Search Documents

By V. Sheikhzadeh, M. Saghaeian, R. Hejazi, A. Abazarpoor

Green pellet strength including drop number and green compression is highly depended on particles size and shape. The used grinding method, ball mill or high pressure grinding roll (HPGR), directly vary the particle size and shape. Considering mode of pellet crack, green pellet is composed of particle-binder-moisture system. Different type of forces existed in such a system to create a stable green pellet: solid particle interaction, magnetic, binding and cohesive forces in the interface of particle-moisture. The pellet failure can occur at solid-liquid interface and/or through liquid phase. Pellet crack occurs when crack passed through entirely the plane failure. In this research, the effect of particle size and shape of pellet feed on green pellet strength has been investigated. It was deduced that green pellet strength highly improved by increase of specific surface area. The drop number and green compression strength of pellets prepared by HPGR was higher than ball mill. It was found that the main reason responsible for this difference was the particle shape. Capillary force is responsible for part of pellet force and improves by decrease of particle size. Also more elongated shape of HPGR particles resulted in higher capillary force due to thinner interface layer between particles. The less circular shape of particles lead to improvement of capillary force as can be occurred for HPGR particles. The rougher shape of HPGR particles resulted in higher wettability of particles and better diffusion of water among particles. In fact the rougher particles means lower connection angle. Aspect ratio of particles was of great importance of crack growth because the required energy of pellet failure increases directly. The required energy for pellet failure increases by increment of particle aspect ratio. The relationship among crack failure force, roughness and aspect ratio has been modeled considering 3 shape descriptors.

Jan 1, 2018

By James Douglas

THOUGH liberality is not supposed to be a prominent trait of the Scottish character, Canada owes to a Scotchman, Sir Wm. Macdonald, more than to any other of its people, not only wise ideas, but pecuniary help towards extending education: and another Scotchman, in New York, has sumptuously housed under one roof three branches of the engineering fraternity, of which we are one. Having thus created us members of the same family-for the metaphorical meaning of house and kinship is identical-Mr. Carnegie expresses more emphatically than even he could in words, that, by affording facilities for closest intercourse, he in¬vites the mechanical, electrical and mining engineers to par¬ticipate in the freest interchange of idea and experience, and to correlate and combine the results of their studies and activi¬ties; and, being members of the same household, to banish re¬serve and secrecy. And now we, the members of the - American Institute of Mining Engineers, meet as though we were at home in a neighboring country. For whether we meet in Canada, in England, or elsewhere, the Institute is always received as though no political or geographical divisions separated its members from those of congenial associations in the land of its host. In truth, the title which we have assumed claims for the sphere of our activities the whole American continent, regardless of such trifling details as boundary-lines. The fact that we meet to communicate each other's experience, to discuss our difficulties, and to seek each other's aid in solving the intricate problems that so often present themselves in the course of our profes¬sional life, is an acknowledgment of our individual helpless¬ness; and therefore an argument for united effort. But no effort can be of any value if there is an underlying suspicion

Jan 9, 1907

By Cavallaro. J. A.

This report presents the results of a washability study of 455 raw coal channel samples with special emphasis on sulfur reduction. These raw coals contained on the average 14.0 percent ash, 1.91 percent pyritic sulfur, and 3.02 percent total sulfur. The complete washability data are presented for each sample processed. A statistical evaluation is included for the coalbeds from which more than 10 samples were collected and for the geographical coal producing regions. A graphical summation is presented for the coal producing regions and selected coalbeds. Finally a similar statistical evaluation is included showing composited data interpolated at Btu recovery levels of 50, 60, 70, 80, 90, and 100 percent. Graphical summations are presented for the various coal producing regions. Generally, significant sulfur reduction is achieved if the coals are crushed to a finer size and the higher specific gravity increments are removed. If a 50-percent Btu recovery was acceptable, then 32 percent of the samples tested could be upgraded to meet the current Environmental Protection Agency (EPA) standard of 1.2 pounds of S02 emission per million (MM) Btu when crushed to 14-mesh top size and separated gravimetrically.

Jan 1, 1976

By C. E. Smith, R. J. Inglis, M. I. Brittan, J. H. Schoeman

"AbstractA number of plant-scale cyanidation tests were carried out in an attempt to uncover the causes of fluctuations in gold recovery efficiency. Computer analysis of the results verified that the treatment of reclaimed waste rock material along with normal run-of-mine ore was having an adverse effect on metallurgical efficiency. Mineralogical examinations showed that the free gold particles in the dump material were heavily tarnished or coated (predominantly by hydrated iron oxide), accounting for the lower recoveries. The indications were that the effect could be counteracted somewhat by using higher than normal cyanide and lime reagent concentrations and adequate leaching times. Analysis of the results showed further that leach pulp density was an important operating parameter affecting extraction efficiency. A variable activation energy leaching model provided an extremely good characterization of individual batch leach runs.IntroductionFLUCTUATING METALLURGICAL EFFICIENCY at the gold concentrator of the S.A. Land and Exploration Company Limited (""Sallies"") prompted some plant-scale cyanide leach studies in an endeavour to uncover the causes of the variations in gold recovery."

Jan 1, 1977

By T. Sterry Hunt

WET processes for the extraction of copper from its ores have of late attracted much attention, especially in Europe, where the use of oupriferous iron-pyrites as a' source of sulphur prevails. These processes present an additional interest, when, as is often the case, the copper-ore's to be treated contain one or both of the precious metals, since the complete separation of these from copper, either in the dry or the moist way, is, as is well known, one of the difficult problems of metallurgy. The principal wet processes hitherto in use for the extraction of copper from its ores, may be included under three classes : I. Those in which the copper in snlphuretted ores is rendered soluble in water by calcining them, after a preliminary roasting which removes the greater part of their sulphur, with a portion of common salt, by which the copper is converted into a chloride (Long-maid and Henderson), or with sulphate of soda, by which sulphate of copper is formed (Monnier). Allied to these is the method by which a portion of the copper is got as soluble sulphate by roasting without addition, a process which is sometimes modified and rendered more complete by re-roasting the lixiviated residues with the addition of a portion of raw sulphuretted ore (Bankart). II. Those methods in which free chlorhydric or sulphuric acid is used to dissolve the copper- from oxidized or roasted ores. These, while simple and efficient, are too costly, except in certain localities where chlorhydric acid is a waste product. Related to them is the plan which consists in exposing the oxidized and moistened ores to the slow action of sulphurous acid mixed with air, by which a soluble sulphate of copper is formed. It has been proposed to decompose the solutions of sulphate or chloride of copper by sulphuretted hydrogen, thus separating the copper as sulphide, and using the liberated acid to dissolve fresh portions of oxidized copper. It is, however, the almost universal practice to throw down the copper from its solutions, however obtained, in the metallic state by the use of iron, either scrap-iron or iron-sponge; although in some instances

Jan 1, 1882

By H. G. Schneider

The Smackover oil and gas field lies in Ouachita and Union Counties, Ark., in the south-central part of the state, in T.15 and 16S., R.15, 16, and 17W. It is 10 miles north of El Dorado, the principal business and railroad center in the south-central part of the state, and extends from the town of Smackover in all directions. Productive territory having an aggregate area of nearly 55 sq. mi. has been developed. It extends 6 miles from north to south and 12 miles from east to west. The field may be divided into two parts: the east district, which includes the area east of the town of Smackover, and the west district, which includes the area west of the town of Smackover. In relation to the neighboring productive fields of Arkansas, the Smackover field is 9 miles north of the El Dorado oil and gas field, 8 miles slightly west of north of the East El Dorado oil and gas field, and 16 miles east of the Stephens oil field. Smackover was discovered in May, 1922, when the Oil Operators Tnist drilled well No. 1 on the J. T. Murphy farm, in the southeast quarter of sec. 8, T.16S., R.15W. This well is located almost on the top of the Norphlet dome and was drilled to the Nacatoch sand. It had an initial open flow volume of 30,000,000 cu. ft. of dry gas per day, and a rock pressure of 950 lb. per sq. in. In July, 1922, the V. K. F. Co. completed the first oil well in sec. 29, T.l5S., R.l5W., on the north side of the Norphlet dome, and 3 miles north of the first gas well. This well started the active development of the north and northwest sides of the dome.] The field was extended westwards, when oil was discovered in the Nacatoch sand, during October, 1922, in sec. 4, T.l6S., R.l6W., 51/2 miles west of the original gas well. This opened up the so-called "light oil district," which is located on a structural terrace, for which the name

Jan 1, 1924

By P Klaric

Drive stability is critical in block/panel cave mining. The production levels of these mines are required to remain operational for the entire life-of-mine which can be in excess of 30 years. It is therefore necessary to produce high-quality drives to ensure sustainable mine productivity. In order to improve the quality of drives and pillars, this gold mine instigated a project to optimise their development blasting. The initial aim was to reduce the level of overbreak in development mining, which was averaging approximately 18 per cent. The benefits of controlling overbreak not only included an increase in the integrity of the drives but also a decrease in load and haul and ground support costs. Through focusing on the optimal blasting of the perimeter holes, overbreak was reduced to 4.5 per cent. The project’s focus then moved to drill and charge designs for development faces. In optimising these designs, it was anticipated that improvements in the development mining cycle would be achievable.CITATION:Klaric, P, 2015. Improving drive stability through efficient development blasting design and practices, in Proceedings 11th International Symposium on Rock Fragmentation by Blasting, pp 327–332 (The Australasian Institute of Mining and Metallurgy: Melbourne).

Aug 24, 2015

By J M. Nel

Richards Bay Minerals (RBM) has recently undertaken an exercise to improve orebody basal definition. Older mineralisation bases (pre-2005) were highly erratic using old protocol reverse circulation (RC) (also known as Air Core) drilling and interpretation. The erratic nature of this estimation tool, resulted in impractical modelling towards the end of mineralisation (EOM), difficult mine design and unrealistic reconciliation. New protocol drilling, which includes high definition SONIC drilling data, has provided more data to be available for more accurate interpretation. Recent advances in drilling methodology has allowed for better penetration depths for SONIC drilling, which was previously restricted to approximately 80 m (pre-2005). This improved data set allows more emphasis on geological interpretation with factors such as lithology, slime content, and mineral species distribution. These factors when in unique combination enable more accurate and realistic EOM data points. These points are modelled into surfaces using conditional simulation techniques and result in a vastly smoother surface and consequently, a more representative resource.

Oct 5, 2011

By D. K. Deardorff, Laurance L. Oden, H. Kato, M. I. Copeland

The equilibrium relationship of hafnium and tantalum was studied to facilitate alloy and property investigations . The data obtained are presented graphically as the equilibrium diagram . Alloys were prepared from high - purity components by nonconsumabletungsten- electrode arc melting . Data for the equilibrium diagram were obtained by optical pyrometry , metallography , X- ray diffraction of heat treated specimens , X- ray diffraction using a high- temperature camera , and differential thermal analysis . During the investigations , a casting apparatus was constructed , and the existing melting - point apparatus was improved . The bcc allotrope of hafnium forms a continuous series of solid solutions , called beta , with tantalum which exhibit a minimum melting point at 20 wt pct tantalum and 2,110 ° C. At 24 wt pct tantalum and 1,050 ° C , hafnium- rich beta decomposes in a monotectoid reaction to yield alpha and tantalum- rich beta . The monotectoid isotherm extends from 5.5 to 60 wt pct tantalum . The twophase region , in which hafnium- rich and tantalum- rich beta solid solutions are immiscible , rises to a maximum temperature of 1,265 ° C at 38 wt pct tantalum . The solubility of hafnium in tantalum decreases from 40 wt pct at the monotectoid temperature to 20 wt pct at 750 ° C.

Jan 1, 1964

By I. I. Inculet, M. A. Bergougnou, J. D. Brown, R. M. Quigley, D. K. Faurschou

"Coal from Hat Creek, B. C., has been successfully beneficiated to remove ash while retaining calorific value by a dry electrostatic separation process using a fluidized bed for triboelectrification. Two samples of Hat Creek coal used in the beneficiation experiments have been characterized particularly with respect to their mineral and clay content. The calorific value recovery exceeds the simple dry coal recovery because of the heat lost in altering the clay minerals of the ash fraction during burning. Recoveries and ash contents of the beneficiated coal are comparable to recoveries by water washing, but the dry process avoids the potential water pollution problems.IntroductionProposals have been made which suggest an interim Canadian coal production target for the end of the century which will require a four- to fivefold increase over the 1977 production. This implies a nominal annual growth rate of about 6070 to achieve a production rate of 115 to 140 million tonnes/year. Whether or not this target is realistic, it is becoming increasingly evident that coal must be utilized more fully on a world-wide basis. The crisis in Iran has provided the latest stimulus to renewed interest in Canadian coals for domestic use and export.The emerging new era of coal utilization for fuel and chemical feedstock will not, however, generally tolerate the level of coal technology which prevailed prior to the availability of abundant cheap oil and natural gas. Economic, environmental and resource management considerations will necessitate, for example, the use of improved methods of beneficiating coal. Continued development and application of conventional processes, principally aqueous, involving heavy media and cyclones will solve many of the problems. However, less conventional processes may be preferable or necessary in certain applications."

Jan 1, 1980

By Soner Top

In this study, a process to separate manganese and iron from manganiferous iron ores by reductive acid leaching followed by magnetic separation was conceived and experimentally tested. In the leaching process, sulfuric acid was used as lixiviant and oxalic acid was used as reductant. The experimental results showed that the manganese and iron separation was optimum when the concentration of the sulfuric acid and oxalic acid were 0.75 M and 30 g/L, respectively, at a temperature of 80 °C, a solid/ liquid ratio of 67 g/L, stirring speed of 400 rpm, and leaching duration of 60 min. Under this condition, 90.49% and 6.78% of Mn and Fe were dissolved, respectively, from the ore sample with a size fraction of − 106 μm. It was determined that the leaching of manganese from the ores was a second-order reaction with an activation energy (Ea) of 53.38 kJ/mol. The leaching residues obtained under the optimum condition were subjected to high-intensity wet magnetic separation tests to recover the remaining iron content. This separation process produced a concentrate containing 56.20% Fe and 1.79% Mn with iron and manganese recoveries of 56.83% and 66.73%, respectively. A magnetic separation test from an unleached ore sample was also carried out as a benchmark. To the best of our knowledge, this is the first time that a magnetic separation process was used to a residue obtained from reductive acid leaching of manganiferous iron ores to recover iron.

By Paul Weir

PRODUCTION METHODS in the bituminous coal mines in the United States are undergoing many changes. Although the primary object of these changes is the production of a better product at a cheaper cost there are other considerations such as the improvement of conditions under which miners live and work. The almost complete unionization of the manual workers in the industry during the past four years has 'brought with it a stabilization of wages on a relatively high basis and of hours of work per day and per week on a relatively low basis. With differentials in wages between producing districts firmly established, and with a nation- wide and uniform application of in-

Jan 1, 1937

By R. Casagrande, L. R. Moore, J. Sessoms, P. Macy

"A South American nickel mine has been producing a nickel concentrate that also contains copper and iron (15 percent nickel, 5 percent copper, 28 percent iron) for the end use of stainless steel production. A critical specification for the quality of the steel is the iron to magnesium oxide ratio (Fe:MgO), with a ratio of below 3.5:1 having a substantial impact on the nickel value. This nickel concentrate producer historically mined the sulfur-rich mineral pyroxenite but had been forced to move to a magnesium oxide/silicate (MgO/SiO4)-rich mineral harzburgite, resulting in nickel recovery being reduced to around 50 percent to maintain the required Fe:MgO ratio. In this study, ore mineralogy, water chemistry and the chemistries associated with the flotation process were evaluated in an effort to increase nickel recovery while maintaining the Fe:MgO ratio according to specification. New chemistry was introduced with the specific goal of complexing with the MgO and retarding its interaction with the bubble flow in the flotation process.IntroductionGlobal nickel production in 2011 totalled to 1,770,000 tons, according to the Raw Materials Group (2013). From 1950 to 2005 there was an aggregate 4.4 percent increase in nickel production, with a market-value high of $1,000/ton nickel reached in 1998 (Mudd, 2010). Approximately 65 percent of nickel production is consumed by the stainless steel industry, with 12 percent of the remainder used in super alloys or nonferrous alloys. Nickel imparts corrosion resistance, temperature resistance and other benefits to its alloys (Table 1) (Oberg, 1996; Cáceres, 2003). The dominant types of nickel-bearing ores are classified as either laterites (limonite: nickel-iron oxide; and garnierite: nickel hydrosilicate) or magmatic sulfide deposits (pentlandite: iron-nickel sulfide) (U.S. Geological Survey, 2013a, 2013b). In 2012, about 40 percent of the nickel ore mined was sulfidic-type ores and about 60 percent was laterite. The processing of laterite ore is expected to continue to increase as the availability of sulfidic ores continues to decline."

Jan 1, 2015

By Frederick Laist, H. J. Maguire

The ore from the Butte mines of the Anaconda company is quite siliceous; that is, it contains considerably less iron than is needed for the fluxing of the silica. The direct smelting of this ore, therefore, requires the addition of considerable fluxing material. The only flux available in sufficient quantity at Anaconda is limestone containing no metallic minerals. This necessarily makes the direct smelting of the ore so expensive that it is more profitable to subject the ore to a preliminary treatment by concentration. However, as the losses in the concentrator were very high, compared with the losses in the blast furnaccs, it was found, by calculation, more profitable to send ores containing more than a given percentage of copper to the blast furnaces, in spite of the higher costs of treatment. The dividing point was dependent on various factors, the chief of which was the prevailing market price of copper. It was customary, therefore, to segregate the ore as mined into two classes. Ore in the first class generally contained 5 per cent. or over of copper; this was sent directly to the blast-furnace department. Second-class ore contained less than 5 per cent. of copper, and averaged around 3 per cent.; it went to the concentrating department. With the introduction of the flotation process, it mas found possible to maintain an average recovery, as concentrates, of 96 per cent. At the same time, the cost of reverberatory smelting was very materially decreased by the use of pulverized coal. These two changes in practice made the blast furnaces unable to compete with the reverberatory furnaces on a cost basis. The distinction between first- and second-class ore was therefore abandoned, and the ore was treated by concentration. Obviously, on ores as high in iron as the Butte ores, the concentrator produces a product that is self-fluxing; in fact, it is possible to produce a concentrate containing more iron than is necessary to flux the silica. Under these conditions, the amount of basic fluxing materials required for the smelting operations became very srnall and the converter slag lost its former value as a blast-furnace flux. For this reason, a less expensive method of recovering the copper, silver, and gold contained in it than smelting in blast furnaces had to be found. It was not feasible

Jan 1, 1921

By W. T. Robson

LAKE Shore Mines, Limited, with property at Kirkland. Lake, Ontario, was incorporated m 1914 and commenced product10n m March, 1918, with an initial capacity of about 60 tons per day. Since that time a gradual expansion has taken place until a present capacity of approximately 2,300 tons per day has been reached, giving a total annual gold production valued at approximately $16,000,000. At the time of incorporation of the Company, the Lake Shore property consisted of four mineral claims and four fractions, totalling 171.05 acres. In 1934, four claims belonging to the Minaker Kirkland Gold Mines, Limited, were acquired and a controlling interest purchased in the Kirkland Hudson Bay Gold Mines, Limited. Both these properties adjoin the Lake Shore claims on the south. From the accompanying compilation of the world's largest gold mines for 1934 (1), it will be seen that, while Lake Shore holds sixth position in respect to the number of ounces produced, there are twenty-five gold mines which treat a greater tonnage of ore. This is an indication of probably the most unusual feature of the Kirkland Lake deposits, namely, the high aver-age tenor of the ores. coupled with the comparatively large tonnages available.

Jan 1, 1936

By N. L. Wimmler

The Metaline mining district was examined in 1942 by Bureau of Mines engineers with the cooperation of geologists of the Federal Geological Survey. Because the demands of war production were rapidly depleting reserves in some of the older zinc-lead producing districts of the United States, it was decided that an attempt should be made to increase the proved or indicated ore reserves of the Metaline district. With this objective, certain areas to be explored by diamond drilling were selected. The Bureau of Mines began work on this program. January 15, 1943; work terminated September 8, 1945. Forty-five holes were drilled in seven areas, The holes ranged from 154 feet to 2,050 feet in length. Their combined length was 38,150 feet. Thirty-three of the holes were drilled in. the North Pend Oreille and the West Bella May areas.

Jan 1, 1949

By S. J. Broderick

[The discovery in 1934 that carbonaceous materials, such as coal, peat, and lignite, when treated with concentrated sulfuric acid, were given cation or bate-exchange properties opened up a field of new materials with many interesting possibilities. These hydrogen exchangers, as they were called, were so superior to the impractical hydrogen exchangers made from natural and synthetic zeolites tat they were accepted immediately by the water-conditioning industry. A noteworthy advantage of the carbonaceous exchangers is that they are nonsiliceous, hence they do not give up silica to the water as do the natural and synthetic Zeolites, especially if the hard water is corrosive. For this reason the new exchangers found immediate favor in boiler-feed-water conditioning.]

Jan 1, 1941

By R Carstens, L Munsamy, L Prinsloo, C Huang, C Zhang, R Liang, I Canbulat

With recent advances in machine learning, data-based automated decision-making has been proven successful with promising outcomes in many industries. Its application in the mining sector can be ground-breaking; particularly the data-driven analytics that can support decision-making in the mine and consequently enhance mine safety, efficiency and sustainability. However, the current data science applications in the mining industry are mostly noncomprehensive, especially inadequate in evaluations, thus often requiring further study to be potentially beneficial to operations. Furthermore, it was found that existing literature paid limited attention to understanding the role of such techniques. Hence, in this paper we aim to build an end-to-end machine learning powered data-driven framework for intelligent geo-hazard analytics to support decision-making for underground mining data management, where multiple types of data are managed in an integrated database to support the unified machine-learning model. Here we present the framework as data-driven workflow processes, including data gathering, data preparation, data processing and data presentation. In addition, we present an application scenario of the proposed framework, where falls of ground (FoG) is managed with the historical FoG investigation reports, related panel hazard ratings, compliance data and rock mass characteristics. A demonstration can be dynamic forecasting of high-risk areas for FoG based on both spatial and temporal factors, including a 3D data visualisation. Such a scenario shows the potential of the proposed framework to establish connections among various data fields across a diverse category of data. Furthermore, the results offer important insights for ground instability management and can potentially optimise the underground operation flow.

Nov 29, 2022

By Matt Fenwick, Douglas R. Schwarm

Stringent underwater noise limits at the Fairview Ave. N. bridge replacement on Seattle's Lake Union required an innovative approach to supporting the 470-ft long drill trestle. Screw-in foundations would reliably mitigate noise, but the required 500-kip bearing resistance were an order of magnitude higher than typical helical piles. This paper describes the process of designing, installing, monitoring, and testing 22-inch dia. helical piles 80 feet long with 36- and 42-inch dia. helices. A static load test confirmed that the conventional k-factor design method can be extended to capacities of up to 800 kips. A 375 ft-kip drive mechanism had to be fabricated because no North American equipment could supply the necessary installation torque. Real time installation monitoring and later "re-torque" tests confirm that helical piles gain capacity with time, analogous to driven piles under restrike. Underwater and on-land measurements confirmed very low noise, suggesting the possibility of widespread adoption where similar limits are imposed. Overall, this case history illuminates the need for innovative deep foundation types in response to ever-increasing regulatory constraints.

Sep 8, 2021

By Henry Fay

The erosion of guns is a complex problem which can be solved only by a detailed study of all the factors involved. In the present paper it is proposed to submit the results of observations and experiments which have extended over several years. Certain facts have been established and preliminary reports have been published.' Conclusions have been reached since then which, it is believed, are worth presenting for the purpose of promoting discussion, and with the hope that ultimately the whole truth will be known. The particular phase of the problem which will be presented here is the hardening of the inner surface of the gun tube. It has been known for a long time that after firing a large caliber gun for some time, the surface of the metal becomes hard and brittle, cracks, and wears away.2 The character of this phenomenon may be learned by an examination of some results which are typical, obtained from a detailed study of a 12-in. gun. The gun was trepanned and rings representing various sections of the gun were cut out for examination. The first ring represented the metal at the muzzle end, and the other rings were removed at approximately 120, 240, 325 and 337 in., respectively, toward the breech end. The appearance of the surface of the metal in the different sections is shown in Fig. 1. The greatest amount of wear and maximum amount of hard surface layer was found in section marked E, which is nearest the powder chamber and at the beginning of the rifing; the maximum amount of cracking in D; the maximum amount of copper deposition in C. Section B showed heat cracks on both lands and grooves, and section A showed heat cracks on the driving edge of the lands only. There was much flow of metal in both sections E and D, and there was progressively less effect noticeable toward the muzzle end. The polished cross-section of these same pieces is shown in Fig. 2. The extent of the erosion is greatest nearest the powder chamber, and diminishes toward the muzzle end.

Jan 1, 1917