Search Documents

By Simon C. Dominy, Ian M. Platten, Garry F. Johansen, Alwyn E. Annels, Brian W. Cuffley

Abstract - Coarse gold-bearing veins are characterized by high grades that are localized and erratic. Effective sampling of coarse gold-bearing veins is difficult because of the low concentration and erratic nature of the gold particles. Diamond drilling is an effective measure of geological continuity, however, grade distribution can only be reliably obtained from underground development (including close-spaced sampling, bulk sampling, and trial mining). Comparison between surface and underground drilling, underground linear/panel and bulk sampling indicate that drilling and linear/ panel samples generally understate bulk sample grades. Bulk samples are likely to be the closest estimators of true grade. It is unlikely that anything above an Inferred Resource category can be estimated from surface drilling alone, and at best the grade will only be a global estimate. Underground development, in-fill drilling and bulk sampling/trial mining will be required to delineate Indicated and Measured Resources. Closely spaced development and bulk sampling is likely to be the only way to determine Reserves. The resource estimation process must driven by a clear geological model that should attempt to understand both geological and grade continuity. Traditionally, grade interpolation has been undertaken using classical methods, though more recently, computer-based block modelling techniques have been used with some success. Three case histories are presented, documenting the problems of estimating resources in coarse gold-bearing veins and the practical ways in which they were approached.

Jan 1, 2000

By Jr. Blanchette, Claude Bazin, Paul Blatter

"Reporting of the gold production during a month implies to estimate the mass of gold that is processed by a mill, recovered as dorés, lost to the tailings and the change of gold inventory within the plant equipments. Carbon in pulp and/or carbon in leach tanks may contain gold inventory as large as 72 000 g (2 200 troy oz) of gold which needs to be estimated from measurements obtained from sampling the material contained in a tank. Sampling and analysis introduce measurement errors that could be systematic and/or random. This paper examines the effect of measurement errors on the estimation of the gold inventory and on the variation of the inventory in the carbon in leach tanks of the Agnico-Eagle Mines Limited Lapa concentrator. Results of the analysis demonstrate that the inventory is measured with ±10% reproducibility, while the variation of inventory used for mass balance is known with ±300% reproducibility.INTRODUCTIONThe preparation of a production report implies to reconcile measured data with the constraint of material balance over the considered period of production. The sampling and data reconciliation problems were recently reviewed by Morrison (2008) and Jansen (2009a; 2009b). Typically a production report is prepared to cover a period over which the mass conservation hypothesis can be considered valid. In gold processing plants, the change of inventory can seldom be neglected and plays a significant role in material balance calculations. This is particularly true for circuits using carbon adsorption, where the complete cycle of carbon is measured in days.In a 2000 t/d gold ore processing plant, the inventory can be as large as 140 000 g of gold in gravity concentrator bins, carbon in pulp tanks and in the electrolysis circuit. This is equivalent to the mass of gold that fed the circuit during 10 days of operation with 7 g/t feed assays. The change of inventory may therefore strongly impact the production figures and especially the plant recovery at the end of a week or a month. Despite the financial impact associated with gold inventory, few publications are devoted to the subject. Bruington (1991) and Morrison (2008) discussed the problem and stressed the need for the development of reliable methods to estimate the gold inventory in carbon in pulp or carbon in leach circuits."

Jan 1, 2011

Unrelieved residual strains can influence the states of stress in rock masses. If cheap, stable structures are to be constructed in Tock in which large pressures are exterted by unrelieved residual strains, the pressures must be either utilized or neutralized. A photoelastic technique has been used to show that the release of residual strains in rocks is influenced by the mechanical properties of anisotropic rock fabrics. The relationships between directions of residual strain release and the attitudes of minor geological structures have been determined. For engineering design purposes it is possible to estimate the principal stress directions in rock containing unrelieved residual strains if the attitudes of such minor geological structures as foliation planes, lineations, and joints are known.INTRODUCTIONQuarry workers have long been aware that near-surface rocks may be highly stressed. In many cases the stresses are too large to be the result of gravitational loads. Dale (1923, pp. 34-35) gives many instances of deformation produced by the violent release of residual strains in quarry rocks. Coates (1963) has described powerful rock bursts and cases of rock deformation underground implying the existence of much larger stresses than are likely to result from present-day gravitational loads. The only explanation is that pressures of up to several thousands of pounds per square inch may be caused by unrelieved residual strains in rock.Instruments have been developed for measuring states of stress in rock masses. Most stress determinations in rock have been made on the assumption that rock behaves as a simple elastic material. It is now known that rock is a...

Jan 1, 1964

By P. A. Cornejo

Blastings always generates induced vibrations through seismic wave, these are lost energy from the blasting and therefore causing the disturbance of slopes and structures near the blast source. Most of the components of an open pit mine have to reach a physical stability, especially during its closure process. Aiming to develop a vibration prediction model using the historical records from an open pit mine located on Peru, this research has two main focuses depending on the distance from the blast source to the structure: to protect the final slopes and to protect the villagers houses. Both are based on the Devine model of the Square Root Scaled Distance (SRSD) to predict the Peak Particle Velocity (PPV). The Mid field model, where is found mainly final slopes from the pit, limited by the failure criteria of McKenzie based on the geo mechanical properties of the rock mass. The Far field model ranges distances where are located the most community households, limited by the German rule DIN 4150 for highly sensitive structures. Both scenarios resulted on a distribution abacus to limit the weight of used explosives to get the best operational advantage and structure safety.

Mar 2, 2022

By M. Hokkanen, M. Hagfors

The Finnish Defence Forces has destroyed old and spoilt munitions in open surface mass detonations at Hukkakero fell in municipality of Kittilä, Lapland, since 1988. The technique was learned from Sweden. The destruction camp is organized every year at autumn and it lasts 4-5 weeks. During the disposal period one or two charges are exploded per day. The gross weight of a typical charge varies from 150-250 tons including 15-25 tons of explosives. They mainly consist of TNT, but may also contain smaller amounts of amatol and hexotol. As a booster there have been used 2-3 tons of emulsion explosives.

Jan 1, 2015

By Y. Fu, C. Aldrich

This study investigates the use of dense convolutional neural networks to estimate bubble size distributions in flotation froths. 20 videographically captured images of froths from an industrial platinum flotation plant in South Africa were labelled semantically with an image annotation tool, LabelMe (https://github.com/wkentaro/labelme). As a result, pixels in the images could be categorized into three classes, namely ‘bubbles’ (class 0), ‘bubble boundaries’ (class 1) and ‘background’ (class 2). A U-net architecture with a ResNet34 backbone was trained to segment the images via classification of the pixels in the images. 5 of the 20 images were used to validate the performance of the neural network. Explicit inclusion of the bubble boundaries as a class proved to be significantly more effective than training the network with two classes only (i.e. bubbles and background). However, the thickness of the boundary layer regions were critical. If too thin, it was difficult to distinguish between some contiguous bubbles and if too thick, some fine bubbles may not be identified. The results were promising, although there is still significant scope for improvement, among other by making use of more images for training and validation. Keywords: Flotation Froths, Convolutional Neural Networks, Deep Learning, Bubble size Analysis

Jan 1, 2020

By G. J. Arseneau

Mineral resource statements are reported as tonnage and grades. While great care is placed on generating a reliable estimation of grade, the determination of bulk density for resource estimation is often overlooked. Often mineral resource statements are prepared using a simple average value of bulk density. While a simple average estimate of bulk density can be appropriate for some deposit types, an over simplistic estimation of bulk density can result in serious errors in tonnage estimates and consequently to an over or under estimation of metal content. The construction of a reliable bulk density model requires an adequate bulk density database in order to develop a good understanding of the variability of bulk density with grade and geology. A reliable bulk density database is also essential in determining the appropriate modeling technique to minimize the errors in mineral resource statements caused by incorrect bulk density determinations. This paper summarizes bulk density measuring techniques and draws on examples from various deposit types examining the basic principles of bulk density modeling and domain construction for each deposit type. Examples are presented of simple bulk density domains and more complex grade variable bulk density domains with a discussion of best approaches for minimizing errors in density estimation.

Aug 1, 2013

By B. Oraee

Accurate estimation of equipment costs is a key factor in feasibility study and evaluation of design alternatives of mining projects. In this paper, capital and operation costs of backhoe loaders are estimated using multiple linear regression (MLR), based on principle component analysis (PCA). These cost functions are consisted of five independent variables; bucket size, digging depth, dump height, weight and horse power. The MLR is conducted in two steps. At the first correlation between independent variables is omitted using PCA technique. Thereafter, MLR functions are established using selected significant PCs and total cost functions are constituted as functions of initial variables. At the end, accuracy of functions are evaluated using mean absolute error rate method.

Jan 1, 2011

Careful monitoring of the position of the cave front is essential for well-managed block cave mining. Seismic monitoring systems yield 3-D information about the location of fracturing ahead of the cave. However, because the a seismic gap between cave surface and seismogenic zone is of unknown extent, the cluster of micro-seismicity can only yield an upper bound for the cave height. The large velocity contrast between the fractured rock in the aseismic gap and the broken loose material within the cave can be used by an inversion procedure to find a few simple geometric parameters (such as height of the top of the cave) if a simple 3-D shape?for example a paraboloid?is assumed for the cave. Ray-tracing must be used as the large velocity contrasts cause seismic waves to bend around the cave, but the inversion takes the form of a minimization problem in only a few dimensions and so the computational time is feasible. This technique has been successfully tested on realistic synthetic data.

Jan 1, 2007

By F. Nakhaei, M. Le Roux, Q. P. Campbell, S. Botlhoko

The results of coal sink and float tests are plotted as washability curves. Analysis of these curves indicates the most effective method by which the coal can be cleaned. Sink and float experiments are samplingdependent, destructive, and use toxic chemicals. The present research introduces an alternative method based on a 3D imaging system called RhoVol, in which the density of individual particles is determined from mass and volume measurements. A quantitative error analysis was conducted on coal sample density measurements obtained by RhoVol, and the main factors influencing the errors identified and investigated. The results show that the drawback of this technique is related to its not being able to detect hidden concavities in the particle, resulting in an overestimation of particle volume. Furthermore, it is difficult to capture the impact of porosity and cracks on the volume of coal particles obtained from 3D silhouette images. This mismatch between the RhoVol data and sink and float test results has led to attempts to apply neural network (NN) and linear regression techniques to produce a fully reliable model for correction and estimation of coal density. The NN approach offers superior predictive capability over linear regression, and the estimated density distribution is in line with the sink and float analysis. The root mean square error in estimation of density using the NN model was less than 0.05 g/cm³.

Aug 2, 2022

By Lizbeth A. Escobar, Phol M. Llanos, Maritza Y. Avila

The present research develops the estimation of critical parameters such as the distances and times between holes to avoid failure due to dynamic pressure in electronic detonators based on the near-field ground vibration model of Holmberg & Persson and the hydrodynamic model of Neyman.

Feb 1, 2020

By Firat Atalay

Estimating the spatial distribution of ore grade is one of the most critical steps to proceed with an investment decision on a deposit. Kriging is the most widely used method to estimate the ore grade, but alternative techniques are being developed. Machine learning algorithms can be used as alternative methods to classical kriging. In this paper, the iron (Fe) grade of a deposit is estimated with the XGBoost algorithm, and the results are compared with kriging estimation results. For the estimations, samples collected from drillholes were used. Due to the different natures of the estimation methods, different steps were taken to perform the estimations. The results show that XGBoost produced higher-range estimates, which is a desired result in ore grade estimation, and the minimum and maximum of the estimates were lower and higher than the kriging estimates, respectively. However, like kriging, the estimation results were smoother than composites, and the variance of the XGBoost estimates was lower than the variance of composites. This means that even though estimation with XGBoost mitigates the smoothing effect, the estimation results suffer from smoothing like kriging.

Dec 1, 2024

By Kwangmin Kim, Junnhyeok Park

In the hard-rock mining industry, the generation of fines during blasting is an important issue because it has a significant effect on production processes. Although various approaches have been used to estimate fines generation in recent decades, as a practical matter, measuring the amount of fines is nearly impossible given the technical limitations and the scale of mine blasts. In this study, a dynamic PPV damagemodel was developed and applied to predict fines generation as a result of blasting. This modified version of the crushed zone model (CZM) considers dynamic rock properties and peak particle velocity (PPV), which represents the induced stress in blasting. Dynamic rock-fracturing tests were conducted using a split Hopkinson pressure bar (SHPB) to determine the parameters that represent the relationships between rock strength and strain rate as well as the relationships between fines amount and strain rate in the model. A case study was conducted that entailed determining the radius of the crushed zone for given blast designs and rock properties and then measuring the fines amount within this crushed zone. This case study indicated that the dynamic PPV model can provide a rough estimate of fines generation for given dynamic rock properties and blast design parameters and that this method has the potential to be a useful indicator, or index, of fines generation.

May 16, 2020

By M F. L Gibson

The issue of flyrock is very important in the operation of all mineral extraction sites that use blasting. Unlike the other environmental effects of blasting, ground vibrations and air overpressure, flyrock is (or should be) a rare event, with severe consequences which may include damage to property, injury or fatalities. These physical consequences are often matched by a loss in public confidence in the management of the operation, mistrust and bad press. In this work a deterministic analysis of flyrock was carried out using the equations of a particle in motion with drag. The solutions to these equations were used to reproduce the curves of Lundborg et al (1975) for the relationship between flyrock boulder size and maximum ejection distance. A probabilistic approach based on Monte Carlo simulation was built up with the analytical model using published data together with local case study material. A typical quarry blast situation was modelled and estimates of safe standoff distances were determined using risk assessment techniques.

Jan 1, 2001

By Jogeshwar P. Singh, Mohamed Ashour

The Soil-Foundation-Structure-Interaction (SFSI) problem involves Kinematic Soil- Foundation Interaction occurring during large (cyclic and permanent) ground deformations as well as Inertial Foundation-Structure Interaction occurring during shaking all of which take place while the soil and possibly structural properties degrade with time. The design of the deep foundations for liquefaction induced lateral spread loads include estimation of passive pressure applied by the liquefied material as well as the amount of anticipated displacement. Coupled analyses for SFSI problems that account for liquefaction induced lateral spread and their interaction with the deep foundation system can be performed using sophisticated numerical modeling techniques based on finite element and/or finite difference methods using computer programs such as FLAC or OPENSEES. Such complete SFSI analyses are very tedious and expensive and not feasible for most projects. In 2011, California Department of Transportation (Caltrans) released guidelines for foundation loading and deformation due to liquefaction induced lateral spreading. This simpler prescriptive approach combines empirical and analytical methods through decoupled analyses using computer programs such as LPILE and SLOPE-W together with liquefaction induced lateral spread estimates based on published empirical equations. Pile group effects and pile caps are accounted for using a concept of super-pile. In this paper, we discuss the solution of the same lateral spreading related SFSI problem using computer program CGI-DFSAP (CGI-Deep Foundation System Analysis Program) based on the strain wedge theory that can account for liquefaction, lateral spreading, pile groups, pile caps and nonlinear behavior of piles in a more coupled manner to analyze the foundation system under the lateral spread loadings. CGI-DFSAP can be used to analyze piles in sloping ground conditions as well as piles beneath the abutments near a free-face slope. We have included Strain Wedge Model (SWM) solutions using CGI-DFSAP of several field and laboratory case studies of lateral spreading related to pile foundations. Finally, we have also included the SWM solution of one of the Caltrans (2011) example problems.

Jan 1, 2015

By T. Wettainen

Malmberget town is located in northern Sweden, approximately 70 km north of the Arctic Circle. Parts of the town overlie more than 20 iron orebodies, consisting mainly of magnetite with smaller quantities of haematite. The mine is operated by the mining company LKAB. Mining started in the 17th century, but not until the railway to the coastal city of Luleå was completed in 1888 did large-scale production commence. Around 1920, mining proceeded underground and today sublevel caving is the only mining method used. Sublevel caving causes subsidence of the ground surface, and buildings and residential areas have been relocated due to the mining activities for more than 50 years. The number of seismic events accompanied by strong ground vibrations is now increasing. In 2008 the mine received a permit from the Environmental Court of Sweden to increase production to 20 Mt of crude ore per year. A prerequisite for the permit was that the mine conducts a number of investigations regarding the environmental impact on the residents of Malmberget. One of these investigations concerned how seismicity will change as production increases and what measures could be taken to reduce inconvenience to the town residents. Today the mine possesses an extensive seismic monitoring system with more than 180 underground and surface geophones. For this study, eleven seismically active volumes in Malmberget mine were identified, and for each of them, a yearly future maximum magnitude interval was estimated based on the current production plan. Relationships between historical seismic events and measured ground vibrations in the town of Malmberget were established, and future ground vibrations caused by expected seismic events were estimated using a probabilistic approach. The outcome was the number of intervals of expected ground vibration per year and per monitoring point. Possible measures to reduce inconvenience for the town residents include blast restrictions, sequencing, and possibly preconditioning. The ultimate long-term solution is an almost complete relocation of Malmberget town. This process has recently been formalized and LKAB is taking an active part in realizing this goal.

Jan 1, 2014

By Rameshwar Sah, Bharath Kumar Shanmugam, Govinda Raj Mandela, Marutiram Kaza, Harsha Vardhan, Harish Hanumanthappa

Iron ores obtained from different sources differ in their chemical and physical properties. These variations make the process of grinding a difficult task. The work carried out in this context focuses on three different samples of iron ore, viz., high silica high alumina, low silica high alumina, and low silica low alumina. The grinding process for all the three iron ores is carried out individually in Bond’s ball mill and the total retention time taken by each iron ore sample is calculated. The present investigation focuses on utilizing the calculated retention time of the iron ore as a standard grinding reference time to the laboratory ball mill for optimizing the grinding time of each ore. The desired P80 (150 μm) with an acceptable range of hematite liberation (> 75%) was obtained in the laboratory ball mill after reducing 6 min from the total retention time taken in the Bond ball mill.

By Zhiye Zhao, Jianping Sun, Zhipeng Xu

"The subsea rock cavern project is an underground storage facility. For cost-effective and safe excavation, the risk caused by the groundwater seepage into the rock cavern must be addressed. As one of the essential steps of the assessment work, the hydraulic conductivities of the fractured rock masses should be reliably determined based on the limited measured data and geological information from the few boreholes. On the basis of site investigation data and monitored water flow and water pressure data during excavation, this paper presents an estimation method for hydraulic conductivities of the rock masses containing water bearing zones.INTRODUCTIONIn the development of underground rock caverns/tunnels, the uncertainty of hydro-geological conditions plays a critical role in terms of construction cost and construction safety. Excessive groundwater seepage has been identified as one of the key safety factors which always lead to vast loss, such as the Seikan Tunnel in Japan which encountered sudden water inflow and then catastrophic face collapsing, tunnel flood during excavation (Xu et al., 2013). As a key factor in the risk management associated with the groundwater leakage into rock cavern, the hydraulic conductivity determination is very important for cost estimation, reliable water control and construction management.The case presented in this paper is an underground project consisted of a cluster of rock caverns/tunnels for storage of hydrocarbon products. These caverns are located in heterogeneous sedimentary bedrock, at a depth of 130 m beneath a basin, and the crowns of caverns are located at about 100 m below the sea bottom. The caverns/tunnels are designed at two levels, the water curtain tunnels are at the upper level, and oil storage caverns are at the lower level, as shown in Figure 1. A number of horizontal and vertical water injection holes drilled from the upper water curtain gallery tunnel will establish continuous horizontal and vertical water curtains for providing a stable pressure distribution around the caverns, to keep oil/gas inside the lower storage caverns. As the hydrostatic pressure of groundwater around the storage caverns is higher than the pressure of stored oil and gas, the stored oil and gas will not leak out of the cavern (Lu, 2010; Xu et al., 2015). Groundwater pressure gauges are installed in the water curtain gallery tunnels, and water from the operational and access tunnels is injected continuously into the curtains to maintain the pressure. Some monitoring data are recorded at different construction stages. This paper proposes suitable methods to predict the hydraulic conductivity in rock mass and to determine the hydraulic conductivity in the water bearing zones."

Jan 1, 2015

By Danell. R. E.

A drill monitoring system has been developed which measures performance characteristics of a rotary drill and relates this to the material strength and rotary specific energy. The drill monitor is computer based and is used to produce a profile of the relative in-situ material strength as a function of depth. This information has been used in determining the placement of explosive in selective loading of blastholes and estimating the amount of lump product in a particular section of an ore body.

Jan 1, 1989

By R. Benito

This paper describes the application of modern data-analysis tools, such as data mining, to determine a representative cost driver for haulage activity. This cost driver was subsequently used to trace incurred haul costs by bench for openpit optimization. For this purpose, large amount of previously unused cost and production data from an open pit mine was extracted, loaded, transformed and integrated into a data warehouse. Predictive modeling was performed using Microsoft Decision Trees algorithm to compute expected unit haul costs by bench for future mine expansions. Finally, a sensitivity analysis was carried out to determine the effect of two cost drivers (if any) for incremental haul costs in the final pit outline process.

Jan 1, 2008