Search Documents

By John A. Ames

Webster's dictionary nearly equates portland cement with its current primary definition of cement. While such equation may be a triumph of common usage, the confusion between the terms cement and concrete unfortunately is very widespread. Full definition of cement properly would extend to many substances, but for purposes of this chapter cement is restricted to "varieties" of portland cement including "portland-pozzolan" cement. Among the usually recited historical notes concerning cement is that it was developed by the Romans. Their use of cement in the great structures of Rome, and even in the far corners of their Empire, such as Hadrian's Wall in the north of England, testifies to the antiquity of a major cement industry. From English sources we know of John Smeaton's carefully proportioned hydraulic cement mixed in 1756 and of Joseph Aspdin's patented cement to which, in 1824, he gave the name portland cement. (He claimed that his cement mortar looked like the famous dimension limestone quarried from the Isle of Portland on England's south coast.) The year 1971 marked the centennial of portland cement manufacture in the United States. A fair historical account of the many real contributors to the cement industry between 1756 and 1972 unfortunately would be too extensive for this chapter. Development of rotary kiln practice and burning at higher temperatures clearly was the major move into today's technology. As part of the 4th edition of Industrial Minerals and Rocks, this chapter on utilization focuses on inferred interests of the anticipated readers. The main period for which statistics are cited generally is that since the 3rd edition (1960) of this same volume. Cement manufacture is the processing of selected and prepared mineral raw materials to produce the synthetic mineral mixture (clinker) that can be ground to a powder having the specified chemical composition and physical properties of cement. Cement making incorporates a number of distinct steps, but it is characterized by the key pyroprocessing (or "burning") step, which brings about the necessary changes in the raw materials. This pyro processing is a chemical process. Conversion of large volumes of mineral raw materials into specification-controlled cement at relatively low cost is the objective. Statistics on the cement industry are of uneven quality. The U.S. Bureau of Mines (USBM) is the principal collector of available data. These statistics can be only as good as the quality of information supplied by the reporting companies. The data perhaps most often questioned relate to productive capacity of the industry. The questions submitted by USBM are valid, nevertheless, and the information gained from them, as good as is available, is what is generally used. The Portland Cement Association also is a source of much quality information. The year 1972 witnessed a test of the adaptability of cement industry statistics. This was the year of conversion from barrels and sacks of cement as the units of measure to tons and hundredweights. The barrels and sacks related to volumetric measures in that a sack was judged to be 1 cu ft (weighing a "standard" 94 lb). A barrel was 4 sacks. Volumes of concrete are stated in cubic yards, and the number of cubic feet of cement in a cubic yard of concrete was a sensible approach to the mensuration involved. However, cements do not weigh the same amount from plant to plant or from type to type, and the buyer was more interested in getting an assured weight for his money than in "buying air." Thus, a sack, in accepted practice, became 94 lb instead of 1 cu ft, and a "barrel" became 376 lb, regard- less of whether it would fit into 4 cu ft of space. With the de facto dealing in weight rather than in volumes, the awkward numbers made little

Jan 1, 1975

By S. A. Swensrud

NINETEEN Thirty-Six was the biggest year in volume in the history of the oil industry, and unquestionably the best since 1929 in respect to profits. The quota of new and difficult problems to face seemed a little smaller during 1936 than in most recent years, and a somewhat better balance existed between the different branches of the industry. 'Domestic consumption of gasoline in 1936, estimated at about 475 million barrels-the highest on record, was approximately 11 per cent above 1935, and 26 per cent above 1929. Demand for crude oil increased in about the same proportion: over 1935, and the average production, of about three million barrels daily,' constituted a new high level.

Jan 1, 1937

By William Mikulas Lars Thomassen, Clair Upthegrove

Little work has been done in the field of the nickel-tin binary system. The complete diagram has been investigated on two occasions, but the results are in very poor agreement. The structure of a compound NiSn and the solubility of tin in nickel have been determined by X-ray methods. Several investigations have been made of the liquidus curve, and the compositions of the eutectics have been determined. A study of the literature, however, shows that there is no satisfactory agreement between the data presented in the various papers. The present investigation was undertaken in an attempt to determine the equilibrium conditions, and to correlate X-ray with metallographic data. The equilibrium diagram was determined by thermal and metallographic means, and the solubility of tin in nickel further investigated with X-rays. Powder diagrams were made of the intermetallic compounds, and three of the structures were partly determined. NO attempt was made to determine the range of homogeneity of the intermetallic compounds. Review of the Literature Vigorouxl prepared alloys of tin and nickel containing 74 and 92 per cent of tin. By treating these alloys with very dilute nitric acid or with sodium hydroxide he was able to separate residues, both of which analyzed 66.7 per cent tin. This corresponds to the compound NiSn and was reported as such. Guillet2 made the first investigation of the complete binary diagram Fig 1) The maximum point of the liquidus is at 43.5 per cent nickel. Several solid solution areas were found, and the compound NiSn was included on the strength of Vigoroux's work. The investigation was

Jan 1, 1937

By R. O. Williams

It is shown that Ni3Al precipitates homogeneously from nickel-rich alwminum alloys as plates on the (100) planes. Prior to actual precipitation a process occurs which is believed to be one of increasing short-range order. After precipitating the Ni3Al plates enlarge through competitive growth. Discontinuous precipitation can occur simultaneously with the above processes. Recent ideas of the origin of precipitation strengthening appear adequate to explain the hardness changes. REMARKABLY little appears to be known about the precipitation process in Ni-Al alloys in spite of their technical importance. This investigation originated to supply additional information about precipitation in general, this system in particular. Information on the structures and kinetics have been obtained through the use of hardness, X-rays, microscopy, calorimetry, and resistivity on high-purity alloys. PROCEDURES Six alloys, Table I, were prepared by melting carbonyl nickel and high-purity aluminum in alumina crucibles in vacuum and casting into 1-in. graphite molds. All rods were homogenized at least once at 1300°C for 24 hr prior to swaging and this was repeated on the first three alloys after 75 pct reduction. Alloy 4 could be reduced only 10 pct at 1000°C (probably in two-phase field) prior to fracture but 1/4-in. samples quenched from 1100°C were readily reduced cold. Alloy 5 was reduced 15 pct cold but failed on the next pass while alloy 6 of essentially the same aluminum content failed inter-granularly without apparent flow up to 1000°C. The alloys were heated in hydrogen at the elevated temperatures and formed thin, coherent aluminum oxide coatings which provided excellent oxidation resistance at lower temperatures. However, freshly prepared surfaces showed considerably less resistance at 500"to 700°C in air and apparently resulted in internal oxidation. As a consequence, low-temperature agings were carried out in evacuated tubes. RESULTS The isothermal hardening behavior of these alloys at 500"and 565C is given in Figs. 1 and 2. These results were obtained from samples cold worked 75 pct, recrystallized at 1000°C (1100°C for the 7.8 pct Al) and quenched in water. This recrystallization was used to give smaller grain sizes so as to obtain more uniform hardness values and the points represent an average of five readings. The electrical resistivity was measured on 1/16-in. wires quenched from 1000°C during aging at 495°C to give Fig. 3. The energy release and its rate are given in Fig. 4 for the 6.9 pct Al alloy during aging around 500°C. Inasmuch as this was a single run, its accuracy is not known but certainly the general shape and magnitudes are correct. The method used to obtain these results is described elsewhere.' Data for the aging at 600°, 700°, and 800°C of these alloys cold worked 50 pct are given in Fig. 5. Supplementary information from microscopy and X-ray diffraction have been included to indicate recrystallization, discontinuous precipitation and the appearance of superlattice lines from the Ni3Al. The hardness of these alloys as annealed, aged, cold worked, and cold worked and aged is given vs composition in Fig. 6. Those samples which were isothermally aged, Figs. 1 and 2, were reaged at 532°C and at successively higher temperatures for the indicated times to give the data of Fig. 7. These results as well as certain others, support the idea that the level of hardness reached for temperatures above 600°C are equilibrium values more or less independent of path. This being the case, the breaks in the curves would be the complete solution of the Ni,Al. The electrical resistivity versus temperatures for some of these alloys, both aged and unaged, is given in Fig. 8 along with those data from heating slowly (10 deg per day) to high temperatures. Interesting points include the lowering of the Curie temperature (the change in slope), the lack of any indications of a solubility limit and the large temperature coefficient for the Ni3Al. A slight break for Ni3Al around 100C shows up but this is not a Curie temperature as Ni3Al is not ferromagnetic down to -190°C. Metallographically both the nickel-rich solid solution and the Ni3Al appear very much like pure nickel. Profuse twin boundaries are present both

Jan 1, 1960

By T. E. Watkins, M. D. Nelson

A method of determining the lime content of drilling muds proposed by Battle and Chaney* has been examined both in the Field Research Laboratories of Magnolia Petroleum Co. and in field drilling operations. This proposed analytical method is condensed as follows: Titrate one ml of mud and one ml of mud filtrate, each with 0.02 N H2SO4, to the end point with phenolphthalein. The volume of acid in ml required for the mud titration is recorded as Pm and that of the mud filtrate titration as Pf Calculation of the lb lime/bbl mud is accomplished by use of the formula, lb lime/bbl mud = 0.6 (Pm - 1.3Pf) - 0.6 The analytical procedure outlined by Battle and Chaney was found to be quite satisfactory when reasonable precautions are used in mud sampling technique. Calculation of lime concentration by the method proposed in the above publication, however, has given values that varied as much as 100 per cent from the known lime content, for muds of either high or low solids concentration. By use of a formula, lb lime/bbl mud = 0.26 (Pm-F1.Pf), based on chemical equivalents of acid required for titration, both laboratory and field data have been satisfactorily correlated with the known amounts of lime present in the muds. In this formula a correction, F1, or volume fraction of liquid in the mud, is made for the solid content of the mud. Pm and Pf * Paper on "Lime Base Muds," presented by J. L. Battle (Humble Oil and Refining Co., Houston, Tex., and P. E. Chaney (Sun Oil Co., Beaumont Tex.), at the 1949 Spring Meeting of the Southwestern District API Division of Production, Dallas, Tex. The analytical procedure has also had wide distribution in a trade journal, Kembreak, published by the Accuracy Rig and Tool Co., Houston, Tex, †Derivation of equation lb/bbl — grams/350 ml 1 ml 0,02 N acid = 0.000741 grams of lime 1 ml 0.02 N acid/ml lime suspension = 0.26 lb of lime bbl lb of lime/bbl = 0.26 (Pm -F1, Pf ) have the same significance as in the Battle and Chaney equation. Table I shows typical data employing both formulas in calculating the lime content of various lime muds. This suggested procedure ignores lime which is in solution in the mud. The solubility of lime in water, normally about 0.5 lb per bbl, is greatly decreased by the addition of caustic soda, and, in a lime mud, it is probable that the concentration of dissolved lime is of the order of 0.1 lb per bbl. Such an amount of lime can be ignored, since it is within the range of accuracy of the method.

Jan 1, 1950

By E. P. Pfleider, W. G. Pariseau

This paper reports the theoretical and experimental results of an analysis of ore pass drawdown as a problem in soil plasticity. The method of analysis developed appears to be a promising technique of investigating the mechanics of fragmented materials. A reconstructed plasticity theory brings experiment and theory into reasonable agreement, and hence may provide the foundation needed for rational ore pass design. The gravity-induced movement of broken rock through ore passes is an economical method of transporting fragmented materials through large vertical distances in underground mines, provided the ore pass is well designed. Unfortunately, the design procedures now followed do not very often result in ore passes that function properly. On the one hand, the huge accumulation of empirical data pertaining to the movement of materials in gravel bunkers, storage bins, deep grain silos and other structures that have features in common with ore passes remains uncorrelated. On the other hand, there is a dearth of experimental data substantiating theoretical models of material behavior. Consequently, reliable generalizations encompassing the broad spectrum of problems involved have not been forthcoming. AS an initial step towards improving the situation, a detailed case study of two-dimensional ore pass drawdown was made. The mechanics of the process were analyzed theoretically as a problem in soil plasticity and tested experimentally through observations of a laboratory ore pass model. THEORY Taylor1 defines a soil as a "mass commonly considered to consist of a network or skeleton of solid particles, enclosing voids of varying size." Broken rock in an ore pass is such a material. In plasticity theory this material is replaced by an idealized substance that deforms elastically up to some state of stress at which slip or yield occurs,2 For com-pressive states of stress, the Mohr-Coulomb criterion is a satisfactory yield condition. After the yield point is reached, the ideal soil deforms in accordance with the concept of plastic potential.4 Under plane strain conditions the governing equations separate into a system of stress equations and a system of velocity equations.' Only stress variables appear in the stress system. The velocity system, however, contains a stress variable. The two systems are, therefore, coupled. Both systems are hyperbolic, and each has two real, distinct families of characteristic curves that coincide according to the classical theory of ideal soil plasticity. The

Jan 1, 1969

(Under this heading will be published notes sent to the Secretary of, the Institute by members or other persons introduced by members.) Development and production. Member open for engagement. Past year in war mineral field work, operation and production. Ten years' thorough experience in both subordinate and executive positions, gold, silver and copper in the United States, Mexico, and South America. Age 37, married. Minimum salary $3600. Interview. South or East. A-577. Engineer and executive. Captain U. S. Army, age 36, expects discharge before July 1. Seven years' mining and metallurgical experience. Six years with large technical college as assistant professor of mining and directing extensive experimental work. Desires position as executive or research engineer with reliable company or as professor of mining with first-class college. A-909. Mining engineer. Twenty-five years' experience including mine and works operation, construction, plant installation, and mine examination. Speaks Spanish. A-1173. Metallurgical engineer. Twelve years' laboratory and operating experience, covering copper blast furnace, roasting, reverberatory smelting and latest Cottrell treater practice on copper dust and zinc dust. Technical graduate. 36 years of age, married. Highest recommendations. Open for immediate engagement. A-2354. Mining engineer. Junior Associate. At liberty about July 1. Captain Heavy Artillery. Single, age 24. Go anywhere. Best refer-ences. A-2843. Engineer. Age 26, single, graduate Missouri School of Mines, B. S. in Mis. Met. E. Desires position with oil or metal company in any capacity. Location immaterial. At present engaged, desires change. A-2844.

Jan 7, 1919

By LEE H. CORNELL

Drilling activity in Kansas during 1945 continued at an accelerated pace to meet the demands of war; 1810 tests were drilled as compared with 1856 tests the previous year, 1771 tests in 1943, and 1516 tests in 1942. During the year 31 new oil pools and 10 new gas pools were discovered. Of a total of 364 wildcats drilled, 323 were dry; and of 1446 tests drilled in pools, 825 were oil wells, 197 gas wells, and 425 dry holes.

Jan 1, 1946

By C. David Mann

Metal prices continued to improve in 1964, resulting in the opening of new mines and re- activation of old ones. Larger and deeper shafts are being bored. At the AEC's Nevada Test Site, a 72-in. diam shaft is being bored to 3200 ft. In New Mexico a 1650-ft shaft was bored with a 48-in. bit, then reamed to 124 in. to 935 ft, and 102 in. to 1650 ft. Deviation was only 26 in. In California a shaft sinking machine has been developed which combines drilling and mucking equipment into one mobile, integral unit.

Jan 2, 1965

By Charles H. Hopkins, J. M. Lohr

A COMMERCIAL development requiring a suitable alloy or alloys for sealing into various grades of glass made it desirable to have a more exact knowledge' of the expansion characteristics of the nickel-iron alloys than appeared in the published literature. The expansion of the more common [ ] types of glass under consideration varies from 2.7 to 10.4 X 10-s per degree Centigrade, over the range 20° to 400° C. For reference, the thermal expansion curves of several types of glass used for sealing by the lamp and tube industries are given (Figs. 1 and 2). This paper covers a detailed investigation of the thermal expansion of a series of nickel-iron alloys in which the nickel content varies from 30 to 70 per cent. Particular stress is laid on the alloys that contain between 40 and 55 per cent Ni.

Jan 1, 1938

By Arthur Forsyth

A SEARCH through the available literature shows that the cobalt-nickel-silicon system has not been systematically studied. This seems rather odd because all three elements are fairly abundant and have better than average corrosion resistance; however, they have high melting points, which makes experimental work difficult. Cobalt and nickel form a continuous series of solid solutions and each metal dissolves about 7.5 per cent silicon. Many ternary alloys have been studied and the results have been published in a great many different ways, but to present such data in a short but clear and understandable manner is still a problem. The following pages describe the results of the study of 76 alloys from which the partial ternary diagram (Fig. 1) of the cobalt-nickel-silicon system has been drawn. The alloys were chosen according to the plan in Fig. la by taking definite ratios of cobalt and nickel, such as 70 Ni, 30 Co or 40 Ni, 60 Co, and adding silicon in various percentages. Each section was studied in detail by means of thermal analysis and microscopic examination, but only two of these detailed sections were thought necessary in the point perspective drawing to give a comprehensive picture of the liquidus, solidus and solid transformations that take place over the temperature ranges studied. Complete thermal analyses and microscopic data are given in Table 1.

Jan 1, 1940

By W. P. Frey

The briquet plant of the Lehigh Coal and Navigation Co., at Lansford, Pa., has previously been referred to.' It has passed the stage of experiment and now rests on a foundation practically and financially sound. The purpose of this' paper is to explain how success has been attained, and to set standards for the development of hard-coal briquetting which may eliminate future failures. All the conclusions are based on actual results, avoiding speculation and theory. Up to three years ago, the L. C. & N. Co. used coal-tar pitch as a binder in making hard-coal briquets. The plant operated with a monthly output of 300 to 600 tons and was unable to meet its own expenses. Naturally, no one was satisfied, especially as the ideals of a smokeless briquet were still unrealized. Then came the Dutch oil process, introduced in this country by the General Briquetting Co. of New York. This process uses as binder a small percentage of heavy residue of petroleum, which, being liquid, permits great economy of handling; it also did away with the most objectionable fumes of the former briquets. This oil2 is not absolutely smokeless; heating the fresh fuel to red heat liberates some low-volatile hydrocarbons, yielding a white smoke which is not objectionable. On the other hand, this oil adds to the briquet about 7 per cent. of volatile matter, giving the product a heating capacity equal to that of the best free-burning anthracite coal. The average silt used in the manufacture of these boulets averages 16.5 per cent. chemical ash, corresponding to 12,000 B.t.u. per pound of coal. The Hydrolene used as binder runs about 17,500 B.t.u. per pound, and approximately 7 per cent. of it enters into the mixture, hence 12,000 X 0.93..........................11,160 B.t.n. 17,500 X 0.07.......................... 1,125 B.t.u. Total.....................................12,385 B.t.u. per pound of briquets

Jan 1, 1918

Jan 1, 1970

By Charles H. Hopkins, J. M. Lohr

A commercial development requiring a suitable alloy or alloys for sealing into various grades of glass made it desirable to have a more exact knowledge of the expansion characteristics of the nickel-iron alloys than appeared in the published literature. The expansion of the more common types of glass under consideration varies from 2.7 to 10.4 X 10-6 per degree Centigrade, over the range 20" to 400' C. For reference, the t,hermal expansion curves of several types of glass used for sealing by the lamp and tube industries are given (Figs. 1 and 2). This paper covers a detailed investigation of the thermal expansion of a series of nickel-iron alloys in which the nickel content varies from 30 to 70 per cent. Particular stress is laid on the alloys that contain between 40 and 55 per cent Ni.

Jan 1, 1939

By Charles H. Hopkins, J. M. Lohr

A commercial development requiring a suitable alloy or alloys for sealing into various grades of glass made it desirable to have a more exact knowledge of the expansion characteristics of the nickel-iron alloys than appeared in the published literature. The expansion of the more common types of glass under consideration varies from 2.7 to 10.4 X 10-6 per degree Centigrade, over the range 20" to 400' C. For reference, the t,hermal expansion curves of several types of glass used for sealing by the lamp and tube industries are given (Figs. 1 and 2). This paper covers a detailed investigation of the thermal expansion of a series of nickel-iron alloys in which the nickel content varies from 30 to 70 per cent. Particular stress is laid on the alloys that contain between 40 and 55 per cent Ni.

Jan 1, 1939

United Engineering Trustees, Inc. GEORGE D. BARRON ARTHUR S. DWIGHT R. M. ROOSEVELT The Engineering Foundation H. C. BELLINGER E. DEGOLYER J. V. N. DORR

Jan 1, 1932

DISCUSSION, G. Derge and D. J. Girardi presiding N. A. Gokcen (Michigan College of Mining and Technology, Houghton, Mich.)—While the authors present very interesting results on the effect of sulphur upon the solubility of oxygen, they have not established convincingly the lines of two-fold saturation (the valleys in Fig. 2) and the ternary eutectic point. Their contention that sample S-7, upon melting at

Jan 1, 1954

By R. E. Ogilvie, J. I. Goldstein

The a and y solubility limits in the Fe-Ni phase diagram have been redetermined at temperatures above 500°C. Both a diffusion-couple and a quench and anneal technique were used. The solubility limits were measured with an electron-probe micro-analyzer. The nickel concentration at the phase boundary is increased below 700°C and the a solid-solubility range is much larger than had been previously measured. The solubility limits are also extrapolated to 300°C. It is suggested that the solubility of nickel in a FeNi reaches a maximum in the temperature interval 400o to 500°C. THE equilibrium diagram has been of great use in the field of meteoritics, where the phase relations between kamacite (a) and taenite (?) in metallic meteorites can be described by means of the Fe-Ni diagram. The study of metallic meteorites by electron-probe microanalysis has cast some doubt on the accuracy of the presently available Fe-Ni diagrams.1, 2 Recent thermodynamic studies of the Fe-Ni system also suggest that the diagram may be in error.3 For these reasons the high-temperature (800o to 500°C) part of the diagram was redetermined. INTRODUCTION The currently accepted Fe-Ni diagram is that of Owen and Liu.4 Above 910°C, there is a region of complete solid solubility, ? (fee). Below 910°C, the a (bee) phase is stable in pure iron. The effect of increasing amounts of nickel is to stabilize the y phase. The phases that form when Fe-Ni alloys are heated or cooled bear little relation to the equilibrium diagram. If an alloy is cooled from the y state and held at a temperature within the a + y field, no evidence has been found for the occurrence of the y - a transformation.5 If the alloy is cooled to low enough temperatures the y phase breaks down into a supersaturated bee phase called a2. In fact, in alloys over 27 pet Ni the y phase is retained at room temperature.6 The a, phase has the same composition as the original y. The temperature at which a2 forms, the M, temperature, has been determined experimentally.' A state of equilibrium can be approached by cooling below M, to form a2 and then reheating the alloy into the two-phase region of the diagram. The y phase will then begin to precipitate out of the a2 phase and grow. The growth of the phase, however, is quite slow. Using the interdiffusion coefficients -Da of Goldstein et al.,8 it is estimated that it takes about 1 year to grow a 10-µ-wide region of ? at 700°C in a 5 pet Ni alloy. Owen and Liu4 used the technique just described to form the equilibrium phases. The phases present were determined by means of X-ray analysis. The accuracy of their diagram depends on the number of alloys available near the phase boundary at a given temperature. In this study two different techniques were used to determine the a/a + ? and ?/a + ? solubility limits. The inherent accuracy of both techniques was greatly improved over that used by previous investigators because the phase-boundary compositions were measured with an electron-probe microanalyzer. PROCEDURE The two methods used to determine the Fe-Ni diagram are the diffusion couple (D.C.) and the quench-and-anneal (Q.+A.) techniques. In the first method, diffusion couples whose diffusion path goes through a two-phase region of the phase diagram were used. A description of the technique used for making the diffusion couples has been described in a previous paper.' After the diffusion treatment a discontinuity in the resultant concentration vs distance profile was measured. The nickel concentrations in the a and ? phases at the interface of the discontinuity are the solubility limits of the a and ? phases in the phase diagram at the diffusion temperature. If the interface com-

Jan 1, 1965

By C. T. Holland

This discussion is concerned with those com-J- paratively infrequent bumps that eject material from the failed mass with enough energy to wreck heavy machinery and seriously injure or kill people. In such cases there is a loud report, and the pressure waves set up in the mine atmosphere wreck ventilation controls several hundred feet from the site of the failure. Vibrations are induced in the earth's crust which may be detected several miles from the point of origin, and explosive gases may be released in large quantities." Usually a dense suspension of dust is produced from the failed mass, together with dust placed in suspension by the resulting air blast. These rock bursts or bumps have also been called mountain shots, bounces, pillar bursts, and crumps. Such occurrences, when extensive, are serious catastrophes, and may force abandonment of the mine.' CAUSES OF ROCK BURSTS Rock bursts in coal mines have occurred in sev-eral places under diverse geological conditions, mining methods, and/or practices. A study of many such occurrences indicates that they depend on the following conditions: 1) cover thickness of 500 ft or more, 2) cover composed of strong members and having a strong member lying close to or on the coal bed, and 3) a floor that does not readily heave. The minimum thickness of cover under which coal mine rock bursts occur is based on observation. Under unusual circumstances involving very strong roof members and/or mining systems, and/or practices especially likely to induce bursts, a burst conceivably could occur under less roof. Under special circumstances rock bursts have occurred in marble quarries almost at the surface.' Furthermore, under very heavy cover it is likely that controlling roof members need not be as strong and as close to the coal bed to cause bumps. Just what constitutes a strong roof or roof member is open to question. Obviously the strength of a roof member is determined by its thickness, the distance between bedding planes across which little or no bond exists, the spacing and development of joints, and the inherent strength of the rock material. Strata sections showing roofs under which rock bursts in coal mines have developed are illustrated in Fig. 2. In several instances the strength of rocks composing the roof of mines developing bumps has been tested. The tests (Table I) show that these rocks develop considerable strength in compression, bending, and shear and that in some cases the strengths are very high. No quantitative data are available concerning joints in roof members over coal beds where rock bursts have occurred, but the author has examined many mines subject to bursts and has received the impression that joints in the shales immediately over the coal beds are always present and well developed. In the main sandstone or conglomerate members themselves, however, where observed, he has yet to see a fracture system indicating well developed joints. Undoubtedly joints are present, but the evidence indicates that they are few, poorly developed, and widely spaced. In a mine subject to bumps, moreover, actual measurements of roof spans up to 225 ft between supports under a cover of about 800 ft have been made before roof failure occurred. Such spans signify strong roof. In coal beds subject to rock bursts, the bottom rock is usually a sandy shale. A few tests (Table I) have been made on the floor rock of coal beds. Observations in mines indicate that they are usually weaker than roof rocks and that under load they show more plastic effects. Observation in ground subject to bursts invariably indicates a relatively strong floor highly resistant to heaving, although some heaving may occur. Other factors listed as favorable to rock bursts are: 1) mountainous surface; 2) steeply dipping beds; 3) proximity of faults, folds, and other similar geologic structures. Many bursts have occurred in mines under mountainous country, although many bursts in mines under a surface with little relief indicate that thickness of cover, not relief, is the essential factor. In steeply dipping beds, many bursts occur in mines that are essentially flat, and rock bursts have been recorded in beds dipping as much as 35o Most rock bursts in the U. S. occur in beds that dip less than 10 pct. It is also true that most U. S. coal is produced from beds dipping less than 10 pct. Geologically, faults are regarded as providing stress relief, but in the vicinity of faults, as well as sharp folds, it would seem that residual stresses could be present. The fact that rock bursts occur in areas remote from faults and sharp folds suggests that their presence is not necessary for this phenomenon to take place. There have been many bursts close to faults and folds, and it may be assumed that occasionally such structures accentuate their occurrence3 The conditions and circumstances under which they occur, coupled with the effects produced, leave In the Crowsnest Pass coalfield, British Columbia, a rock burst is estimated to have released 1.8 million cu ft of methane.

Jan 1, 1959