Search Documents

By Graeme Lamson, Rene Dobbe, Merle Kalstrom, Janice Baron, Robert Thorpe, Chris Martin, Snezana Bajic, Amanda Palko

"This paper presents early results from a case study at the Mount Polley porphyry copper mine in British Columbia, Canada. Daily mineralogical data was acquired and reconciled with daily plant assay and recovery balances to interrogate the mineralogical controls on plant performance. After commissioning, the Minesite machine operated smoothly and reliably, creating daily data to generate mineralogical balances around the circuit for a period of several weeks. The key data, on final products were produced within 13 hours of the end of the respective operating period. The resulting information allowed for a greater understanding of the opportunities and limitations of the Mount Polley metallurgy. In this paper, how this information can be used to help long term strategic planning for the operation, as well as mine planning and production forecasting and the fine tuning of day to day mill operations is described. Early data available at the time of preparing this paper indicated that roughly 40% of the sulphide losses, or 32% of total copper losses existed as low grade inclusions in +75 micron tails. This represented the dominant form of long term copper losses, demonstrated repeatedly in the routine analytical data. However, on a day-to-day basis, the amount of copper present as chrysocolla is the major driver behind variations in total copper recovery. The degree to which this influences mill recovery is valuable to the geologist or mine planner. As they are armed with mineral speciation information for future ores, this will allow for improved mine planning, prediction of copper recoveries and production forecasting. The data was also started to find value in the mill by showing what caused sulphide losses. In the roughers, for example, the data showed poor recovery was the result of poor flotation performance and not any changes the speciation of copper sulphides or their liberation characteristics."

Jan 1, 2014

By D. J. Zhao

The reduction and nitridation of preoxidized ilmenite pellets were investigated with the prospect of creating grain refiners containing titanium oxides and titanium nitride dispersoids. All the reduction and nitridation reactions were conducted in a furnace with a thermogravimetric analysis system under a H2-N2 gas mixture of 5 L/min. The ilmenite pellets were reduced at 1100°C for 4 hours under 80%H2-20%N2 gas mixture, and then respectively nitrided at 700°C, 800°C, 900°C and 1000°C for 12 hours under 10%H2-90%N2 gas mixture. The products were characterized by X-ray diffraction (XRD) and Scanning Electron Microscopy (SEM), and analyzed by Energy Dispersive Spectrometers (EDS) and Electron Probe Mapping Analysis (EPMA). Iron oxides were reduced into metallic iron, and oxides of titanium were found, but titanium nitride particles were not identified in the products under the prevailing conditions.

Jan 1, 2014

By G. Bonifazi

Mining activities belong to human history, they are lost in the mists of time. They usually impact on large portions of the territory, both directly (exploitation areas and logistic activities directly related to exploitation) and indirectly (exploited material handling, processing and storage areas). As a result, at the end of mine life cycle, dismissed areas usually cannot be directly re-utilized without efficient and well addressed rehabilitation actions. Main goal of the proposed approach was to analyze the results obtained when a set of procedures, based on the collection and the analysis of remote sensed data, is developed to evaluate the effect of rehabilitation of land contaminated by extractive industry activities and to perform a monitoring and registration of the environmental impact of such operations by the application and the integration of modern information technologies, as Earth Observation (EO) and Geographic Information Systems (GIS). The study was developed in the framework of the EU-Asia program of the European union (EXTRACT-IT Project-Contract No: CN/ASIA IT&C/006 (89870): Application of Information Technologies for the sustainable management of Extractive industry activities) and it was addressed to three different dismissed mine areas in India, Thailand and China, respectively. The results, in terms of methodologies selected, strategies applied and results obtained, are described with particular reference to Fuxin coal mine area in China.

Jan 1, 2014

Hydrofluoric acid (HF) is traditionally used in plant operation to separate feldspars from quartz. It acts during the conditioning step as a pH regulator and an activating agent providing an excellent selectivity. However, a corrosive chemical environment and the ecological impacts from the potential fluoride ion contamination are limiting factors to use this chemical agent at an industrial scale. Investigations of the amine-feldspar interactions in the absence and presence of non-linear long-chain alcohols were carried out on two ore samples. The presence of alcohol molecules led to form a closed packed layer only on the surface of feldspars. The highest order and packing at the feldspar surface were observed with a particular alkyl chain structure as Guerbert alcohols. A new flotation process was developed and tested at both lab and pilot scales on different feldspar ores. The Guerbert alcohols allow to perform a feldspar concentrate with very low content in quartz (2.3-3.5 %) and feldspars recovery from 78 to 81 % while the aliphatic isoalcohols show a more higher recovery level (88-89 %). New HF-free process is strongly dependent on the pH range (1.7-2.5) and sulphuric acid consumption during conditioning steps. Optimal amine/alcohol weight ratio is ranged between 0.8 and 1.0. The values below 0.8 decrease strongly the flotation selectivity and the quartz recovery within the feldspar concentrates is higher than that occurs with HF.

Sep 1, 2012

"The most common comminution testing methods and the associated analytical procedures provide average values and do not capture the variability in breakage behaviour of ore particles. Therefore, plant designs and process performance predictions use only average values; hence, the inherent variability that exists within an ore domain translates into variation in process performance. This paper discusses the drawbacks of using the average approach to impact comminution testing and plant design. It considers breakage intensity from a single-particle testing perspective. To clearly delineate breakage behaviour of each broken particle, an experimental technique using the JK Drop Weight Tester (JKDWT) has been developed by which actual comminution energy per particle and breakage intensities were measured directly. Four different ore types in ranges of -45+37.5mm and -22.4+19mm were tested at three energy levels of 0.1, 0.25, 0.5 kWh/t and six energy levels of 0.5, 1.0, 2.0, 2.5, 3.0 and 4.0 kWh/t, respectively. Sieved results of individual particles, notionally, t5, t10 and t50 showed the presence of large variability in their breakage behaviour, which was not captured by the standard JK impact breakage testing approach. In addition, the original JKMRC model is modified by introducing a third parameter, ‘Emint10’ which describes the minimum energy required to produce the minimum percent t10 for a given ore type and its significance is tested statistically. These results suggest that the analysis of breakage variability may lead to more robust plant designs. Perhaps, the most useful insight from this work is to extend single particle breakage model developed by JKMRC to a comminution percentile envelope which may provide a more detailed picture of ore breakage variability. INTRODUCTION Rock consists of mineral grains, pores, voids and cracks that may differ markedly in size/density which make rock a heterogeneous material by nature. The degree of heterogeneity in rock affects its breakage characteristics where these variations increase local differential strain and stress within a stressed rock (Scholz, 1968) and it depends on many factors such as rock genesis, alterations, etc. One important aspect of this heterogeneous feature is presence of a large amount of experimental scatter in rock breakage results when considered from a quantitative perspective (Bieniawski, 1967; Liu, Roquete, Kou, & Lindqvist, 2004)."

Jan 1, 2018

By Y. P. Lu

PET/kaolin nanocomposites were prepared with melt blending and the structure of the composites was characterized through IR techniques. The results showed that there is stable chemical bonding between PET and kaolin. In addition, the dispersion condition of nanokaolin was observed by means of SEM and TEM. As a result, the thickness of the dispersed nanokaolin particles in PET matrix is mainly from 20mm to 200nm, and the nanosized particles congregate to some extent in the samples. In addition, it was found the kaolin particles almost homogeneously dispersed in the PET matrix, without preferential orientation because of the shear effect due to melt blending. Furthermore, the mechanical properties of PET/kaolin nanocomposites are dependent on the modification efficiency of nanokaolin. With the poor efficiency, the nanokaolin particles tend to aggregate at a high content and have a bad interfacial compability between nanokaolin and PET matrix, which lead to stress concentration, so the mechanical property descends. Experiments of different contents of modified nanokalin indicated that the intensity of PET/kaolin nanocomposites decreased firstly, then increased, and then decreased, with the nanokaolin content of 4 to 5 percent, the rupture stress of the composites showed a maximum value, which was higher than that of pure PET. On one hand, nanokaolin decreases the function among PET macromolecules, which will decline the mechanical properties, on the other hand, the piece structure of kaolinite improves the intensity of the nanocomposite. When the nanokaolin content is low, the former has the dominant effect on mechanical property, while the latter will play a leading role with the content increasing. But if the nanokaolin content still increases, the agglomeration and stress concentration will be obvious.

Jan 1, 2014

In this paper, the recovery of copper minerals and molybdenite in the rougher/scavenger banks of two separate mineral processing plants is compared with recovery in laboratory (batch) cells. A number of metallurgical surveys were conducted in two plants treating porphyry copper and molyb-denum ore. Batch flotation tests were performed on plant conditioned flotation feed slurry, in parallel with the plant metallurgical surveys. The effect of flotation feed density on recovery of copper minerals and molybdenite in the plant and batch laboratory cells was investigated. Changes in flotation pulp density had little effect on copper mineral recovery, in both plant and laboratory cells. The ultimate recovery for molybdenite, on the contrary, increased considerably when the pulp density was reduced, both at plant and laboratory scale. Recovery of copper minerals in the plant and laboratory cells was comparable, while molybdenite recovery in laboratory flotation cells was 3%~13% higher than in plant. It is hypothesised that the flotation rate of molybdenite is sensitive to cell hydrodynamic conditions (i.e. local turbulent energy dissipation) more than the flotation rate of copper. The flotation behaviour of molybdenite particles is likely to be predominantly controlled by bubble-particle collision frequency, which is enhanced in high local turbulent energy dissipation environment. For chalcopyrite, the flotation recovery is mostly determined by attachment and stability efficiencies. An increase in local turbulent energy dissipation, while increasing collision efficiency, may negatively impact on attachment and stability efficiencies. The outcomes may have a considerable impact on increasing molybdenite recovery from porphyry ores.

Jan 1, 2014

By L. S. Zhou

Germany-made AFX-100 multi-grade fluidized classifier, with exquisite built-in structure and unique novel design, materializes not only advanced and efficient classification, but also precise particle distribution and safe operation, in addition to lower energy consumption and operational cost. It, simultaneously in place of screw grader, water cyclone, and fine mesh sieve, is equipped with multi-function, and indicates superior advantage in comparison with other else grading facilities. As the third-generation equipment, which is efficient and modern, it, in China and therefore all over the world, will find wide application in processing both fine and ultra-fine iron ore resources. Industrial test, carried out in Benxi Steel?s Waitoushan Iron Mine, shows efficient classification, i.e., ?74µm particle contents in coarse, medium and fine products come to 6.09%,11.82% and 82.73%, respectively, while qualitative and quantitative efficiencies end up by 74.38% and 87.66%, respectively. Its separation efficiencies amount to 2.2 times as many as spiral classifiers, 1.8 times as hydrocyclones, and 1.3 times as vibration high-frequency fine screens. Another industrial test, which was practiced in Anshan Steel?s Donganshan Sinter Plant, indicates that classification qualitative efficiency comes to 65.06%, equal to 1.51 times as many as hydrocyclones; yield, 55.79%; -71µm particle contents in medium and coarse products, 42.02%; -71µm content in fine product, 95.42%; and concentration, 12.33%. Estimated separation indexes are as follows: Yield for concentrate at 31.22%, up 0.22% than that from general processing line; concentrate grade at 66.42%, up 1.13%; recovery ratio comes to 65.84%, up 1.28%, which, based on pilot study, results in 20 million yuan of economic return for Donganshan Floatation Plant.

Jan 1, 2014

By A. Bahrami

Andalusite, silimanite and kyanite are polymorphs of aluminosilicate known as kyanite group. Minerals of this group are used in refractory. Sarband andalusite deposit is located in Markazi province in Iran. Mineralogical studies showed that the main minerals of this sample are andalusite (5%~10 %), mica, quartz and feldspar. Processing of the samples was carried out using two sizes, i.e. +500 µm and -500 µm. In the processing of +500 µm scrabble test was carried out on -4750 +500 µm and -2380 +500 µm. For -500+300 µm shaking table and magnetic separation were applied. For -300 µm fraction, flotation was carried out. The results of this tests showed that the processing of Shazand andalusite was not possible, because silica and iron impurities were diffused inside the andalusite mineral grains.

Jan 1, 2014

By F. M. Zafar Kaifi

Magnesite producing companies rely on high grade purity MgO. The mineralogy of magnesite has a significant impact on the manufacturing of magnesite bricks. The manufacturing process of bricks companies utilizes 90% magnesite from 85%~90% MgO. The main objective of the magnesite bricks was to utilizes the indigenous raw material available in bulk quantity. With better under-standing of chemistry of the mineral species and strict control of bricks manufacturing/processing conditions can be better produce efficiently, a low cost high quality bricks with minimum detrimental environmental impact.

Jan 1, 2014

By Ö. Genç

In this paper, full-scale mill data collected from three different circuit configurations in cement plants demonstrated the size reduction performance of dry ball mills in relation with the ball size classification achieved by the classifying liners used in the grinding compartments. Findings indicated that, breakage rates at the last 2 ~ 4 meters of the mill are not sufficient to give reasonable size reduction ratios for the production of fines. Application of only damming rings without classifying liners resulted in inverse ball size classification which lead to reduction in size reduction performance thus specific breakage rates.

Jan 1, 2014

By Jun Xiao, Jianwen Yang, Yanhong Dong, Daixiong Chen

"MANGOTE Surface oxidizied mine is a gold-bearing deposit in Congo's Mai-Ndombe Province. In the past decades, the ore has been exploited in a crude way by local farmers, which directly causes the severe processing difficulty and lack of economy of cyanide leaching. Hunan Research Institute for Non-ferrous metals (HRIM) is currently investigating options to recover gold through understanding the mineralogical characteristics of MANGOTE surface oxidized ore. Chemical analysis , X-ray diffraction(XRD) and Quantitative Evaluation of Minerals by SCANing electron microscopy (QEMSCAN ) were conducted on samples from mine site. The bulk analysis results showed that gold separation of gold from gangue minerals was feasible after grinding the ore samples to 86.53% -74µm and adjusting the pulp specific weight at 100g/L, and gravity separation was conducted with using spiral chute after grading of fine particles. The research results highlight relationship between mineralogical characteristics of abandoned ore and economy of comprehensive utilization that might be directive experience to similar cases. INTRODUCTION The occurrence state of gold in natural gold minerals determines the basic recovery and recovery effect of gold(Komnitsas, C., & Pooley, F. D. 1989). The occurrence of gold in the ore refers to the form of gold in the ore.Generally the existing state of gold can be divided into four major categories according to the viewpoint of the dispersion system: independent minerals and gold, a coarse dispersed system;colloidal gold, a colloidal dispersion; solid solution gold, is atomic ion dispersion; adsorptive gold, according to the adsorbed particle is the micelle or ion, respectively is a colloidal dispersion or ion dispersion(Mao, S. D., & Yang, R. S. 2009; Liu, J. J. et al. 2007; Horisberger, M., & Rosset, J. 1977; Li, Y., & Audétat, A. 2013).With the increasing technology of comprehensive utilization of resources,varied plants can now process different gold ores with various occurrence states(Neesse, T. 2014; 7. Ogundare. et al.2014)."

Jan 1, 2018

By Å Sandström

Loussavaara-KiirunavaaraAB?s (LKAB) iron ore mine, Kiirunavaara, is of magnetite type and associated with apatite, a calcium phosphate mineral. Rare earth elements (REE) have been found in connection with the apatite but also in other minerals. An apatite concentrate was produced by flotation, in order to investigate the possibilities to extract REE from the apatite by hydrometallurgical means. The content of rare earth elements calculated as oxides (REO) was 0.44% in the concentrate. Initially, leaching was done with large excess of sulfuric acid, hydrochloric acid, and nitric acid to study the REE solubility with different acids. In sulfuric acid, the leaching recoveries of the different REE varied between 35% and 69%. The reasons for the relative low recoveries, were that the rare earth elements precipitated together with huge amounts of gypsum, that was formed during leaching. However, both hydrochloric- and nitric acid gave almost complete extraction of the REE into solution. Analysis of the small leaching residue obtained, revealed, that mainly the light REE remained in the residue indicating that a small part of the light REE were present in the concentrate, in some acid, insoluble mineral other than apatite. With the aim to separate the REE from the phosphoric acid, the solutions from hydrochloric- and nitric acid leaching were neutralized to a pH of approximately 2 with ammonia. During neutralization, REE were precipitated as phosphates with good recoveries and a concentrate of 10.5% REO suitable for further upgrading was obtained, while the remaining solution can be used for fertilizer production. In further solvent extraction studies, it was shown, that after re-dissolution the REE could be effectively extracted by Cyanex 923, thus allowing for purification from calcium and phosphate ions. The REE were finally stripped from the organic phase and a REE nitrate solution was produced. Keywords: apatite, rare earth elements, REE, leaching, solvent extraction, cyanex 923

Sep 1, 2012

By D. R. Nagaraj

The traditional (popular) approach to reagent development and optimization, which has become entrenched in both research and practice for many decades, is highly reductionistic, incremental and costly, does not produce robust chemical solutions, and is not sustainable. There are also several other key drivers and changes in the mineral industry that clearly warrant a new approach. In response to these, an intense effort was made at Cytec over the past fifteen years to develop a meaningful, efficient, and sustainable approach to develop new reagents and robust chemical solutions in mineral separations. This effort culminated in three important outcomes: a) development of the holistic approach, called FLOTATION MATRIX 100? process, b) novel reagents, and c) enhanced application knowledge for reagents. In this paper, the efficiency, utility and power of the holistic approach will be demonstrated with examples using both existing and new reagents in the processing of Cu-Au, Au and Cu-Mo ores.

Jan 1, 2014

By A. Deepatana

This study compares the use of an aminophosphonate chelating resin Purolite S950 and a sulphonic acid strong cation exchange resin Purolite C160H to recover nickel and cobalt from the synthetic bioleach solutions. Batch adsorption of metal citrate is performed as a function of metal citrate concentration (15~2000 mg/L) and solution pH (0.01M and 0.1M citric acid). The adsorption mechanism as inferred from the fit to Langmuir (LM), Freundlich (FM) and Redlich-Peterson (RP) models demonstrates that nickel citrate adsorbed onto C160H appears monolayer in comparison to that adsorbed onto S950 exists multilayer. In contrast, cobalt citrate adsorption onto both C160H and S950 agrees with the LM model. The results have shown that loading capacities of nickel and cobalt citrate onto C160H are 60 and 65 mg/g whereas those observed onto S950 are 16mg/g and 6 mg/g respectively. These are achieved in metal citrate prepared in 0.1 M citric acid. Resin regeneration is conducted in using 2 M nitric acid solution. The results show high efficiency of nickel and cobalt citrate elution from Purolite C160H. Findings from this study suggest Purolite C160H is a suitable adsorbent for recovering nickel and cobalt from bioleaching solutions.

Jan 1, 2014

By V. V. Lodeischikov

This paper covers some problems of treatment of complex gold ores and concentrates containing copper, lead, zinc, antimony and uranium. All these metals may present in ores both as major and associated components. The paper suggests recommendations for early determination of relative value of the metals in ores, which largely contribute to the selection of concentration and metallurgical flow sheets. The focus of this work is on treatment of complex ores in gold plants. The metal extraction technologies used by such gold plants are typically aimed at maximum gold recovery. Therefore, production of base metals as by-products requires significant updating of the existing process flow sheets based on economic, environmental and other considerations. At the same time, cooperation between gold plants and base-metal plants is of great importance. This paper provides an overview of worldwide existing and innovative processes of base metals extraction from complex and refractory gold ores.

Jan 1, 2014

By P. Coghill, P. Revell

"Geo-sensing is the ability to routinely measure or infer rock properties at a range of scales using non-destructive scanning devices. These devices can provide information and data flows to support integrated decision making across the minerals value chain. Geo-sensing has been used for many years in minerals exploration using technologies such as satellite imaging and airborne geophysics but to date the use of sensing technologies in between the mine and mill has been very limited. The aim of this research is to develop new mineral sensing technologies that can enable grade engineered ore streams through the rejection of low value gangue or deleterious elements either in the pit or during transfer to the mill. This paper focusses on the research and development being carried out by the Cooperative Research Centre for Optimising Resource Extraction (CRC ORE) and its associated members on new sensing technologies. At the start of their development these technologies were classified on the Technology Readiness Level scale at 3 (TRL3), equivalent to a proof of concept. Laboratory testing and design is aimed at progressing these sensing technologies to TRL5 which is ready for testing at an end users site. New data obtained during the laboratory testing and design phases are presented in the paper demonstrating performance specifications such as detection limits and measurement speed which can be used to infer the pod size of gangue that can be rejected at scale. Grade Engineering® and coarse gangue rejection are considered as key drivers for the mining industry to address the higher mining, processing and environmental costs associated with ore bodies with declining grades. This is an industry wide problem and the development of mineral sensors is seen as key enabling technologies for Grade Engineering® and coarse gangue rejection. The technologies presented in this paper enhance value through early decision making of ore quality and improve the overall resource value case through improved grade and throughput engineering."

Jan 1, 2018

By J. W. Ahn

Calcined calcium oxide from limestone is hydrated to manufacture calcium hydroxide, and then precipitated calcium carbonate is synthesized from calcium hydroxide. Precipitated calcium carbonate of various morphologies is obtained by controlled conditions of each step and characteristics of limestone, calcium oxide and calcium hydroxide. Especially characteristics of calcium hydroxide have the most influence on morphology of precipitated calcium carbonate character. In this study, therefore, it is investigated the specific surface area of calcium hydroxide and morphology of precipitated calcium carbonate according to hydration temperature. Calcium hydroxide hydrated at high initial temperature has larger particle size than that hydrated at low initial temperature. And formation yield of aragonite precipitated calcium carbonate has improved by using calcium hydroxide had larger particle size (hydrated at high initial temperature).

Jan 1, 2014

By S. Castro

The aim of the present paper is to review the current technologies for arsenic removal in the copper metallurgical industry, and to show the progress made with some promising emerging strategies, involving the pre-treatment of copper-arsenic concentrates by selective flotation or by alkaline leaching with Na2S-NaOH solutions. The advantages of alkaline leaching are the selective dissolution of arsenic and antimony, the high extraction yield, and the practically nil loss of copper, which remains in the clean copper concentrate as a CuS residue. Alkaline ?cured? ? a procedure similar to ?acid-cured? in conventional leaching ? improves arsenic extraction and reagent consumption, making this an interesting technology for the treatment of high As-Sb concentrates. In light of recent advances on the selective flotation of enargite and tennantite, the importance of redox conditions to control the flotation/depression of enargite-tennantite from copper sulfides has been well established. However, problems of selectivity must still be solved. Additionally, the limitation imposed in some cases by the need to recover the molybdenum by-product must be considered. The application of selective flotation to obtain a high-arsenic or ?dirty? concentrate and its sub-sequent treatment by the alkaline leaching process is suggested.

Jan 1, 2014

"The JKMRC researchers have been redeveloping the suite of breakage testing techniques to more realistically align with the breakage environment in comminution equipment. Computational modelling, especially with the DEM, has highlighted the modes of impact and contact, along with the energies of these interactions in a wide range of equipment. This has driven the need to match these modes in laboratory ore characterisation tests. The objective has been to build a continuum of breakage tests and associated models that span the required range of breakage forces and environments. The ore sample is subjected to a sequence of tests covering: abrasion (surface breakage) of angular through to smoothed particles; single particle impact from high energy single impact breakage down to low energy; incremental breakage covering 60 mm down to 50 µm particles; particle strength distribution; and bed breakage over a range of compression and energy input. These tests are suited to multi-component ores with a distribution of strengths. A modelling technique is also being developed to utilise primary breakage for modelling breakage under different types of confinement. This paper draws together these techniques, previously published by the authors, into a coherent whole. The suite of tests is approaching the point where a fixed set of tests can be used to model the performance of a full range of equipment, removing the need to conduct repeat ore characterisation tests for each different piece of equipment incorporated in a circuit design. This path should open up the application of different circuit configurations and novel equipment in comminution circuit design.1 INTRODUCTION Comminution modelling utilizes a wide range of ore characterisation techniques and methods to deal with the range of comminution equipment available for production use. In general, the tests have been designed to correlate with specific pieces of equipment, using calibration factors to provide the link between the small-scale test and industrial-scale performance. The majority of tests focus on estimating the power requirement of process equipment to achieve a desired throughput and grind size. Overall, this approach has achieved acceptable outcomes in the past, but the reality is that equipment sizing suffers from considerable uncertainty, in excess of ±20% is not uncommon. This uncertainty is covered by safety margins and automatic scaling factors in equipment selection. In striving to improve reliability, the industry tends to add more adjustment factors to the tests, and design margin rather than question the basic validity of the testing process. As the tests are often correlated to production equipment via carefully guarded databases, they are also locked into historic operating inefficiency, not providing direction or revealing opportunities to improve production efficiency in comminution equipment or circuits. Safe in the knowledge that they are achieving the expected production based on test work predictions, a site is unlikely to set their sights on reducing specific energy consumption by 10%."

Jan 1, 2018