You are on page 1of 11
anon ALLIS-CHALMERS Metal Wear In Crushing and Grinding FRED C. BOND Senior Engineer ALLIS-CHALMERS MANUFACTURING COMPANY 54th ANNUAL MEETING OF AMERICAN INSTITUTE OF CHEMICAL ENGINEERS Houston, Texas December, 3, 1963 Not released for publication except by permission of Alli-Chalmers, Technical Press Bureau o7P 1701 METAL WEAR IN CRUSHING AND GRINDING “The two principal direct operating costs in commercial ‘crashing and grinding are power and metal wear. Power costs ean now be analyzed and predicted with reason able aceuracy, but metal wear costs ore still difficult to estimate with exactitude, This paper descrites a nnethod of testing different ores and other materials for their relative abrasive wear, and gives equations which relate the laboratory tested abrasive wear index to the actual metal wear in commercial machines. ‘An expression for commercial metal wear which can be applied equally to all machines and all crushing oF grinding operations ig necessary, The common term of pounds of metal wear per ton ground is not satisfactory, since it does not allow for differences in product and feed sizes, or for differences in grindability. ‘The term uused here is pounds of metal wear per kilowatt-howr, of Ib/kwh, Tt permits a ditect comparison between differ- cent machines, different feed and product sizes, and dif ferent materials. It can be obtained from the wear in pounds per ton and the energy input required in kilo- watt hours per ton. The metal weat figures given in this paper include the metal worn away by abtasion ot removed by dissolu- tion and the scrap loss, such as the weight of crush- ing of mill liner plates discarded and the grinding rads and alls removed from the mill end discarded in rormal ‘operations. The percentage of scrap loss must be known before the net metal weight lost by abrasive wear and dissolution alone can be found from these figures, Very approximate values for the average serap loss are 60% of the totel metal wear for ceusher liners, 35% for rod angi ball mill liners, and 20% for crushing roll shells. ‘The setatch hardness of quartz and that of most was ores is higher than that of the average motal while that of limestone is silic used in contact with them, usually lower. For this reason the metal wear in grind- ing Limestone and dolomite is much less thon that in the grinding of ores containing large amounts of silica. TThe metal wear loss in wet grinding is several times that in dry grinding and in crashing, This is caused by actual dissolution of iron from the highly active nas- cent fresh metal surfaces produced in grinding. Mage nets in the mill discharge will recover only a fraction of the iron consumed in wet grinding. Most of the iron hhas been dissolved as iron hydroxide, and has intro- daced hydrogen into the pulp from decomposition of the ‘water. Metal consumption is increased further by grind= ing in en acid pulp, especially when the pi becomes lower then about 5.5, Some measurements have indicated that the potential ‘of the pulp in wet grinding may be positive to the mill shell by as much os half a yolt, and is probably also positive to the geinding rods and balls. This potential difference should increase the metal loss by con ‘The subject of pulp and mill potentials in ge has been much neglected. When used dry grinding balls are examined under the microscope they are seen to be covered by pits and seratches which avorage from 3 to Smictons in width, ‘and about half that in depth. The smoother surface of wet grinding balls indicates the importance of corro- sion caused by dissolution in their greater wear rates. ABRASION TESTER ‘After several other devices had been tried in the Allis: Chalmers Research Laboratory and rejected for various reasons, the abrasion tester shown in Fig. 1 and d scribed below was adopted in 1956, It has been in continuous servico since that time. Its design is adapted from an article published in 1955 by the Pennsylvania ‘Crusher division of the Bath Iron Works.) It measures, the weight loss of a steel paddie continuously impact- ing falling ore portictes for one hour under standard conditions. A flat paddle 3'x1"x1/4", of SAE 4325 chrome- nickel-molybdenum steel hardened to 500 Brinell, is inserted for one inch into a rotor 4.Sinches indiameter, which rotates on a horizontal shaft at 632 through fall ing ore particles. Two square inches of paddle surface are exposed to abrasion, and the paddie tip, with a radius of 4.25 in., has a linear speed of 1410 feet per minute; sufficient for a good impact blow. ‘The rotor isenclosed by @ concentiic drum 12 inches in diameter and 4.5 inches deep, which rotates at 70 rpm, of 90% of critical speed, in the same direction ‘a5 the paddle. The inner circumference of the drum is wrod een Fig. 1 ~ Abrasion tester fined with perforated steel plate to furnish a rough sur- face for continuously elevating the ore particles and showering them through the path of the rotating paddle. In operation, screened particles passing 3/4 inch square and retained on 1/2 inch square ate used as feed, (It should be noted that in the published article referted to, the feed is 3/4 inches x 3/8 inches). Four hundred grams of 3/4” x 1/2" foed are placed in the drum, the end cover is attached, and abrasion is con- tinued for 15 minutes; then the drum is emptied, aa other 400 grams are added, and the abrasion continued. In each complete test four 400 gram samples are each abraded for 15 minutes, ‘Thus the paddle is abraded for a total of one hour, after which it is weighed to the tenth of @ milligram. The loss of weight in grams is the abrasion index Ai of the material. Since the exposed outer corners of the paddle are visibly worn after @ single one-hour test, the paddle ig reversed for a second test and is then discarded. ‘The abraded material is united, screen analyzed, and the microns P which 80% passes is determined. ‘The equivalent feed size F of the 3/4””x1/2” feed is 42,200 microns.\?) Of 31 samples with an average work index Wi of 14.25, the average product size P was 80% passing 13,250 microns, with en average abrasion index Ai of 0.2440 gram. Substituting in the work index equation :) ve J0WE . 10WE “ Ve VF 14 142.5 Vi3,280 42,200 = 0.544 kwh per short ton. ‘The transposed work index equation wie w 70 30 © (WP VF or Wi = 0544/C10*. gear @ ‘AB yl ‘can be used to calculate a very rough average open circuit impact crushing value of the work index from the 80% passing size P of the abrasion test product. Since 1600 grams, or 0.001763 ton, are impacted in tone hour, the useful power-output of the abrasion tester is 0.960 watts. An abrasion index of 0.2440 gram, or 0.000539 pound, gives the weight of metal abraded away by 0.960 watt-hour, so that one kwh abrades 0.561 pound of paddle metal. In any test the abrasion index multiplied by 2.30 gives the pounas of paddle metal abraded per kwhr. ‘The tester can elso be used to evaluate the abra- sion resistance of different metals, Paddles 3”7x1”” 1/4” can be made of any desired metal and given any desired heat treatment. The abrasion loss can be com- pared with that of the standard paddle, using a standard rock material for testing. However, this article deals with the comparison of the abrasiveness of rock ma- terial only. ‘The Table headed ABRASION TESTS lists the abra- sion index values of 171 different materials. The 80% passing size P in microns of the combined sbraded Produets wos not determined forthe earlier tests. The work index values listed as RN- were obtained from Standard A.C rod mill gindability tests at the mesh shown, and those listed az BM- were obtained ftom A-C ball mill grindability tests ot the mesh shown’) Those not designated were obtained from impact crushing fests, ant those in parentheses were calculated from the 80% passing size by equation (3). CORRELATION WITH PLANT EXPERIENCE Plant operating data from which metal wear in pounds per Kilowatt hour could be obtained were collected from all possible sources, both published and unpub- Hished. References are given following this paper to the published sources used. The work of deriving relationships between the metal consumption in Ib/lewh ‘and the abrasion index Ai was laborious, and many anomolies were found which required discrimination. ‘The most important cause of discordant data was the excessive amount of scrap metal discarded in some plants, probably resulting from breakage. This was followed by inexact reporting of the energy input re- quited in koh ton, particularly in crushing, The use of metals especially resistant to abrasive wear caused some discrepancies, and Ni-Hard was excluded from the averages. Finally, no ebrasion index tests had been made on many of the ores whose metal consump- tion in Ib/iwh was available, end it was necessary to treat these as average members of a class. Tt was found that the wear reported for extremely non- abrasive materials tended to be higher than that cal~ culated, and an atbitrary lower limit of 0.020 was, estab- lished for the abrasion index used in calculations, ‘The equations given below were derived from careful consideration of the available plant performance data WET ROD MILLS Rods tb/towh Liners WET BALL MILLS ~ OVERFLOW AND GRATE DISCHARGE Balls Ib/kwh = 0.35 (Ai-0.015)'/5 35 (Ai ~ 0.0202 += 035(4i - 0.015)3°. . o © = 0.026 (Ai- 0.0152 ----- Liners DRY BALL MILLS ~ GRATE DISCHARGE (Ai < 0.22) Bells Ib/iwh = 0.05 Var ®) Liners = 0.005 VAT ® CRUSHERS - GYRATORY, JAW AND CONE TYPES Liners Ib/kwh ~ Ai + 0.22 wees (0) it * ROLL CRUSHERS 7 Roll Shells tb/kwh » (ALY ay 10 In Fig, 2 the metal consumption in ib/kwh is plotted ‘against the abrasion index Ai, and the curve numbers, are tabulated below in the order of decreasing wear Curve Wear Equation Mech No. (1) Wet Rod Mills “Rods @ Q) Wer Boll Mills Balls (6) (3) Roll Crushers. Shells. = (1) (4) Crushers. Liners: (10) (5) Dry Ball Mills: Bolls (8) (6) Wet Rod Mills Liners (5) ” Wet Bali Mills Liners a (8) Dry Ball Mills Liners (9 Table I was compiled from the table ABRASION ‘TESTS at the end of the paper. Different types of ma teri listed in the order of increasing abrasion index, and the successive columes include: 9 cam Le/om Fig. 2 ~ Abrasion Index plotted opeinst me Wowh. ‘The number of fests averaged The average specific gravity Sg ‘The average work index Wi when available. The average mictons P 80% of the product passes. ‘The average abrasion index Ai ‘The conelation between the abrasion inex and the work index is very poor. However, when the work index is higher than the average of 13 the abrasion index is higher than 0.20, with the exception of cement elinker, ‘The abrasion index is actually a parometer of the of fective scratch hardness of a material, measured pon metal of 500 Brinell hardness, while the work index is TABLE | ~ ABRASION AVERAGES |_No. | Moterial Ave. | So | wi _| G) | Dolemite 5 | 27 ~ (2) | Shole 5 | 26 | 99 @) } LS. for Cement wo | 27 | 17 (@) | Limestone 9 | 27 | 7 (5) | Coment Clinker 8 | 315 | 135 (| Magnesite 3 | 30 = @) | Heavy Sulfides 10 | 356 | wd (8) | Copper Ore wm | 295 | 7 9) | Hematite 7 | 417 | as (10) | Mognetite 2 | 37 | 130 (1) | Grovel 4 | 268 | 15.4 (12) | Trop Rock 2 | 20 | 178 (13) | Gronite | 272 | 166 (14) | Taconite 7 | 337 | 163 (15) | Quartzite 3 | 27 | wa (16) | Alumina 7 | 39 | ws | | | | TABLE I] - WEAR AVERAGES ~ POUNDS PER KWH Wor Red Mills Wet Boll Mills Dry | Dy | Dry No. Reds | Lining | Bolls_| Lining | Balls | Crushers | Rolls aw 110) 0075 060 0053 | .0050 | .0220 | .0160 Q a 0076 | 061 soosa | 0051] 0221 | 0161 @ 1130, | ‘0084 | c074 002 | 10070 | ‘0224 | -o180 (a) 1150 0100 | .090 0074 | 0088 | .0230 | .0215 (S)Dry| 0175) | (00175) | (.0135) | (.00135) | .0135 | .0265 | .0380 6) “196 | .0152 138 | 0112 | -o140 | 10270 | .o400 ” 225 010 | .169 034 | .o1e0 | .0318 | .0550 (a) 1232 0190 178 ‘0140 | 0190 | °0333 | -060 0) "298 | .0198 | 186 | 0147 | .0205 | .0350 | .066 (10) 1255 219 | ‘207 | 0161 | .0235 | .o410 | .080 Mm) 270 0238 228 | 0176 ~ | .046 | 094 (12) (83 | 10256 | 247 | 0190 - | 056 | -109 (13) 1286 0262 | .264 | 0194 = | [066 | cn 4) 1297 0302 | (316 0224 = | 1076 | ‘158 (15) 320 0322 330 | 0240 = | 090 | 381 (16) 1336 | 0336 340 0248 = [100 | -198 a porameter of the resistance to breakage. These are distinct, although related, physical properties, and a close correlation between them should not be expected. In Table II the different materials ate listed in the same order of increasing abrasion index used in Table I, with the metal consumption averages in pounds per kilowatt hour, as found from equations No. (4) to (11). ‘The first four columns list the metal wear for wet grind= ing tod and ball mills, except for cement clinker which is always ground dry, Because of lack of data the ball wear in dry mills was not continued beyond an abrasion index of 0.2217, ‘and equation (8) should not be used tentatively for more abrasive materials. DISCUSSION It should be remembered in this discussion that metal ‘wear is considered in terms of pounds perkilowatt hour, ‘and not in terms of pounds per ton ground. When the plant metal wear in Ib/kwh is knowa the abrasion index Ai can becalculated by transposing the equations given, and when Ai can be checked by the plant wear in two or more stages of crushing or grinding it can be used with considerable confidence. However, abrasion index tests are advisable, both for predicting, metal wear in new installations end for checking that in existing plants. ‘The metal wear data represent the averages from operating plonts, excluding Ni-Hard, Ni-Hard will r duce metal weat considerably in wet grinding, and much more in dry grinding, Publication of comparative abra- sion tests on wear resisting alloys, and more plant data on their use, is anticipated charge in plat Wear data now available on autogenous mill linings are not considered adequate for deriving @ dependable equation In wet rod mills the rod and shell liner wear per unit of surface area exposed are approximately the same, while in wet ball mills the ball wear per unit of sur face area exposed is less than half that of the shell liner wear, Contributing couses of this difference are the lighter weight of the balls, their higher Brinell hardness, and the greater slippage of the ball charge against the liner, Fig. 4=Jaw crusher waors mate! fomewing jaw end stotionary jaw plates; ofse check plates. The data available did not show the difference in metal wear per kilowatt hour between grate discharge bali mills and overflow bell mills which have been ex- pected from the weardifference in pounds per ton ground, ‘The metal wear in wat grinding is 8 to 10 times that in dry grinding. ‘The metal wear in wot rod mills is ftom 15% to 40% higher than that in wet ball mills. In wet ball mills the bell wear is about 13 times that of the mill Liner wear, and in dry ball mills it is 10 times. Ja wet rod mills the rod wear is 11 to 12 times that of the Liner wear. The metal wear in ctushing rolls is about twice that in gytatory and jaw crushers, and that in crushers is, about 75% higher then that of dry grinding mills. In Table I the abrasion index Ai multiplied by 2.30 gives the amount paddle metal abraded in Ib/kwh. These values ate larger than all the dry operations in Table ti, and areexceeded only ivy the wet rods where Ai < 0.085, and wet balls where Al< 0.045, ‘The highvetocity impact, of the abrasion tester wears the hard paddle metal rapidly. Much the same action con be expected in im- pact crushing of abrosive materials. The abrasion index Ai represents direct impact ab sion, and the difference between direct impact and other methods of size reduction is shown by the fact that in equations (4) to (L1) the exponent of the Ai term is less than unity in all cases except that of crushing, where the Ai term is divided by 11. In other ‘words, an increase in abrasive hardness has much more effect upon the metal wear in impact rushing than in F4g, 5 — Gyratory crushing abrades metal from mantle ond other methods of size reduction. conceve liners Within the normal operating range of rod and ball mills the metal wear in [b/kwh is roughly proportional to the mill speed. The cost of metal consumed in cents pet pound divided by the cost of the energy in cents per kilowatt hour gives @ metal cost factor. When this metal cost factor is multiplied by the metal consumption in.tb/kwh the tatio of metal cost to energy cost is obteined for ‘each machine; and the calculated energy cost per ton ‘can then supply the total energy plus metal cost per ton, Fig, 7 ~ In bell mill the balls, shell fr, alsa the discharge grote if pe SUMMARY Correlation has been established between laboratory abrasion tests and metal wear in commercial crushing, and grinding machines. Equations have been derived by Which laboratory abrasion tests can be used to predict, crushing and grinding metal consumption in planning commercial installations, and to check the efficiency of metal use in operating plants. REFERENCES 1 10, u 2. 13, 14 15, 16. 47, . Norman Weiss Benj, J. Burbank: “Measuring the Rotative Abra siveness of Rocks, Minerals and Ores,"" Pit and Quarry, August 1955. Fred C. Bond: ‘New Ideas Clarify Grinding Prine ciples,”” Chemical Engineering, February 5, 1962, E.R, Johnson: Crushing Practice at Braden Cop- per. AINE Tech. Pub. 2180 (1947). “Rod Milling-Plant Tech. Pub. 2175, Myers, Michaelson and Bond ‘and Laboratory Data,” ADI July 1947. B.H. Cody: “Milling Practices at Concentrator of Moronci Reduction Works,"” AIME Tech. Pub, 2194, Joly 1947 7, E. Norman: “West Tests on Grinding Balls,” AIME Tech. Pub, 2319, 1948. Myers and Lewis: “Progress Report in Grinding at Tennessee Copper Co.", AIME Trans. Vol. 187, Mining Engineering, June and November 1950, B.S. Crocker: “Screened Ore Used for Fine Grind ing at Lake Shore Mines,”” AIME ‘Trans., Mining Bnginoering, November 1952. “Silver Bell Concentrator,”” AIM San Francisco meeting, Sept. 1984. Lewis and Goodman: “Grinding Practice at the Tennessee Copper Company's Isabella Mill,’ AINE, New Orleans Meeting, Feb. 1957. C.B. Kettering and K.L, Power: “The Mettalurgi- ccal Story at Inspiration,” AIME, New York meets ing, Feb, 1988, “A Study of Wesner, Pobereskin and Campbell Grinding Ball Wear Employing Radioactive Tracers, AIME New York meeting, Feb. 1958. C.E. Schmidt and F.M, Hiwell: ‘‘Single-Stage vs. ‘wo-Stage Grinding at Homestake,"" AIME, New York meeting, Feb, 1958. “A Study of Different Types nax,”” AIME, New Windoph and Duggan of Steel for Grinding Media at C! York meeting, Feb. 1958. Wesner, Pobereskin, and Campbell: “Study of Grindiag Ball Wear Employing @ Radioactive Tracer Technique,” AIME Trans., Vol. 217, 1960. M. J. Sather: “Concentrator Operations at the Bunker Hill Company,”” AIME, St. Louis meeting, Feb, 1961. J. L. Green: “Economic Comparison in Grinding Phosphate Rock,” AIME, Dallas meeting, Feb. 1963, 18, 19. 20. a. 22. 2. 24 2s, 26 21. 29, 30. 3h 32, 33, 34 36. J. ¥. Moody: ‘The Rodbep Mill," AIME, Dalles, mmecting, Feb. 1963. Komadina and Hemlund: "Pima Mining Company Concenteator, “Rocky Mountain Mineral Conference, Denver, October 1957 Milling Methods at the Concentrator of the San Manuel Copper Corp., AIME, Arizona Section, May 17, 1958. Chapter V Milling ~ International Nickel Corp., Canadion Mining Journal, May, 1946, B. G. MeDermid: “Rod Milling at Snow Lake,"* Canadian Mining & Met, Bulletin for Sept. 1953, page $35, Hollinger Mill Staff, “Hollinger Crushing Plan! Canadian Mining & Met. Bulletin for Sept. 1953, page $50. B. S, Crocker: ‘Pine Grinding with Screened Ore at Lakeshore Mines,” Canadian Mining & Met Bulletin, March 1954, “Technical Advences During Canadian L. E, Djingheuzian 1955 — Milling and Process Metallurgy. Mining Journal, Feb. 1956, Symposium ~ Grinding Practice, Canadian Institute of Mining & Met,, Bulletin, August 1957. Keys ond Eichholz: “Measurements of the Wear Rate of Cost Grinding Balls Using Radioactive Teacers."" Dept. of Mines, Ottawa, TBI8. F.C. Bond: “New Grinding Theory Aids Equip- ment Section ”” Chemical Engineering, October 1952. ©. ©. Van Zandt: “New Grinding Installation at Lone Star's Nezareth Plant.” American Mining Congress, Cleveland meeting, May 1957. F.C. Bond: “Action in @ Rod Mill."” Engineering and Mining Journal, March 1960, Wilhelm Anseka: “Technique of Grinding," Compania General de Asfaltos y Portland Asland, Pub. by Ortega, Atibau 7, Barcelona, Spain, R. D, Haworth: “The Abrasion Resistance of Metals, No. 42", American Society for Metals, Philadelphia meeting, October 1948, James Lampman: ‘Selection and Performance of Abrasion Resistant Materials in Mining and Quarry- ing Industry,”” Pit and Quarry, March 1963, Noranda Staff: “Notes on Ball and Rod Mill Liner Practice in the Noranda Organization, CIMM Toronto meeting, April 1950. ‘Taggart: Handbook of Mineral Dressing, Sections 4&5, john Wiley and Sons, N.Y, (1945). soxason ret sme tek

You might also like