Search Documents

By D. M. Larsen

Since start-up in 1985 the Cannon Mine has produced and shipped a 170 g Au/tonne (5 oz Au/ton) flotation concentrate to various smelters for gold and silver recovery. Several processed have been investigated to economically produce dore' at the mine site from the refractory concentrate. This paper describes recent laboratory and pilot scale test work utilized in the process design and engineering feasibility study of a 73 tonne/day (80 ton/day) pressure oxidation and CIL facility. The designed pressure oxidation circuit utilizes a series of vertical autoclave units operating at 170°C (338"F) and 1.4 MPa (200 psig). Projected operating costs are discussed.

Jan 1, 1993

By Dwight R. Coles

The solving of complex metallurgical flowsheet calculations by hand is, at best, a tedious job which often proves that engineers, too, make mistakes. The high cost and long lead time associated with manual flowsheet solutions have led several organizations involved in research, development and construction of extractive metallurgical processes to search for computerized systems which will solve flowsheet calculations. Although design flowsheet programs have been developed by the chemical industry, the ability to solve flowsheets involving particulate solids is usually overlooked. A new computer applications system has been developed at Fluor Mining and Metals which has been designed to accommodate the type of design flowsheets required by the minerals industry. By employing the sequential modular computation technique, a library of metallurgical pro- cessing routines may' be accumulated which are used by a universal executive system. This technique enables rapid solution of new process) .flow- sheets with only a minimum of additional programming effort. The implementation of the FLEXMET flowsheet solution system is the subject of this paper. Some of the major features of the system which are examined include interactive input data handling, automated flowsheet calculation order arrangement, flexible accomodation of unit operations models and advanced management of recycle and specification control calculations.

Jan 1, 1982

By R. K. Davey

Before the introduction of computerized procedures, mineralization modeling (compilation of level maps), ultimate pit studies, ore reserve compilations, and long-range mine planning all were performed manually. In fact, much of the basis for today's computer programs was derived from manual procedures and actual experience of mining engineers and geologists.

Jan 1, 1979

By S. K. Kawatra, A. K. Bakshi

Mineral suspensions in water show Newtonian flow properties at low percent solids. However, this changes to a non-Newtonian regime as the pet-cent solids increases. For Newtonian fluids, viscosity is independent of shear rate, hut, for non-Newtonian fluids, viscosity varies with shear rate. Therefore, if the slurry is non-Newtonian, attempting to measure the viscosity at a shear rate different from the shear rate experienced by the fluid within the process is futile. This is especially true for a hydrocyclone, where the rate of shear is different at different points within the hydrocyclone. In this case, it is very important to determine the flow type of the suspension before attempting to establish any correlation between the viscosity and the process efficiency. A new technique to distinguish such changes inflow type was developed. This involves two viscometers: a coaxial cylinder-type viscometer that measures apparent viscosity at low shear rates and a vibrating sphere-type viscometer that measures the apparent viscosity at high shear rates. For each of these instruments, special precautions were taken to allow, for the measurement of the viscosity of rapidly settling mineral suspensions. Because the vibrating sphere viscometer is operated at a much higher shear rate than the rotating viscometer, the two instruments together could determine the shear-rate dependency of the viscosity. Ground silica (80% passing 65 µm) in water was used to prepare slurries at concentrations ranging from 10% to 70% (by weight) solids. The viscosity of each slurry sample was measured simultaneously by both viscometers, and the results were compared. With this technique, it was found that all of the silica-slurry samples tested were in the Newtonian flow regime for up to 70% (by weight) solids. Methocel and Cellosize polymer solutions showed clear non-Newtonian behavior using this technique. This confirmed that non-Newtonian behavior of the silica slurries, if present, would have been detected by this arrangement.

Jan 1, 1997

By Wisecarver D. W.

Today the longwall is getting increasing attention in the United States as an excellent way to mine coal. This is not to say it is the best way in every mine because conditions, capital costs, and markets may well dictate another method. However, operators with working longwalls are often showing preference by phasing out room-and-pillar sections and replacing them with longwall. In the United States there are about 90 longwalls in operating coal mines and the number is increasing. In Europe, 95% of underground coal mining is by longwall compared with only 5% in the United States. In the Western States about 25% of the underground coal mines use longwall mining. There is logic behind the increase in longwalling. Compared to room-and-pillar mining, straight single-pass longwall has important advantages that show up in underground coal mining in any part of the country.

Jan 1, 1978

By D. K. Dysinger

Capital resource budgeting is the means by which managers of publicly held companies fulfill their fiduciary responsibility to maximize the wealth of stockholders. Managers maximize shareholder wealth by allocating investment capital to the most productive of competing alternative projects in a manner that maximizes the net present value of the firm's future cash flow. The techniques of cash flow analysis are the primary tools available for selecting a corporate portfolio of projects that meets or exceeds the rate of return expected by shareholders. The reliability of cash flow analysis hinges on the ability of managers to forecast the future values of individual components. The accuracy and reliability of forecasting determine the soundness of capital investment decisions and, ultimately, determine the value of the firm. The present paper will explore tools for forecasting the spot market price of a commodity, silver, to demonstrate the difficulty and uncertainty involved in forecasting the future. The commodity price is only one of a number of components of the cash flow equation. Each component has a unique and uncertain future and must be individually analyzed and forecast to arrive at a robust cash flow model.

Jan 1, 1997

By E. J. Jacques

A typical problem in ventilation networks with multiple fan systems is the possibility of them interacting. Multiple fan systems are no exceptions in complex tunnels, and may also be met in mines where underground fans are installed. When aiming at controlling a multiple fan system in a ventilation network, e.g. with a view to maintaining an acceptable air quality level, yet at minimum power consumption, transient flow conditions frequently occur. Although in most applications these transients are short compared to the time needed to achieve the planned actions on the ventilation, one should, however, keep in mind that it is important to avoid any bouncing effect between different (adjacent or distant) fan systems. Therefore a simulation of the unsteady-state flows could allow us to learn more about the propagation of transients within the network, and their possible secondary effects Among others, one question that often is raised, is to find out in which sequence and at what time schedule each of the fans in a network should e.g. be started up? Another typical problem, into which an unsteady-state simulation could provide some insight is the following: how should the facilities of variable pitch angle and/or changeable RPM of several fans operating in parallel be handled and combined in order to obtain a flow rate that changes gradually from zero to its maximum value i.e. the sum of the individual throughput flows of all the fans. The literature is not very abundant on the topic of simulating transient flows in networks. There is no doubt that numerical models have been developed (Pollak, Christensen, 1987), most of the time for particular applications (Wala, Kim, 1985), (Tanaskovic et al., 1984), (Schultz, Sockel, 1991), etc. The aim of this paper is to present a model that also simulates one-dimensional compressible flows in unsteady-state conditions with, however, the specific intent to handle networks with more than one fan. PROBLEM STATEMENT The mathematical model chosen to describe the unsteady flows in a ventilation network is based on the dynamic and continuity equations applied to each of its airways. The following basic assumptions are made : - the flow in the airway is considered to be one- dimensional, - - the flow is assumed to be fully turbulent, and hence the ATKINSON equation for calculating the friction losses will be applied. Whether this equation remains valid for transient state flows remains an open question. However, this may reasonably be assumed since the interesting flow frequencies are low. A discussion may be found in (Kämpf, 1987) ; - the airways are assumed to have a constant cross- sectional area along their length. If obviously this would not be the case, it is always possible to subdivide the concerned airway into sections for which the previous assumption holds. Applying the basic equations of mass and energy conservation, the flow in each branch may be described by the Euler equation for Newtonian flows.

Jan 1, 1995

By V. K. Karra

Research by various workers has shown compressive comminution to be significantly more efficient than impact comminution. Based on the work being done at Nordberg, wet compressive crushing in cone crushers producing finer and finer products, shows the feasibility for partial substitution of conventional tumbling mills. For further substitution, major challenges are processing of fine feed slurries in cone crushers to realize a significant size reduction of the particles, and understanding the limitation of the gyratory type units in terms of finest product achievable. The paper addresses the challenges involved, efforts at Nordberg, and the implications to conventional comminution flowsheets.

Jan 1, 1989

By S. K. Kawatra, H. L. Dilley, W. W. Wen, T. C. Eisele

To improve the sulfur-rejection capability of a coal-cleaning plant, it is first necessary to evaluate the performance of the existing circuit. This information can then be used to decide whether there would be significant benefits from improving the operation and control of the existing circuit, or whether further improvements would require more radical process changes. This paper describes a simple model for evaluating heavy-media separation circuits, and presents a case-study of such an analysis of an operating heavy-media coal cleaning plant. Efficiency curves were calculated and modelled using Lynch's method, and a simplified expression was derived for calculating the Ecart Probability value from the model parameters.

Jan 1, 1995

By Jue Kou, Fazhao Sun, Jiang Liu, Cuili Gui, Baoqiang Yu, Chunbao Sun

A novel special magnetic separator was developed based on the special properties of precious metal-bearing alloy in highgrade nickel matte. The recovery of Au, Pt, and Pd was 77.79%, 84.79%, and 81.79% respectively at optimum operation conditions of magnetic separator feed of 2.5t/h, pulp density of 45%, and flush water of 1.5t/h, and the recovery significantly improved comparing with normal magnetic separator. Magnetic field simulation of the special magnetic separator indicates that magnetic intensity of 0.6 T can be generated on the surface of magnetic drum, which can strengthen the recovery of heavy alloy that contains precious metals. The alternate strong and weak magnetic field can also make magnetic particles rolling on the surface of drum, which will reduce the generation of magnetic agglomeration to a large extent and decrease the content of impurities in magnetic concentrate.

Jun 16, 2022

By Robert G. Moorhead

Fine-coal screening can be one of the most critical operations in a modern coal-cleaning plant. The yield and quality of clean coal produced in fine-coal circuits is controlled as much by how efficiently 0.15 mm (100 mesh) particles are sized as on the performance traits of the fine-coal cleaning device itself. In addition, the increased interest in effective fine-coal sizing operations indicates that many operators have also discovered its importance in affecting the quality and yield of clean coal produced in fine-coal circuits. This paper describes the three most prevalently used fine-coal screening technologies and discusses their sizing mechanism, performance, and application.

Jan 1, 1988

By George R. Huebner

The White Pine Copper Company, subsidiary of Copper Range Company, is located in the Upper Peninsula of Michigan. The offices, plant, and mine are six miles south of Lake Superior in Ontonagon County. Employing 1700 people, White Pine is the largest single industry in the Upper Peninsula. 970 underground employees produce 16, 700, tons of ore daily. From this tonnage 363,400 pounds of refined Lake Copper are produced in various shapes as specified by the customer. The following paper explains the geology of the ore body and briefly outlines present mining methods. It enumerates the reasons for considering longwall mining of a hardrock* ore body and the preliminary investigations leading to the decision to undertake the initial test panel. Selection of the initial test panel location is covered in detail. Equipment choice and reasons for that choice are described as well as selection and training of personnel.

Jan 1, 1964

By Jim Merza

The recent heightening of environmental issues has forced more and more mining operations to focus on developing new methods to control dust problems encountered at their concentrate dryers, stock piles, tailings ponds and during the loading, transportation and unloading of their products. The development of dust control agents requires a quick reliable and most of all accurate method to evaluate the performance and to optimize the formulation of these products in the laboratory, thus avoiding expensive field trial evaluations. This paper gives an overview of the industry requirements for specialty dust control agents, application techniques and current laboratory evaluation procedures. The paper will also outline the development of a unique dust suppression testing apparatus called the "Dust Box, which is used to evaluate chemical dust suppressants at the laboratory level. Laboratory results from two test programs on different substrates are presented in the paper.

Jan 1, 1994

By D. Bizier, B. H. Gardner, Y. P. Wang, S. Serata

As part of an ongoing effort to develop more cost-effective and “user friendly” hardware for geomechanical measurements required for underground mine design, Serata Geomechanics, Inc. (SGI) has investigated the potential of borehole pressure cells (WCs) for in situ measurement of stress states and continuous monitoring of changes in ground stress. In particular, because of certain advantages gained thereby, WCs utilizing an elastomer, instead of metal (e.g., copper), for the skin of the pressure cell have been developed. A data analysis technique for determining ground stress state directly from BPC data has been developed. This technique has been tested for salt ground through both in situ applications and computer simulation analysis of the BPC/borehole boundary interaction. This paper discusses the data analysis technique, the results of field and computer simulation testing of its validity, and evaluates the advantages of elastomeric BPCs in relation thereto. It utilizes field data from recent elastomeric BPC applications in the salt mine at Goderich, Ontario operated by the Sifto Division of Damtarn Inc.

Jan 1, 1986

By Shouci Lu, Keji Sun, Dianzuo Wang, Yiqun Li, Chuang Jin

The dephosphorization of iron ore and reduction in Fe loss in process are currently the foremost technological challenges facing the mineral industry of Meishan Iron Ore. These problems can be resolved to a greater extent by an apatite flotation process which depends on selective flotation reagents. The primary objective of this research work is to develop new type of flotation collectors for dephosphorization of weak magnetic iron ore. Detailed discussion is given on the influence of collector dosage, pH, dispersants and pulp temperature on the flotability of apatite, siderite and hematite using the collectors of Oleyl sarcosinate(OSS), Alkyl polyoxyethylene ether phosphate (AEP) and the mixtures of OSS and AEP A key feature of the flotation was the use of the mixtures of OSS and AEP as a selective collector for phosphate. The effects of chemicals and operational variables on the phosphorus flotation from the weak magnetic iron ores have been investigated. The test results shows that floating the phosphorus minerals with new type collectors can reduce the phosphorous content in the flotation concentrate to an acceptable level of less than 0.20% P with increased Fe recovery.

Jan 1, 2002

By W. M. Ma, Daniel W. H. Su, K. Centofanti, Yi Luo, W. L. Zhong, Syd S. Peng

10.1 I~RODUC1'ION In order to understand the effects of surface subsidence due to underground mining on the ground water hydrology, a brief introduction to the rate, direction, and general pattern of ground water movement in aquifers is necessary. Ground water movement in an aquifer is controlled by aquifer lithology, structure, permeability, water levels, topographic set¬ting, and pumping patterns (Sgambat et al., 1980). Fig. 10.1 shows an idealized flow pattern of ground water in a hilly terrain (Coe and Stowe, 1984). The hill is made up of a series of interbedded permeable aquifers (mostly sandstone) and confining units (mostly shale). Since sandstone is porous and permeable, ground water moves through it downward and laterally until it reaches the interbeddings of the confining unit immediately below. The ground water then moves laterally along the interbeddings, be¬cause the confining unit is less permeable, and comes out as springs on the surface at the intersection of the aquifer and the confining unit. If there is a ground water table in each aquifer, it is a perched water table (e.g., right side of Fig 10.1). On the other hand, if the ground water recharge is larger than can be drained out, the interbedded permeable aquifers and confining units will be filled with water and a ground water table will exist only at the uppermost permeable aquifer. This is the semi-perched water table (e.g., left side of Fig. 10.1). There are three ground water flow systems: shallow, interme¬diate, and deep. Shallow flow systems occur near the land surface and the surface drainage system. The flow depth ranges from a few to several hundreds of feet with a resident time of a few weeks to a few years after entry, depending on the permeability of the rock strata. In a shallow flow system, the drainage coincides with sur¬face run-off water divides. Intermediate flow systems occur in minor hills some distance away from surface drainage. The travel distance ranges from several hundreds to several thousands of feet, the travel time from several months to tens of years. These systems are less dependent on seasonal recharge or discharge. In deep flow systems, the ground water travels from thousands of feet to miles, and transfers between surface basins. The travel time ranges from several tens to several hundreds of years. The systems generally receive recharge over a large area on major mountains and dis¬charge into major stream systems. 10.2 CASE STUDIES There are two ways of monitoring water wells for studying the effects of dewatering due to underground mining, i.e., water level and piezometric level. The water level is the water elevation in a water well that has penetrated two or more aquifers. Thus it indi¬cates the combined flow of all the aquifers that the water well intersects during the time of measurement. The piezometric level is the level to which water rises in a piezometer cased to a specific aquifer, and coincides with the pressure level of water in that aquifer. Therefore piezometric level reflects the water condition in that particular aquifer to which it is confined, while the water level represents the total yield of a water well without regard to the source of water. For instance, if a water well intersects two aqui¬fers, and if the lower aquifer is affected by undermining and goes dry while the upper aquifer does not, the water level in the well will drop, but may never go dry. Under this condition, the data obtained by monitoring the water level do not indicate the true behavior of the individual aquifers, especially if they are located at widely separated heights. 10.2.1 WATER LEVEL EFFECTS The effects of underground mining on ground water flow have been the subject of many investigations (Booth, 1984; Coe and Stowe, 1984; Owili-Eger, 1983; Rauch et al., 1984). However, the problems are very complicated, as have been exemplified by the conflicting conclusions reached by various investigators in var¬ious case studies. The most dramatic examples were described by Coe and Stowe (1984). In the first case (Fig. 10.2), the coal seam was 5 ft thick with cover from 400 to 800 ft deep. The majority of water supplies came from a sandstone aquifer in most ridge tops. When the face of the Panel No. 3 passed under it, dug well WW-1, which was approximately 670 ft above the coal, went dry, while the water level in Borehole B- 1, which was located at the center of Panel No. 4 with the bottom reaching the coal seam, also dropped about 160 ft. In Panel No. 5, the developed spring DS-2, which originated from the surface of the sandstone/shale contact and was approximately 750 ft above the coal seam, stopped when the face passed under it. The water level in the dug well W-2, which was dug into the shale about 700 ft above the coal seam dropped about 22 ft, but the well did not go dry. Therefore, nearly every spring or well located above or near the active panels was affected by longwall mining. In the second mine (Fig. 10.3), the seam was 5 ft thick with cover from 200 to 400 ft thick. Dug well DW-2, the bottom of which was 35 ft deep and approximately 330 ft above the coal seam in a perched sandstone aquifer, went dry right after the longwall passed under it. Drilled well W-3, in the sandstone aquifer, about 220 ft above the coal seam, experienced a 75-ft drop in water level when the face was under it, but showed some re¬covery after the face had passed under it. Ponds P- I and P-2 were developed in shale bedrock and were spring fed. They were not affected by longwall mining. Pond P-3 was developed in the sand¬stone aquifer, and its water table showed some decline when the face passed under it. Spring ST-1, which originated from Pond P-1 and was fed from several tributary valleys, was affected by longwall mining, probably because it passed across the panel edges where tension cracks drained off the water. Although the overbur¬den in case 2 was about half of that in case 1, the majority of water resources were not affected by longwall mining. Coe and Stowe (1984) attributed this discrepancy to: (1) a higher percentage of clays, shales, and claystones, resulting in a lesser degree of inter¬connected fracturing, and (2) a less rugged topography resulting in a less localized ground water flow system. Therefore, in order to assess the effects of underground mining, the premining ground water flow patterns, topographical, lithological, and meterological conditions must be known. Booth (1984) also monitored the ground water effects of a longwall mine (Fig. 10.4) in a seam 5 ft thick and 300 to 800 ft deep. There were three major aquifers: Morgantown sandstone approximately 30 ft thick and 630 ft above the coal; Saltsburg sandstone of up to 50 ft in thickness and 520 ft above the coal; and

Jan 1, 1992

By W. M. Schafer

The risk that water quality will be degraded by mining activity can be evaluated using risk analysis techniques. Water quality impacts can occur where water-rock interactions with tailings, waste rock, or leached ore lead to increased loading of constituents into surface or groundwater. Constituent loads contributed by mining facilities are determined by the migration route and chemical evolution of mine water. Mass load contributions from mines can be determined by estimating the prototype pore water in contact with mining waste, predicting the rate and timing of flow, assessing the degree of geochemical attenuation, and modeling chemical mixing with natural surface or groundwater. Mining facility siting and design measures can alter the chemical evolution of mine water by varying the quantity or quality of flow. Stochastic risk management techniques provide a powerful tool for quantitatively identifying the environmental consequences of various siting and design alternatives.

Jan 1, 1998

By A. L. Muler

Since 1956, when Broadbent and Callcott 1 proposed to analyze breakage processes by matrix methods, several authors and their co-workers have proposed various models of grinding that, in principle, could lead to fully automated grinding circuits. In particular, Meloy and Gaudin2, Gardner and Austin3, Kelsall and Reid4, Neloy and Bergstrom5, Callcott and Lynch 6, and Draper and Lynch have, by either computer simulation or by experimental data, or by a combination of the two, verified that comminution is reasonably well described by at least two functions: a selection function and a breakage function. The stage concept proposed by Callcott and Lynch6 requires that a classification matrix be included to completely describe breakage in an open circuit rod mill. The purpose of this work, portions of which are herein described, is to accumulate sufficient information on the breakage characteristics of quartz and magnetite in an open circuit rod mill, so that several of the theoretical and experimental approaches proposed in the literature can be compared with each other. Data is being gathered to compare the experimental method used by Kelsall and paid against the stage approach of Callcott and Lynch.

Jan 1, 1968

By Erling Stranden, Leiv Berteig

INTRODUCTION The measurement of the total activity of radon daughters in the air of mines has become a routine procedure in order to control the radiation exposure in miners due to the inhalation of these radionuclides. Normally the measured concentration is given in terms of total potential L-energy of the short lived radon daughters. The unit most frequently used is 1 WL (working level), which is defined as the potential a-energy of 1.3.105 MeV per 1 of air. The a-dose to the respiratory tract is, however, not proportional to the inhaled potential a-energy. In addition, the degree of equilibrium between the daughter products in air and the fraction of daughter products not attached to aerosol particles (the unattached fraction) must be known. Theoretical calculations of the equilibrium between daughters and the unattached fraction have been presented by different authors (Jacobi, 1972; Porstendörfer et al, 1978). Such calculations must take into account the plateout of daughters on surfaces, the attachment of daughters to aerosols and removal of daughters by the ventilation. The situation in mines will however be very complex and no single model will cover all situations. In most cases the radon and radon daughter concentrations will be described by a combination of different models. Domanski (1979) used the experimental results of Breslin et al. (1969) from uranium mines to relate the equilibrium between the individual daughters to the equilibrium factor, F. Once the equilibrium factor is known, this seems to be a useful approach to assess the individual daughter concentrations. The concentrations of radon and daughter products may vary considerably during the day. Seasonal variations are also frequently found. It is therefore important to have knowledge of the magnitude of these variations and of the factors having the strongest influence upon the concentrations. The unattached fraction of radon daughters in uranium mines was investigated by George and Hinchliffe (1972) and Raghavayya and Jones (1974). Mercer (1975) later modified the results of Raghavayya and Jones. Both these investigations indicated that the aerosol particle concentrations was the main influencing factor upon the unattached fraction. In this paper we will summarize the main results of a study on the radiological characterics of nonuranium mines. We will also discuss the correlations between the unattached fraction of the potential a-energy and the unattached fraction of the individual daughters, and between the equilibrium factor F, and the individual daughter ratios. More detailed discussions of this study will be published elsewhere (Stranden and Berteig, 1981 and Stranden and Berteig 1981 a). MATERIAL AND METHODS During 4 two-weeks periods corresponding to the seasons in the Norwegian mine where the highest radon daughter concentrations occure, the parameters listed in table 1 were studied. TABLE 1 Parameters measured during the study [Parameter Methods Radon Scintillation flask/Continuous ionization chamber Radon daughters Tsivoglou(1953)/Kusnetz(1956) Respirable dust Cyclone Unattached fraction Wire screen (Stranden and Berteig 1981) Atmospheric pressure Continuous baragraph Temperature and Continuous thermohygrograph humidity] At a fixed location, the radon concentration was measured continuously by a ionization chamber (Stranden, 1981), during these two-weeks periods in order to study short-time and seasonal variations. RESULTS Correlation between equilibrium factor and individual daughter ratios. The mean value of the equilibrium factor in all Norwegian mines measured during the last ten years was 0.46 with a minimum of 0.08 and maximum of 0.93 (Stranden and Berteig, 1981). The correlations between the equilibrium factor F, and the individual radon daughter/radon ratios were studied and the results indicated high correlations. In table 2, these correlation formulae are listed together with the correlation coefficients. TABLE 2 Correlation between the equilibrium factor, F, and individual radon daughter ratios [Formula Correlation coefficient RaA/Rn = 1.14 F 0.551 0.76 RaB/Rn = 1.00 F 0.940 0.99 RaC/Rn = 0.77 F 1.120 0.92]

Jan 1, 1981

By S. K. Kawatra, S. A. Moffat, K. A. DeLa’O, T. C. Eisele

It was observed that the efficiency of some autogenous grinding circuits was dependent on temperature, with the energy-efficiency of the grinding process decreasing in the winter months, accompanied by an increase in critical-size production. Several possible causes of this phenomenon were suggested, including: • Increased brittleness of the ore during blasting in the pit at low temperatures, leading to increased levels of internal cracking that would lead to rapid breakage to critical size in the mill; •Brittleness changes of the rock in the mill; and • Changes in the durability of the rock caused by either formation of ice on the surface, or freezing of water inside the pores and cracks of the rock. To determine which mechanisms were occurring, drop tests were carried out using frozen and unfrozen rock with varying degrees of porosity, both dry and water-saturated, and which had been blasted for mining in summer and in winter. These tests showed that there was no difference between the summer-blasted ore and the winter-blasted ore, and that durability changes in dry rock over the temperature range from 25ºC to -25ºC were negligible. However, water-saturated rock showed an increased resistance to breakage upon freezing, particularly when the rock was highly porous, or when a layer of ice formed on the surface. Seasonal changes in the durability of rock in grinding are therefore due to formation of ice in and around the rock.

Jan 1, 1995