Search Documents

By A Minchinton

With blasting, all far-field phenomena including, but not limited to, stress wave propagation, fracture, vibration and muck pile formation are intrinsically and irrevocably dependent upon near-field phenomena or mechanisms that we describe using physics-based mathematics typically transcribed into computer codes. The energetics involved in such near- to far-field transitions are not fundamentally relevant as they are simply progenitors of these processes.The notion of explosive energy as a descriptor for explosive or blasting performance is a case in point. These energies, which are generally used in blasting tables, are almost exclusively based on ideal detonation theory alone and do not include the influence of the rock, charge diameter, blasthole length and the late-time dynamic expansion that can be predicted using codes based on non-ideal detonation theory.With decoupled charges in presplits for example, the pressure drop between the charge and the blasthole wall can be simply described to first order using the energy relationship associated with adiabatic expansion. However, a more realistic mechanistic solution involves interacting fluids and turbulence leading to dynamic processes such as the channel effect.CITATION:Minchinton, A, 2015. On the influence of fundamental detonics on blasting practice, in Proceedings 11th International Symposium on Rock Fragmentation by Blasting, pp 41–54 (The Australasian Institute of Mining and Metallurgy: Melbourne).

Aug 24, 2015

By M J. Thurley, M Wimmer

To assess the present-day functionality of large-scale sublevel caving at LKAB Kiruna a comprehensive measurement program was undertaken involving blast function, fragmentation and gravity flow. As part of this assessment, a fragmentation measurement trial was performed based on 3D imaging of the drawpoint and corresponding bucket load of the underground load-haul-dump (LHD) excavator. 3D image data from stereo photogrammetry was collected and an automated image analysis strategy developed. A number of data sets were collected for each of the drawpoint and LHD bucket, along with sieving results for four of the LHD bucket loads (totally about 70 t). Two of the sieving results were used to inform the automated image analysis strategy, and two were held back as a comparison. Large variations in the visible particles are apparent when comparing corresponding drawpoints and LHD buckets highlighting the impact of sampling location and the need to measure large quantities of data in order to avoid bias from small samples. The results show that 3D imaging and analysis can produce fully automated measurement and analysis of the visible particle size distribution. Although this is not the same as the sieve-size distribution it provides useful estimation of both the larger size classes and a bulk estimate of fine material below approximately 60 mm. The 3D stereo photogrammetry measurement system used produced very high 3D point density but this was achieved using a custom up-sampling technique which significantly smoothed the data, removing small particles, smoothing edges, and this negatively affected the particle delineation algorithms.CITATION:Thurley, M J, Wimmer, M and Nordqvist, A, 2015. Blast fragmentation measurement based on three-dimensional imaging in sublevel caving drawpoints and underground excavator buckets at LKAB Kiruna, in Proceedings 11th International Symposium on Rock Fragmentation by Blasting, pp 763–774 (The Australasian Institute of Mining and Metallurgy: Melbourne).

Aug 24, 2015

DCF Valuations are very important to individuals and companies involved in the mining industry. Such valuations are critical to investment decisions in relation to new and existing projects. In particular they provide fundamental information to shareholders in the event of a capital raising or takeover and to financiers supporting mining companies and projects. Also, DCF Valuations are often used to support asset carrying values in the financial statements of mining companies. This paper presents a Chartered Accountant's (CA's) perspective on the main issues to be considered in preparing, analysing and relying upon DCF methods in mineral valuations and provides some practical examples relating to key issues. The paper makes the important point that DCF methods involve making assumptions about future events and circumstances that inherently contain uncertainty which results in a valuation prepared on a DCF basis providing a range of values that are not perfect or accurate. In order that DCF valuations provide meaningful information the underlying methodology used and assumptions made must be consistent, sound and well supported. Credibility of those involved in preparing mineral valuations is vital and this paper supports the adoption of sound practices by all concerned.

Jan 1, 1994

Nickel-bearing ores in North-east Dundas have been described by A. McIntosh Reid1 as occurring at the margins of narrow dykes of gabbro, norite, peridodite and pyroxenite, which are intrusive into the Dundas series of slates, quartzites and breccias. They occur chiefly in the neighbourhood of the Five Mile, near Zeehan, as small, narrow ore-bodies which are mostly less than 100 ft. in length.The ores are massive sulphides containing pyrrhotite, chalcopyrite, pyrite and the nickel sulphide, pentlabdite.A less common nickel-bearing mineral, gersdorffite, has been recently found in an adit, which is marked on Reid's geological plan of Dundas as the Hecla mine, 30 chains south of the crossing of the North-east Dundas tramway, over the Great Northern Creek, and at an, elevation of about 350 ft. above the tramway. The occurrence is in a lode traversing the Dundas slates and quartzites, and is situated, according to Reid's plan, about 20 chains beyond the northern extremity of a narrow basic dyke which is continuous for a distance of two miles in a south-westerly direction.The mine eonsists of a prospecting adit following the course of a lode for about 275 ft. in a direction a little west of south. The low values of the lode are indicated by a bulk sample from a shallow winze about 100 ft. from the entrance which is reported by Reid to contain:

Jan 1, 1935

By G Gilchrist, S C. Dominy, T Jashashvili, R C. A Minnitt

The grouping and segregation error is one of Gy’s seven errors that cannot be measured in any qualitative manner because of its transient and changing characteristics each time the sample is handled mechanically. Attempts to calculate the mean and variance of the segregation error is considered by some as ‘an exercise in futility’ that would not help us achieve representative sampling. This research is aimed at understanding the process of segregation of dense, finely divided and dilute particulate materials supported in a clean quartz substrate during intense agitation. Several mixtures of fine-grained materials including steel, tungsten carbide and gold grains were used in these experiments. Four hundred grains of each of the minerals and metals at a top size of 2 mm were mixed with approximately 600 g of finely ground quartz material (95 per cent passing 2000 microns) in a 500 ml plastic jar with a sealable lid. The mineral–?metal mixtures were analysed in three steps using microfocus X-ray computed tomography. Step one consisted of rolling the containers on a flat surface in an attempt to concentrate the mineral and metal grains in the core of the supporting substrate; this is a base case that can be reconstructed after each session of analysis. The second step involved gentle and vigorous shaking of the mixtures in order to segregate the denser minerals and metals from the supporting quartz substrate. In some cases, a third step involved splitting the sample through a riffle splitter. Three-?dimensional images of the surface of each grain, its volume and its X-, Y- and Z-coordinates, before and after intense agitation of the container, were interpreted from the reconstruction of the tomographic X-ray data. The spatial analysis of grains using cell declustering and 3D variography provided a useful means of characterising the distribution of the grains in space and the extent to which they had become segregated during the process of agitation. Spatial analysis indicates that in general, grouping and continuity of grains is weakly developed. The modelled nugget effect remains high for most tests, indicating that grains of similar size did not group together strongly. Agitation through rolling provided the best continuity, although not consistently in the long axis of the plastic jar. The nugget effect was consistently highest, and the range of continuity shortest in samples that had been split, indicating that the grouping and continuity were reduced through splitting the sample. Initial results indicate that very little segregation of the minerals and metals occurs in the quartz substrate. This is somewhat counter-intuitive, but suggests that the effects of agitation on mixtures of dense minerals and metals in substrates of lower density do not lead to excessive segregation. Differences in behaviour between dense materials in response to agitation are evident. Grains of different composition and shape did not group and segregate in a predictable manner in response to the agitation methods. Initial results indicate that the degree to which a material groups and segregates relates more to the characteristics of the denser material itself, rather than the agitation method employed.CITATION:Minnitt, R C A, Jashashvili, T, Gilchrist, G and Dominy, S C, 2017. Quantifying segregation of minerals and metals in particulate materials using computed X-ray tomography and variography, in Proceedings Eighth World Conference on Sampling and Blending , pp 173–182 (The Australasian Institute of Mining and Metallurgy: Melbourne).

May 9, 2017

This paper gives an overview of the difficulties associated with determining the strength of hard rock pillars. Although a number of pillar design tools are available, pillar collapses still occur. Recent examples of large scale pillar collapses in South Africa were caused by weak partings which traversed the pillars. Currently two different methods to determine the strength of pillars are used namely, empirical equations derived from the back analyses of failed and stable cases and numerical modelling tools with an appropriate failure criteria. Both techniques have their limitations and additional work is required to obtain a better understanding of pillar strength. Empirical methods are popular and easy to use, but care should be exercise that the results are not extrapolated beyond the range of the data which were used to derive them. An example is the Hedley and Grant formula (derived for the Canadian uranium mines) that has been used for many years in the South African platinum and chrome mines (albeit with some adaptation of the K-value) to design the pillar layouts. Very few collapses have nevertheless been reported and it might even be too conservative an assumption of pillar strength in these mines. As an alternative, some engineers strongly advocate the use of numerical techniques to determine pillar strength. A close examination unfortunately reveals that these techniques also rely on many flawed assumptions. An area where numerical modelling is invaluable, however, is to study specific pillar failure mechanisms, such as the influence of weak partings on pillar strength. In conclusion, it appears that neither empirical techniques nor numerical modelling currently provide a solid basis to conduct pillar design. It is therefore recommended that both these techniques be utilised when confronted with a pillar design to obtain the best possible insight into the problem.

Nov 25, 2010

A single compartment 100 A electrolytic cell charged with AICla-KCl-NaCl at 750°C was studied during an experimental campaign of 4 months. About 10 1b of aluminium were produced in continuous runs which lasted up to 1 week. The highest cathode current efficiency determined was 80 per cent. Anode current efficiency. cathode potential, analyses of melt, sublimate, and product, and reactivity of cell materials were also recorded and interpreted in terms of the role of lower valent aluminium species in the electrolysis mechanism.

Jan 1, 1971

By D Morgan, D Sato, B Lumsden, S Senanayake

Laboratory flotation tests on fresh cyclone overflow samples are a useful tool to troubleshoot and indeed to improve plant performance; however, reproducibility between laboratory flotation test and plant performance is important for interpretation of the laboratory tests and hence their usefulness. Chemical reproducibility can be achieved with the addition of the required reagents; however, reproducing the physics of forces and energy in a 5 L cell compared to a 50 000 L plant cell or a 500 000 L plant cell is not generally studied or discussed. For some tests such as collector evaluation the physics of the flotation process may not be critical, but for some evaluations in the laboratory, it can be important.All flotation plants suffer from losses of valuable <20 µm mineral. Selective mineral aggregation has long been recommended to improve the poor flotation recovery of <20 µm sulfide minerals; however, successfully testing any type of selective sulfide mineral aggregation’s effect on flotation recovery in the laboratory on practical plant samples requires reproducing plant cell physics. The energy input per litre and the forces in a laboratory cell are vastly different to a plant cell. Reproducibility between laboratory and plant tests requires that these differences be minimised. The physics of the laboratory flotation cell is particularly important when fine mineral aggregation tests are carried out.This paper outlines a modified laboratory aggregation and flotation method that closely reproduces plant performance in a laboratory test.CITATION:Sato, D, Lumsden, B, Senanayake, S and Morgan, D, 2014. Development of a laboratory flotation method to test plant mineral aggregation, in Proceedings 12th AusIMM Mill Operators’ Conference 2014 , pp 173–178 (The Australasian Institute of Mining and Metallurgy: Melbourne).

Sep 1, 2014

Evaluation of Grinding Medium Wear in Stirred Mills The standard fine grinding machines in the minerals industry are stirred mills whose performance largely depends on the competency of the grinding medium. Medium wear in stirred mills is one of the key parameters for medium selection. A wear evaluation methodology is developed and applied in this paper to measure the wear rate of river sand using a laboratory batch stirred mill. The procedure is demonstrated by testing the sand medium in a 0.55 kW batch stirred media detritor (SMD). The sand wear rate determined in the laboratory batch mill is successfully scaled up to that determined on a full-scale Metso 355 kW SMD mill.

Sep 13, 2010

By C J. Adams, H J. Campbell, A J. Tulloch, I J. Graham

New Zealand can be regarded as a rifted part of mineral-prospective eastern Australia. Geologically credible comparisons between selected trans-Tasman and SW Pacific geological provinces and units, especially those that contain significant mineralisation in Australia, can help provide a useful exploration framework for New Zealand. As part of a six-year Mineral Wealth of NZ and its EEZ (MWE) research programme, we have targeted selected geological units in eastern Australia for comparison and contrast with their potential New Zealand counterparts. In future years, the programme will focus on geological units in New Caledonia.

Jan 1, 2005

Mount Isa Mines Limited first installed electronic data processing equipment in 1959. The decision to replace this equipment with a computer was motivated by the need for greater processing power and the availability of new and less expensive computers. Consideration of hardware software and delivery features of available equipment, resulted in the selectIon of an ICT 1900 series computer which was installed in March 1966. Special air-conditioning and power stabilization equipment was also required.Systems design and programming was carried out by personnel recruited from within the Company and trained by the computer manufacturer. Early work was hindered by the problems of remote location.Initial applications were basically accounting functions whIch were taken over from the EDP installation. Subsequently computer applIcatIOns of a more complex technical nature have been undertaken. Future work will include both "scientific" and data processing applications.

Jan 1, 1968

By K Firmin

The Mt Tom Price orebody of Hamersley Iron Pty Limited consists of a folded hematite - shale sequence capped by a layer of goethite. Consequently, variations in iron content occur within the ore body, and as the overall grade is below that required to make specification products, selective mining must be carried out. A necessary preliminary to quality control of products based on selective mining is the production of advance data on the tonnages and grades of material to be expected in future mining areas. Bench plans showing zones of different grades and types of material anticipated are prepared from structural geological information. The grades of materials on a bench are essentially governed by the positions and attitudes of the shale bands. Quality control is exercised by blending different grades and types of material in primary crusher feed. Iron grade is the prime grade consideration although sometimes the level of phosphorus content must be considered. Pre-production grades are determined from assay of samples taken from blast hole cuttings. As shales are more friable than hematite, product fines are always lower in iron grade than the lump product. The realisation of contract specification grade is more difficult with fines than with lump, so maintaining the correct fines grade becomes a prime factor in quality control. Control of fines grade is presently effected by aiming for an appropriate feed grade to the primary stockpile. Studies to estimate fines grade directly by qualitatively taking into account material types are in progress. Quality control measures are checked by monitoring the crushing and screening plant output, and taking appropriate compensatory actions in the mine when required. Minor fluctuations in quality of fines can be blended out on stockpiles. The quality control measures described ensure constant quality products, and a good recovery of high grade material from the deposit.

Jan 1, 1971

By J P. Barkas

World demand for iron ore is far from peaking, and the substantial price decline that has captivated commentators since early 2014 is not the result of widespread misreading by producers of the prospects for global market growth. Recent price movements provide little useful insight into the industry’s longer-term fortunes, apart from reminding mining companies that their businesses need to be robust against the volatility and inescapable cyclicality of mineral commodity markets. It is timely to take a strategic view of the history and longer-term prospects for the iron ore industry and of Australia’s place in it.CITATION:Barkas, J P, 2015. Perspectives on the long-term outlook for iron ore, in Proceedings Iron Ore 2015, pp 3–6 (The Australasian Institute of Mining and Metallurgy: Melbourne).

Jul 13, 2015

By G Anderson, V Lawson, P Voigt

Glencore operates zinc concentrators at both Mount Isa and McArthur River. Both orebodies have a history of complexity and required the development of new technologies to maximise the value from these long-life assets. The Mount Isa orebodies were plagued since operations started in 1931 with complex mineral textures making separation challenging and many areas uneconomic. Examples are given in Figure 1 with different grain sizes from the same orebody in different areas. The McArthur River deposit, located in the Northern Territory, was discovered by Mount Isa Mines geologists in 1955 (Logan, Murray and Williams 1990). The extremely fine-grained nature of the ore was quickly recognised and the difficult separation of the minerals was soon obvious. The geographical, metallurgical and economic factors resulted in the orebody not being developed until the technologies for the mineral separation had been developed. The development of the IsaMill allowed for an economic mineral separation flowsheet to be developed and ore treatment commenced in May 1995 (Nihill, Stewart and Bowen 1998). CITATION: Lawson, V, Anderson, G and Voigt, P, 2018. Developing technologies for a complex orebody, in Proceedings Complex Orebodies Conference 2018, pp 116–118 (The Australasian Institute of Mining and Metallurgy: Melbourne).

Nov 21, 2018

Teleoperation is one method by which an operator can remotely control and monitor a vehicle. The operator monitors the status and progress of the remote vehicle by watching the video transmitted from one or more cameras which are mounted on the vehicle. In the case of an LHD, it will be fitted with three cameras. For a twin-boom Jumbo, four cameras are mounted on the drill, three of which will have remote pan, tilt, 15x zoom, focus and lens cleaning controls. The advantage of teleoperation is that an operator can stand in the drive and send the remote vehicle around a corner, completely out of sight and still clearly see what is happening. This means that an LHD returns with a full bucket and a Jumbo can be operated from a safe distance. Standing in the drive teleoperating vehicles, exposes the operator to passing traffic including the remote vehicle. There are two solutions to this problem. One is to equip the operator with a Nautilus æBuddyÆ jacket which will help to protect them from all vehicles. The other is to move the operator into a safe room or area, away from the fumes, the noise and the traffic. In this comfortable, less stressful environment, the operator will be more productive and less likely to accidentally damage the vehicle. The operatorÆs æsafeÆ room does not have to be a formal control room, it can be any area which has AC or DC power available and is large enough to contain a chair and a 1.5 m x 0.75 m table. The æsafeÆ room can be on the surface or underground, it does not matter to the operation of the Nautilus Equipment. If the LHDÆs, trucks, locomotives, etc are automatically guided from origin to destination, the operator is free to operate several vehicles at a time, performing only those tasks which are too difficult for a computer to do by itself.

Jan 1, 1998

By Moorrees C, Everson P

A detailed examination of carbonic acid leaching of crude iron-containing magnesia derived from Savage River magnesite shows that the rate and extent of magnesium and iron dissolution are affected by the slake time, leaching temperature, carbon dioxide pressure, pulp density and rate of agitation. In order to produce high-purity magnesia from the leachate by precipitation and calcination, the optimum leaching conditions, which are those that yield maximum magnesium and minimum iron dissolution, have to be a compromise in terms of the above parameters. For example, iron dissolution can be limited by leaching at an elevated temperature (-45°C), but because the maximum concentration of magnesium biearbonate decreases rapidly with an increase in temperature, it is necessary to use a low pulp density (-2 per cent solids) in order to achieve an acceptable degree of magnesium dissolution.Evidence is obtained that iron dissolution involves the formation of a soluble ferric bicarbonate complex that decomposes at elevated temperatures. The extent of iron dissolution is much higher than might be expected and it is probably because the original magnesite contains siderite in solid solution with the magnesite; on calcination, the hematite whieh is formed is in a very finely divided condition and hence more reactive.

Jan 1, 1983

By Lottrey A. L, Sowerby B. D, Cutmore N. G

On-line monitoring of low rank coal quality during input to a process has the potential to provide benefits through: ò anticipation of coal quality variations prior to utilisation; ò provision of a tool to help determine the influence of coal quality on process behaviour, and ò provide a direct measure of coal quality for pricing. Moisture and ash are the two most significant coal quality parameters for combustion. They determine the energy content which can be derived from the coal and the problems arising from the presence of ash in the process (eg ash fouling, corrosion, erosion, particulate emissions and ash disposal). Extensive pilot-scale testing and full-scale experience has shown that specific coal/ash constituents (eg Na2O/ash ratio, acid-soluble aluminium, silica, sulphur contents, etc) are also important for the behaviour of Latrobe Valley brown coal during combustion, ash formation, deposition and removal of ash in precipitators and furnace ash hoppers. This paper discusses the range of on-line analysis techniques which have been investigated and their application to low rank coal analysis. Preliminary results from a joint SECV/CSIRO project which aims to develop two separate advanced instrument systems are also presented. These systems are based on: ò microwave attenuation and phase shift - for direct determination of moisture content; and ò inelastic neutron scattering - to provide total hydrogen, carbon, chlorine and ash contents, with potential for calculation of moisture content, specific energy, sodium content and Na20/ash ratio using correlations.

Jan 1, 1992

By D J. Black

Ventilation modelling is used to evaluate ventilation system performance. This paper describes modelling to assess potential changes to a single entry longwall ventilation circuit to improve ventilation and gas management. The longwall panel is extracting a 215 metre face in a 1.5 metre coal seam at a depth of 900 to 1000 metres below surface. The 1.5 metre B seam is located approximately 25 metres below the main 3.0 metre A seam and the gas content of both coal seams is approximately 20 to 25 m3/t. The B seam is extracted prior to the A seam and the mine relies on fracturing of the overlying strata to induce gas emission from the A seam to reduce the gas content prior to mining. Longwall operations in the B seam are constrained by limited ventilation and high gas emissions. Due in part to statutory limits on gas concentration and air velocity, the average longwall production rate is approximately 700 t per day.Gate roads in the B seam are single entry, developed using road headers and supported with steel arches. Additional timber cog support is installed to reduce roadway closure from abutment loading. Due to the slow advance of gate road development, mine management were reluctant to consider significant mine design changes, such as two heading gate roads; therefore, the investigation focused on determining the extent to which ventilation and gas emission impacted on longwall production performance, while recommending short-term, low-cost actions to improve ventilation and gas management to support increased longwall production.CITATION:Black, D J, 2017. Application of modelling to improve ventilation and gas management in a single-entry longwall panel, in Proceedings Australian Mine Vent Conference 2017, pp 239–244 (The Australasian Institute of Mining and Metallurgy: Melbourne).

Aug 28, 2017

By B Belle

Dust sampling is pivotal in estimating the ‘dose’ of dust exposure and in deriving dose-response curves in epidemiological studies. Over the last half century, gravimetric sampling has been the fundamental means for dust exposure monitoring using recognised respirable size-selective standards. In Australia, a gravimetric sampling technique has been followed since 1983 using the Higgins-Dewell (HD) type cyclones (AS2985). The re-emergence of ‘Black Lung’ or Coal Workers Pneumoconiosis (CWP) in Queensland, Australia has re-kindled the need for better understanding of personal dust monitoring, compliance determination, accuracy, timeliness of sampling results and dust control systems. Reporting dust levels in real time empowers miners and operators to take immediate action to avoid being exposed to excessive airborne dust. Two decades of intensive research in the USA has led to the introduction of gravimetric based continuous personal dust monitors (CPDMs) called PDM3700 as compliance tools. The new real-time mass based sensor is a continuous mass based dust monitor using the Tapered Element Oscillating Microbalance (TEOM) principle and is regarded as significantly superior to the current gravimetric sampling methodology.Recently the overall respirable dust standard in USA coalmines has been reduced from the historic 2.0 to 1.5 mg/m3. This paper provides the results of a comparative pairwise study of an existing gravimetric sampler and the PDM3700 carried out in 3 different Australian underground mines. The field results consistently suggest that there is a significant ‘measurement bias’ between the current gravimetric HD sampler and the PDM3700 monitor operated using a similar HD type cyclone. The results show that the currently used cyclone measurements are approximately 59 per cent higher than the ‘auditable’ PDM3700 monitor data at the current compliance limit of 3 mg/m3. Based on a review of the extensive data collected, the differences can in part be attributed to the aerosol size selectivity of the samplers with the current gravimetric sampler used in mines, collecting more larger sized particles (D90 of 15 µm). This finding has significant consequences for historic exposure results and their use in medical surveillance programs, as well as the current approach to non-compliance determination.CITATION:Belle, B, 2017. Pairwise evaluation of PDM3700 and traditional gravimetric sampler for personal dust exposure assessment, in Proceedings Australian Mine Vent Conference 2017, pp 87–92 (The Australasian Institute of Mining and Metallurgy: Melbourne).

Aug 28, 2017

By T Davis

Cameco Corporation operates the high-grade 140 tpd fly-in/fly-out underground McArthur River Uranium Mine in Northern Saskatchewan in a joint venture with Cogema Resources Inc. Reserve grade is in excess of 20 per cent U3O8 with current mining grade averaging 16 per cent U3O8. Due to the high radiation potential of the ore, the non-entry mining method of raiseboring was chosen to minimise worker radiation exposure potential. The raisebore and extraction horizons are located in waste rock with the orebody located between horizons. Ore produced by raiseboring is delivered by LHD to an underground SAG mill and is then hydraulically hoisted to surface receiving facilities after underground thickening to 50 per cent solids by weight. Concrete backfilling is utilised. The orebodies at McArthur River are adjacent to water bearing sandstone. Ground freezing is therefore employed to isolate the ore from this water bearing ground. Calcium chloride brine at -35¦C is circulated through a series of cased boreholes to establish freeze walls up to 25 metres in thickness. At 18 000 000 lb of low cost U3O8 production per annum, McArthur River ranks as the worldsÆ largest uranium mine. Over 50 per cent of the workforce resides in this sparsely populated area of Canada with approximately 40 per cent being of aboriginal ancestry. Site operations must cater for winter temperatures as low as -50¦C.

Jan 1, 2002