Search Documents

By K. Adham

Iron chloride pyrohydrolysis (ICP) is used as an efficient way to fix iron as an oxide, for sale as a by-product or for safe disposal. Iron chloride is reacted with water at elevated temperatures to yield iron oxide and hydrogen chloride. Both the steel industry and synthetic rutile producers make extensive use of ICP. Two technologies are available for ICP: fluidized bed (FB) and spray roaster (SR). Both technologies are energy intensive. Here, fluid beds and spray roasters are compared with respect to their fuel and electrical energy consumptions. Roaster size and oxide product characteristics are also considered. Process models are prepared, using the METSIM® software, and specific energy consumptions are compared. The results will help the industry decide which technology best suits its goals and at what energy cost.

Jan 1, 2006

By D. J. Archambeault, A. Saigal

Substantial cost savings have been demonstrated by material design changes after initial product development. As more information about a product's operating environment becomes available to a designer, it is possible to consider a more costeffective material that still meets or exceeds the design intent, such as minimum low cycle fatigue life. This is especially true in the aircraft engine industry, where precise operating cycle data is not available until after the first engine is built and tested. A methodology for predicting estimated low cycle fatigue life has been developed to investigate potential material change candidates for an aft cooling plate. A twodimensional macromechanical finite element model (FEM) has been developed using ANSYS and used to determine the effects of varying material properties. It is shown that slight changes in density, coefficient of thermal expansion and modulus of elasticity have profound effects on component and effective stresses, and therefore, estimated low cycle fatigue (ELCF) life. Aerospace

Jan 1, 2006

Abstract-This paper discusses the fundamental question, "what is an adequate sample size?" in the light of recent developments regarding parameters of Gy’s well-known formula for predicting sampling variance. New understanding of the granulometric factor, g, and the liberation factor, l, had significant impact on the results of calculations involving Gy’s equation. For the granulometric factor, justifications are given as to why nominal comminution size should always be reported as P95 (i.e., the screen size rejecting 95% of the mass). A small but important change in the model to estimate the liberation factor makes the difference between a predictive formula that works or does not. Other parameters and considerations regarding their practical implementation are commented upon in light of the author’s experience.

Jan 1, 1998

By A. L. Mackie

Mine water from base metal mines is generally characterized as having high concentrations of heavy metals and low pH, resulting in the need for treatment prior to discharge to receiving waters. In this study, two high-rate clarification processes were evaluated at bench-scale: the high density sludge (HDS) process and sand-ballasted flocculation. High-rate clarification processes generate larger, denser flocs which settle more quickly than those formed in conventional flocculation processes. The HDS process uses recirculated sludge in order to seed floc formation, whereas ballasted flocculation uses microsand. Both of these high-rate processes require the addition of quicklime or hydrated lime to increase pH and allow for the formation of insoluble metal hydroxides, which can then be removed in the clarification stage. Previous research has shown that cement kiln dust (CKD) can effectively replace quicklime in conventional sedimentation processes for metals and acidity reduction to meet discharge guidelines (Mackie & Walsh 2012). CKD is an alkaline by-product of cement manufacture that has fine particle sizes and varying free lime concentrations depending on the manufacturing process. The lower solubility (i.e., free lime content) of CKD requires larger doses of material to be added to reach treatment pH targets for metals precipitation. The purpose of this study was to evaluate the treatment efficacy of high-rate clarification processes for mine water treatment, compared with conventional sedimentation, at bench-scale with conventional (i.e., lime) and novel (i.e., CKD) pH adjustment additives. Results of the study showed that all treatment processes (i.e., HDS, ballasted flocculation, and conventional sedimentation) were able to reduce target metal concentrations to well below Canadian guidelines (i.e., total zinc and total nickel < 0.5 mg/L, total copper < 0.3 mg/L) using either CKD or lime. Additionally, mine water samples treated with CKD generated lower sludge volumes than those treated with lime (i.e., 50 versus 100 mL for conventional sedimentation and 100 versus 200 mL for HDS) even with increased concentrations of solids introduced during treatment. This is most likely due to slower neutralization rates when using CKD, due to its lower solubility, as well as finer particle size. The sand-ballasted flocculation process was able to reduce final turbidity in CKD-treated samples while the HDS process did not when compared to conventional sedimentation, making the ballasted flocculation process preferable for reducing final solids concentrations in CKD-treated mine water.

Aug 1, 2013

By K. Thuro, B. Sellmeier, M. Krautblatter

"Recent rockfall events, like the Tramin rockfall at January 21st, 2014, underline the significance of a comprehensive hazard assessment. A common way of estimating the hazard along infrastructure is the performance of run-out modelling using 3D rockfall codes like Rockyfor3D or RAMMS::Rockfall (Dorren, 2010, Bartelt et al., 2013). The potential block volumes are often roughly estimated from talus material and an overview of the source area, where the magnitude assessment is usually physically and intellectually demanding. Nevertheless especially for increased magnitudes, the volume estimation is crucial in terms of hazard analysis. Mid-Magnitude-Events exceed the load capacity of the most mitigation measures and their date of failure is hard to predict. Here we define the term Mid-Magnitude-Events as block volumes between boulder falls (10-100 m³) and block falls (100-10,000 m³) (Whalley, 1984).The current contribution provides an approach of integral process analysis for a 200 m3 carbonate block above the federal road B 305 near the city of Salzburg in the Bavarian Alps, Germany. We performed detailed analysis of the critical carbonate block exposed to planar failure. The block volume and the dimensions of the failure surface were mapped in detail, due to the unique occasion of an accessible failure surface. The discontinuity persistence was recorded to assess a minimal degree of fragmentation for a potential failure scenario. A maximum degree of fragmentation was estimated due to recording of block dimensions at the talus slope resulting from frequent rockfalls.We performed parameter studies in terms of run-out modelling using the code RAMMS::Rockfall (Bartelt et al., 2013), taking the recorded data for block fragmentation into account. In our contribution, we aim to demonstrate a comprehensive approach for hazard assessment of mid magnitude rockfalls, taking detachment processes, fragmentation and run-out analysis into account."

Jan 1, 2015

By Atulya Verma, Samantha Espley, Nashel Kalaka

Numerical modelling simulations of proposed mining sequences are used to evaluate mining induced stress concentrations and their migration associated with the progress of mining. This analysis is a standard practice for most prospective mining projects at various levels of feasibility study. The studies are usually initiated at the beginning of the mining cycle, on a global scale, and are maintained and updated throughout the life of mine (LoM) to identify or address local issues and the required remedies as and when needed. Developments in numerical modeling techniques along with advancements in computing technology and data storage/management capacity have given rock engineers access to a huge volume of data that needs to be analyzed and assessed. The most common approach to numerically evaluate stress concentrations and migrations in underground mines is to select the points of interest in a given volume such as the back of an excavation or points within pillars around critical infrastructure, and plot the stresses or a derivative of stresses over time to represent life of mine concentrations and migrations of stresses. The problem with this approach is the sharp gradient of stresses close to excavation boundaries. Slight changes in the location of selected points can change the stress concentration and migration patterns substantially and all generalized statements about LoM risk can be challenged in one way or another. For global stresses, the most common current approach of data analysis is by viewing series of pictures representing stress contours. In this approach, it is easy to overlook some critical detail and/or be masked by personal preferences of contour/picture representation. To overcome the problem of representing and analyzing huge volume of stress simulation data, the authors, through their experience at different mining operations, developed a volume-based technique to reliably evaluate mining-induced stress concentrations and migrations. The volume-based approach is for risk-based assessments of adjacent regions in the rockmass, in and around planned stoping excavations. The volume of interest is the entire ore body that is planned to be mined, but it could be just a part of the ore body depending on the specific use case. The risk identifiers are the induced stresses as compared to the strength of the rockmass. Over the LoM, the risk identifiers are tracked versus the corresponding excavation ratio of the zone. This approach is very effective in identifying the stages of mining that are contributing the most to the mining induced stress risk and the results can be used for proactive planning and mitigations measures such as alternative mining sequences, or stope sizes, or extraction methods. The approach has also been very successful in application at mature mines, particularly in assessing the re-initiation of mining in zone that has been dormant for some time. The volume-based approach allows for an assessment of the current state of the rockmass and the identification of future risks when planning resumption of mining activities.

Jan 1, 2023

By Baseer Abdul, Edouard Asselin

Hydrometallurgical leaching of sulfide concentrates of copper and nickel at medium temperature (150°C) produces residues that contain sulfur and iron bearing minerals and phases. During leaching, and depending on various process parameters, iron may be precipitated as hematite, goethite, jarosite or other oxyhydroxides, which may be more or less crystalline. Hematite is the favored precipitate because it is the most environmentally stable and does not ad/absorb as much copper, nickel or other solution constituents during precipitation. However, the low solubility of iron during the medium temperature processing of sulfide ores can favour the formation of poorly crystalline, nano-scale iron oxide/oxyhydroxide phases. In some cases these phases have been positively identified as the metastable ferrihydrite, which transforms into iron oxides such as goethite, hematite and magnetite over time. A better understanding of what may help drive this transformation during leaching would ultimately result in lower valuable metal losses and more stable leach residues. Higher acid concentrations result in increased copper extractions and favor the formation of hematite during concentrate leaching, rather than other metastable phases. Furthermore, commercially available water displacement formula “WD40®” and other novel reagent(s) affect Fe precipitation and sulfur chemistry, leading to very different process outcomes such as improved extractions and larger, more easily separated, sulfur particles.

Jan 1, 2017

By G. E. Rodger, R. Jugoon

"SUMMARYSamples of Aunor gold precipitates which varied widely in gold content were fluxed to produce a bisilicate slag with various oxidizers and combinations of oxidizers. On the basis of some preliminary tests a flux composition was chosen which appeared to show the most promise.The different components of this flux were then incorporated into a matrix for a 23 factorial design experiment to determine what were the most important variables and what changes should be made to the flux composition to optimize the resultant total bullion fineness.The results of the 23 factorial design experiment showed that the flux composition as calculated was very close to an optimum as only very minor changes were required to give bullion in excess of 995 fine.INTRODUCTIONEarly in 1969 the results from the Aunor Gold Mines refinery showed that wide variations existed both in precipitate grade and in total bullion fineness. An analysis of this data showed that as the precipitate grade increased, the gold bullion fineness also increased (see appendix 1). Therefore, to improve the refinery results test work was initiated to: 1) improve the precipitate grade and 2) to determine, with the help of statistics, a flux composition which would be able to handle wide variations in precipitate grade and give a total bullion fineness in excess of 950 fine."

Jan 1, 1970

By P. J. van Staden

Several differences exist between the columns used for laboratory simulations of heap leaching and commercial heaps. Column walls support part of the weight of the ore and restrict the particle packing arrangements. Columns are typically of a diameter that is smaller than the dripper spacing to be used on the commercial heaps. The stacking practices that cause segregation and stratification on heaps are absent from columns. Intuitively, these physical differences would be expected to be the reasons behind the observed differences between column and heap leaching performance. But how can they be quantified, and what is the magnitude of the differences? Could some of these differences be built into column leaching experiments in order to better simulate commercial-scale heap leaching performance to ease the scale-up calculations from column leaching data? Alternatively, could heap construction practices be improved so that the ideal performance obtained in columns could be emulated on heaps? Data from a few heap leaching cases studies that permit direct comparisons between column and heap leaching performance is analysed, to reveal the extent of the differences in resistance to mass transfer. Discussion is provided on the implications for column and heap leaching practice, and for scale-up design procedures. INTRODUCTION Differences have been known to exist between leaching performance observed in laboratory columns and that of commercial-scale heaps. Both the heap leaching kinetics and finally achievable extent of extraction are generally found to be inferior in heaps compared to that achieved in columns. Heap leaching design engineers extrapolate the anticipated heap leaching performance by multiplying the time-axis of a column-leaching curve by a factor of about 2 and the extraction-axis by a factor of about 0.8, as explained by Jansen and Taylor (2002), Scheffel (2002), John (2011) and Scheffel, Guzmán, and Dreier (2016). The procedure is illustrated in Figure 1 by means of a hypothetical example. The solid line represents the leaching performance observed in a laboratory column, and the dashed line represents the anticipated leaching performance of a heap, extrapolated from the column leaching data using these empirical scale-up factors.

Jan 1, 2019

By T. Nagasaka, T. Miki

Suppression of CO2 and waste such as slags discharged from iron- and steelmaking processes are some of the typical major issues for the protection of global environment and sustainable growth of steelmaking industry. Utilization of active phytoplankton growth will be one of the best options to stabilize and suppress carbon dioxide at high-efficiency. Inorganic minerals such as C, 0, N, Si, P and Fe are necessary for phytoplankton multiplication. It is crucial for supply of nutrition into seawater effectively for phytoplankton multiplication to understand the dissolution behavior of some elements from steelmaking slags into seawater. The morphology of the precipitated phases in steelmaking slags during cooling period was investigated in the present work. The dissolution behavior of some elements from steelmaking slags into artificial seawater has also been studied. The effect of nutrient dissolved into seawater on phytoplankton multiplication, some examples of marine structure made from steelmaking slag, and estimation of CO2 fixation amount by marine phytoplankton are also discussed.

Jan 1, 2004

By Hung-Ming Chen

This paper presents a Cloud-based system framework for providing a web-based building information model (BIM) service. Cloud computing technologies were used to develop a BIM data center, which can handle the big data of massive BIMs using multiple servers in a distributed manner. In addition, the system allows multiple users to access the BIM data center online to submit and review BIMs concurrently. In this study, the advanced technologies/tools of web development and Cloud computing were applied and incorporated to develop a prototype system, named CloudBIM. On the client side, this study applied WebGL and HTML5 in the web-based interface development to achieve the 3D display of BIMs and the dynamic interaction with viewers through browsers. Users can access the services of CloudBIM by using PCs, laptops and tablets anytime anywhere to view the 3D model via browsers, which support WebGL. On the server side, this study used Apache Hadoop, an open-sourced distributed file system based on Google?s technologies, to establish a BIM data center. The Hadoop Distributed File System (HDFS) can utilize multiple servers to provide mass storage spaces in a distributed manner. Based on the use of Apache HBase, which provides Bigtable-like structured storage for HDFS, this study proposes database schema for storing the information from massive BIMs in two bigtables, named Object table and Property table respectively. In addition, Hadoop MapReduce, a Hadoop component for the parallel processing of large data sets, is utilized for model data acquisition by parsing the IFC files of uploaded models in parallel.

Aug 1, 2013

By A. M. Hadian

The use of dolomite refractories in steel making and cement industries has advantages over other basic products particularly in countries where this type of refractories is produced. Due to existence of high-quality dolomite deposits in many regions of Iran, there is a good potential for producing dolomite refractories in the country. This study focuses on the effect of such additives as iron oxide and alumina on hydration resistance and density of Iranian dolomite. It was found that addition of Fe2O3 is more effective than Al2O3 on increasing density as well as resistance to hydration of dolomite.

Jan 1, 2004

Over the years, efficient on-line monitoring of performance and status of an electrolytic copper refining process has been challenging due to the need for complex electrical measurement systems and the management of the enormous volume of data. Recent innovative developments by Kennecott Utah Copper Corporation and Outokumpu Technology in wireless cell voltage and individual cell temperature sensing opens new possibilities to improve and modernize tank house management. This paper describes the powerful Outokumpu CellSenseTM refinery monitoring system for efficient interpretation of data produced by wireless CellSensorTM system with traditional refinery measurements. Examples from commercial tank house installations demonstrate the benefits of the system to improved process control and operation. The Outokumpu CellSenseTM monitoring system is simple to operate and accessible to all the plant operation teams.

Jan 1, 2007

By M. Silva, G. Olivares

In a leach pile, the drainage layer or commonly known as the cover layer has been a matter of discussion due to the multiple phenomena or events that occur within it which impacts negatively the leaching process. All these events are directly related to the loss of draining capacity and irregular flow net due to an excessive compaction and entrapment of fines reducing the flow velocity which in turn cause more fines to get trapped. Over time the problem will continue to worsen rapidly and accelerating continuously like a geometric progression. When the cover layer gets to a certain level of deterioration, the only alternative to continue operating safely is to replace the entire cover material, a huge cost for any mining operation. TESRA has developed a solution to counteract the accelerated advance of the deterioration of the drainage layer. The solution consists of using large agricultural tractors coupled to a proprietary design of a heavy- duty subsoiler-ripper for deep tillage for loosening and breaking up the ground at various depth and with different legs spacing from 0.4 m or above. Also, if the ground still requires additional breaking, TESRA proposes the use of a rototiller that breaks any lumps to homogenize the ground to a certain particle size. With the subsoiler-ripper/rototiller combination equipment a layer of 0.5 m deep is properly treated to match closely the original conditions of the ground thus regaining the flow net as close as to the original seepage condition. Once the ripping is completed the subsoiler-ripper tool is replaced with another that is similar to a plow and thus continue with the installation of drainage pipes. The tool digs a Y-type profile trench allowing the pipe to sit precisely at the bottom of the trench, contributing additional lateral and lower support. This tool controls the installation depth of the drainage pipe at different inclination grades and also eliminating any irregularities that may block the flow of the leaching solution. Sometimes the obstructions block completely the flow of the solution. The main result of this methodology is the reduction of groundwater levels and at least 2% less humidity in the leaching gravel due to an increase in the draining capacity of cover. The methodology developed by TESRA applied during a couple of cycles of a dynamic leach pile has proven to be a real alternative to remedy and recuperate the cover close to its original conditions. The methodology requires only one large tractor and two accessories operated in different stages while replacing the more traditional and less efficient machinery such as the Bulldozer, Motor-grader, and backhoes.

Jan 1, 2019

By N. W. Xu, C. Sha, F. Dai, Y. G. Zhu, B. Li

Underground caverns stability and excavatability of rock masses are of significance in geotechnical engineering, in both design and construction stages. Risks associated with underground group caverns of Houziyan hydropower station in southwestern China are growing as a result of continuous excavation-induced unloading. To improve our understanding of the underground caverns instability and resolve the complex subsurface conditions of the highly fractured rock mass, a high resolution microseismic monitoring system was established in the deep underground caverns, pursuing the relationship between the measured data about microseismic activity and excavation damage zones of surrounding rock mass. Through analyzing the tempo-spatial distribution of microseismic activity, excavation damage zones and potential risk regions in underground caverns were identified. Furthermore, the correlation between microseismic activity and pre-existing geological structures, as well as traditional monitoring was analyzed. In order to validate the correlation between seismicity and potential slip surface of underground caverns, a model was implemented into numerical code to further evaluate the deformation and stability of surrounding rock mass. The monitoring results demonstrate that microseismic events mainly occur at high stress concentration regions, which coincides with the results obtained from numerical analysis. Therefore, the comprehensive method incorporating microseismic monitoring and numerical analysis, as well as traditional monitoring and filed observation has been proven to be very promising in deformation and even instability prediction of surrounding rock mass in underground caverns subject to delineation of potential slip surfaces and excavation damage zones.

Jan 1, 2015

By L. Tapia

The main disturbance for stability in primary grinding using SAG mills is related to continuous changes in the ore being fed to the plant, such as size and hardness. Searching for stability is the first step to accomplish, before trying to maximize the mill throughput. The time variability in the characteristics of the ore being fed from the mine to the plant must be known or inferred to perform adaptive changes in the grinding process control. In day to day operation, such information is not available in real time. Transport times and inventories between the mine and the plant complicate the situation. Under this scenario, Minera Los Pelambres implemented a grinding optimization strategy, based on the real time analysis of certain process variables, such as ore size, mill current consumption, pebble generation and throughput. As a result, both grindability and stability indexes are inferred. This is equivalent to inferring the ore characteristics by knowing its effect on the grinding behavior. These two indexes permit control of the SAG mill total charge level and hence, allow optimization of throughput by means of an expert control system.

Jan 1, 2007

By T. T. Chen

A detailed mineralogical study was carried out to characterize the products generated in a conventional copper refinery anode slimes treatment circuit. The anode slimes, which are first leached in sulphuric acid media under an oxygen pressure to dissolve the copper and much of the tellurium, were found to contain various Ag-Cu selenides, PbSO4, Ag, Au and BaSO4. The resulting Ag2Se, Ag3AuSe2, Se, Au, PbSO4, BaSO4, SnO2 and silicate-rich leach residue is dried and pelletized. It is subsequently roasted to convert the selenium to Ag2SeO3, volatile SeO2 and water soluble Na2SeO4. The roasted product, which was found to consist of Ag2Se, PbSe, PbSO4, BaSO4, Ag3AuSe2, Au and Sn02, is smelted and refined in a Dore furnace to produce a Dore metal that contains a (Ag,Au,Pd,Cu,Pt) alloy together with traces of CuO and a Cu-Pb oxide phase. Parting plant mud shown to contain AgC1, Ag, Au, (Au,Ag,Pd), Ag2Cu2O3 and Ag20, and scrubber mud containing PbSO4 and AgC1 are also fed to the Dore furnace. The Dore furnace sequentially separates a silica slag which was found to consist of Ag, Sn02, BaSO4, Pb-Sb oxide, Fe-Sb-Ni-Cr oxide and a Pb-rich silicate phase, a soda-rich slag containing Na2CO3.H2O and Na2(Se,Te)03, and a niter slag containing Na2TeO3, Na2(Se,Te)03 and a Na-Te-Se-Pb-Ba-Cu oxide phase. The Dore metal is electrorefined in Moebius cells to yield high-purity silver and a gold mud which was shown to consist of a (Au,Ag,Pd,Cu,Pt) alloy and traces of CuO. Silver is leached from the gold mud and the resulting gold sand is cast into anodes and electrorefined in Wohlwill cells to produce pure gold.

Jan 1, 2007

By C. W. Chao, M. Melashvili, F. Hilário

The processing of a bulk concentrate containing copper as chalcocite, enargite, bornite and chalcopyrite along with pyrite and precious metals of silver and gold was studied. A leaching study was undertaken to examine the application of the atmospheric acid leaching process to a bulk copper concentrate for copper recovery followed by cyanidation or thiosulfate leaching for gold and silver recovery. The extraction of copper exceeded 90% under some conditions and the recovery of gold and silver reached a maximum of 44 and 45%, respectively. A sample of bulk concentrate was roasted in air to oxidize the sulfides with release of sulfur dioxide followed by acid leaching of the roaster calcine and cyanidation of the acid leach solution. The copper extraction reached over 92% and the gold extraction exceeded 88%. The sulfur dioxide may be captured for acid production. INTRODUCTION The processing of a large primary copper deposit in Indonesia is under investigation. One of the possible routes for processing involves flotation of a bulk copper-gold concentrate followed by treatment by an atmospheric leach/SX-EW process for copper recovery. The atmospheric leach residue would then be leached for recovery of gold and silver. A second route involves roasting of the concentrate followed by acid leaching and SX-EW for copper recovery. The leach residue from the roaster calcine would then be leached for recovery of gold and silver. The roaster off-gases could be used for sulfuric acid manufacture, enabling the production of a valuable by-product. The Cobre Las Cruces (CLC) process (Munoz, Garcia, & Delgado, 2017) was designed to treat a resource of copper discovered close to Seville, Spain. The process was developed by Outotec with MK Gold (original project owner). INMET Mining subsequently purchased the deposit from MK Gold and built the copper processing facility. The CLC process is similar to the Mt. Gordon Copper process; except that OKTOP reactors have been installed instead of a low temperature, low pressure autoclave process. The plant was designed to treat 1.3 mtpa of ore at a grade of 6.2% Cu to produce 72,000 tpa of cathode copper over a 15 year project life. The main gangue mineral at Cobre Las Cruces was pyrite. The Cobre Las Cruces ore is therefore similar to a rougher flotation concentrate containing copper minerals and pyrite. The roasting of a rougher flotation concentrate can be designed to produce hematite, basic copper oxides and sulfur dioxide off gas. The acid leaching of the basic copper oxides may produce soluble copper for SX-EW recovery of copper. The sulfur dioxide off-gas may be used to make sulfuric acid.

Jan 1, 2019

By K. Torigoe, S. Owad, C. Tokoro, Y Tsunazawa

The mechanism of an abrasion process to recover metal values by removing components from printed circuit boards (PCBs) is modeled using a discrete element method (DEM) simulation. To simulate the abrasion process, a model PCB geometry was constructed consisting of bonded particulates. The simulation allowed connections between particles to be broken if the bonding force calculated from DEM exceeded a threshold level. Simulation results were compared with experimental comminution data using an agitation mill with test PCBs on which nine capacitors were surface solder-mounted. Simulation results obtained from the DEM with the anchoring model corresponded to experimental observations.

Jan 1, 2011

By W. T. Gotheridge

The principal sources of coal-dust underground are, of course, coal, and the working of coal. Coal-dust is most dangerous when it is in the finest state of division. This class of dust is found mostly on the roads, being blown or shaken off the loaded trams in transit, or carried in suspension by air currents. The air current passing from the working faces may also carry dust in suspension into the return airway, and the air current from the surface may carry dust from the shakerscreens into the intake airways. The mosl dangerous dust we have to contend with is, (a) that which is most finely divided, and (b) that which contains the most volatile matter, including the inflammable occluded gases, and also oxygen absorbed from the air. According to experiments, coal-dust in a fine and dry state would intensify an explosion, in air free from fire-damp carrying in excess of 3 ozs. of fine dust or 6 ozs. of coarse dust per cubic yard. The volatile matter in the dust must be between 12 and 31 per cent. in the ash-free coal, the latter percentage giving the highest flame velocity. It is vary difficult to propagate flame where coal-dust contains a large percentage of stone-dust, (say 50 per cent.). Shale-dust or other incombustible dust containing no volatile hydro-carbons or carbonaceous matter will prevent the extension of an explosion if mixed with coal-dust in proportion of say 70 per cent., and if the dust is finely divided and dry.

Jan 1, 1924