7233 Detroit (SthfId), Mich. 48075.. 510 Northland Twrs. E,.. 353-6440 Houston. Texas 77018 ........................P. 0. Box 94110. .861-9161 New York, N.Y. 10017................... 161 East 42nd St.. .687-6700 Philadelphia, Pa. 19102. .3 Penn Cntr. Plz., Rm. 1126. .564-6376 Pittsburgh, Pa. 15237................. 7805 McKnight Road. .931-4490 Portland, Oregon 97201.........1600 Fourth Ave. S.W.. .228-0410 San Francisco, Cal. 94111............. One California St.. .4 14 Seattle, Wash. 98104........................305 Norton Bldg.. .622~t496 Canada: Standard Refractories Ltd. A subsidiary of The Babcock & Wilcox Company Burlington, Ontario and Montreal, Quebec Litho in U.S.A. R-934-4002 5.5M-9-71