Enviable water recovery in a desert environment a case study

Transcription

1 Mine Water Solutions 2013 J. Obemeyer, J. Enriquez and T. Alexieva 2013 InfoMine Inc, Vancouver, Canada, ISBN XXX Enviable water recovery in a desert environment a case study James R. Obermeyer, MWH Americas, Inc., USA Justo Enriquez, Sociedad Minera Cerro Verde S.A.A., Peru Tatyana G. Alexieva, MWH Americas, Inc., USA Key words Water Balance, Water Conservation, Water Consumption, Water Recovery, Tailing Dams, Desert Environment, Make-Up Water Abstract Cerro Verde is a large copper mine located approximately 20 miles southwest of Arequipa, Peru. The area is characterized by an arid climate, rapid urban population growth and a long tradition of agriculture irrigated by surface water. The mine has been in operation since the early 1970s. Initially, the processing circuit included heap leach pads, leachate solution ponds, and a solvent extraction and electrowinning (SX/EW) plant. In 2004, Sociedad Minera Cerro Verde S.A.A. (SMCV) undertook final design and construction of the Primary Sulfide Project involving construction of a 108,000 tonnes per day (t/d) concentrator at its Cerro Verde Copper Mine. An important component of the project is the tailing storage facility (TSF) which was permitted as a centerline method tailing embankment with an ultimate height of 260 meters (m) above the foundation at the centerline and 300 m above the downstream toe. The Primary Sulfide Project was commissioned in November A debottlenecking project was completed at the concentrator in 2010 to increase the concentrator throughput to its current capacity of 120,000 t/d. The mine operation is dependent on having sufficient water supply. In most mining operations, the Tailing Storage Facility (TSF) consumes the most water. The largest water losses are attributed to evaporation and water trapped in the voids of the deposited tailing. The average annual evaporation is 60 times the average annual precipitation at the Cerro Verde Mine where it is a standard operating procedure to account for every drop of water. A water balance model was created for the Quebrada Enlozada TSF during its design. The model has been updated and calibrated on a regular basis since the commissioning of the facility using actual information. This paper describes the TSF design and its operational process, presents the history of the water balance model development and the reasons for the low water losses that make the facility one of the most water-efficient tailing storage facilities in the world. Mine Water Solutions 2013, Lima, Peru 1

2 Introduction Water is used for a range of activities at a mine site: for dust suppression, for mineral processes, slurry transport, and other purposes. With the increased interest in developing large low grade deposits, the consumption of water per quantity of metal produced increases as well. This, together with the growing world s population leads to a rise in the competition for water, especially in dry climates. It is well known in the mining industry that the tailing management represents the largest water consumption at a mine site. Water lost to evaporation, wetting the foundation, unrecovered seepage, and water trapped in the voids of the tailing deposit are some of the biggest water losses. For this reason, the design of the modern tailing storage facilities requires careful planning and incorporation of features to minimize the losses and increase the water recovery and recycling. Described in this paper are the design and operation of the Enlozada TSF at the Cerro Verde Mine, which is one of the most water-efficient tailing storage facilities in the world. The history of the TSF water balance model development and calibration is also presented. Cerro Verde Mine is located approximately 20 kilometers southwest of Arequipa, Peru in the Andes Mountains. The elevation at the mine ranges between 2400 and 2700 m above mean sea level. The area is characterized by an arid climate, with mean annual precipitation based on data from a weather station at the mine of about 37 millimeters (mm), and a pan evaporation of about 6 mm/day. Construction of the Primary Sulfide Project began in 2005 and the Concentrator plant was commissioned in November Following a production ramp-up period of 8 months, the mine reached the planned production of 108,000 t/d. A debottlenecking project was completed at the concentrator in 2010 to increase the concentrator throughput to its current capacity of 120,000 t/d. Processing 120,000 t/d through the Concentrator requires a significant amount of water. Considering the water shortage in the area, an effort was made during the design and planning phase of the Primary Sulfide Project to incorporate features aimed at minimizing the water losses and maximizing the water recycling and reuse. Enlozada TSF Design The TSF was an important component of the Primary Sulfide Project. The design of the TSF began with defining the design objectives, design basis and criteria. The main objectives adopted for the TSF design were: To store the tailing materials in an environmentally responsible manner. To satisfy internationally accepted stability criteria for embankment construction in areas of high seismicity. To satisfy all relevant Peruvian regulatory requirements associated with construction of the TSF. To follow the concept of zero discharge, which means the use of proven and feasible state-ofthe-practice engineering and technology for the design, construction, operation and closure of a facility, in order to minimize discharges from the facilities to the environment. The principal design basis and criteria adopted for the Quebrada Enlozada TSF are presented in Table 1. Mine Water Solutions 2013, Lima, Peru 2

3 Table 1: Enlozada TSF design basis and criteria Total Capacity About 1 Bt Bt Billion metric tonnes Production Rate 108,000 t/d Ramp-up schedule for 1 st 6 months Percent Solids from Tailing Thickeners 55% Range 50-60% Percent Fines in Tailing Slurry 67.5% Weighted Average Tailing Sample Average Compacted Dry Density of Underflow 1.58 t/m 3 98% of max dry density (ASTM D 698); the 1.58 t/m 3 value will vary for different tailing materials over the life of the mine. Start-up Water Requirement 1,000,000 m 3 To provide water for the start of the operations Flood Storage Requirement Seismicity/Earthquake Load Min Freeboard (Vertical distance between embankment crest and max impoundment elevation) PMF MCE 3 m Min Static Factor of Safety (FOS) 1.5 Post-earthquake FOS 1.2 PMF Probable Maximum Flood PMP Probable Maximum Precipitation MCE Maximum Credible Earthquake M9 Magnitude 9 PGA peak ground acceleration Contain within impoundment, min 100 m from embankment crest during flood conditions. 72-hr PMP = 391 mm Canadian Dam Association, Dam Safety Guidelines, 1999; M9, PGA = 0.47g To provide sufficient flood storage and accommodate anticipated earthquake induced settlements. Canadian Dam Association, Dam Safety Guidelines, 1999; USBR, Chapter 4, 1987 ANCOLD Guidelines, 1998; USBR, Chapter 13, 2001 The required large capacity and production rate, high seismicity of the area and dry climate were major considerations in the design of the TSF. More than 10 sites were evaluated and compared in terms of technical, environmental, social and economic features, and the Quebrada Enlozada site was selected as the best TSF site. Conventional, thickened, paste and filtered tailing disposal methods were evaluated in the process of selecting the best TSF option for further development. The main considerations were water conservation, stability of the facility, potential impacts to the environment (seepage, dust, flora and fauna), and operational practicality. The large production rate made the use of tailing filtering and dry stacking, as well as the application of paste, not practical at that time. Conventional tailing disposal was selected as the preferred disposal method. Thickening of the tailing using conventional thickeners was adopted to reduce the amount of the recirculated water delivered to the tailing facility in the slurry and to potentially reduce the water losses. Following the selection of a TSF site and a tailing disposal method, an embankment alternatives study was performed to select an embankment type and embankment construction materials. The embankment selection study included consideration of centerline and downstream embankment types, and a variety of Mine Water Solutions 2013, Lima, Peru 3

4 construction materials including cycloned tailing underflow, mine waste rock and quarried rock. Due to the relatively low efficiency ratio of the selected site, water conservation considerations and economic factors, a centerline embankment type constructed of compacted cycloned tailing underflow was selected as the best alternative. The TSF is located in Quebrada Enlozada immediately north of the concentrator processing plant, as is shown on Figure 1. Figure 1: TSF plan view The TSF consists of an 85 m high zoned rockfill Starter Dam, a 260 m high embankment constructed of compacted cycloned tailing sand by the centerline method, and a tailing impoundment that will cover an area of approximately 453 hectares (Ha) at its ultimate configuration. The base of the Quebrada Enlozada at the location of the Starter Dam is at an approximate elevation of 2400 m above mean sea level (amsl). The elevation of the processing plant is about 2,700 m amsl. The Starter Dam crest elevation is at 2485 m amsl and the ultimate embankment crest elevation will be 2660 m amsl. A cross-section of the TSF embankment is shown on Figure 2. Mine Water Solutions 2013, Lima, Peru 4

5 Figure 2: TSF embankment cross-section The Starter Dam was sized to contain the tailing produced during the first year of operation and allow for raising the cyclone sand embankment to an elevation sufficient to provide the required freeboard. As per the material balance analysis, the estimated embankment rate of rise ranged from over 2 m per month in the first year of operation to about 1 meter per month during the first five years of operation, and reducing to less than 50 centimeters (cm) per month at the end of the operations. Some of the important factors for achieving this high rate of rise were the permeability of the underflow sand and the availability of an extensive embankment underdrains network that facilitates fast drainage of the deposited underflow. As per the design, the underflow was planned to be delivered at about 70% solids by weight, hydraulically deposited in 30 cm loose lifts, and compacted to 98% of the maximum dry density (ASTM D698) of the underflow. The specification for the cyclone tailing underflow called for maximum of 15% fines (particles finer than 75 micron), which was an important factor for achieving high permeability, fast drainage and a low phreatic surface within the embankment. The network of embankment underdrains was a critical element of the design because it promotes rapid drainage of the cycloned tailing sands, maintains the phreatic surface low in the embankment section and expedites recovery of water that drains from the underflow sands after deposition. The underdrains are planned to be constructed in phases, as shown in Figure 3, to spread capital expenditure over the life of the facility and to reduce potential contamination and erosion of the drains, if exposed over a long period of time. The water collected by the underdrains is conveyed to the Seepage Collection Sump. Mine Water Solutions 2013, Lima, Peru 5

6 Figure 3: Embankment underdrains plan The Seepage Collection Sump provides a pond from which the collected seepage and tailing embankment stormwater runoff are pumped to the cyclone station or to the concentrator plant for reuse. The Seepage Collection Sump includes design components to minimize the risk of seepage from the TSF bypassing the sump, and to satisfy the zero discharge concept criteria described above. The main design components of the Seepage Collection Sump are: An excavation through the alluvium across the valley down to fractured bedrock to create a sump. A geosynthetic liner on the downstream slope of the sump anchored to a concrete grout cap at the excavation bottom and side slopes. A grout curtain at the bottom of the excavation extending through the fractured bedrock into low permeability bedrock. A sump with a capacity to store runoff from the 100-year, 24-hour storm plus a small operating pond and operational flows to the sump during an assumed 12-hour power outage. A spillway designed to pass flows from the 500-year, 24-hour storm. Five monitoring wells located downstream of the sump to monitor groundwater quality. The wells are equipped with pumps to intercept subsurface seepage that bypasses the grout curtain. Other aspects of the TSF design that have an impact on the water consumption is the impoundment deposition practice and the size of the reclaim water pond. The main objectives of the impoundment deposition plan that was included in the TSF Operations Manual were to manage the size and location of the reclaim water pond, and to minimize development of isolated water ponds that would be inaccessible for recovery by the reclaim water system. This is achieved by deposition from strategic points around the Mine Water Solutions 2013, Lima, Peru 6

7 impoundment perimeter that force the water to the Southeast corner of the impoundment. Water from the reclaim water pond is pumped using barge-mounted pumps and conveyed to the processing plant for reuse and to the cyclone station for dilution. To provide a minimum barge operating pond depth, but maintain a small pond area despite the very flat beach slope (<0.5%), the barge is located in an excavated channel (see Figure 4). The barge is also designed to skim clear water from the surface of the reclaim water pond enabling the pumping system to deliver water with minimal total suspended solids to the concentrator with only 2.5 meters of pond depth. This is a significant factor in maintaining a reclaim water pond with a very small surface area and a volume that has normally ranged from only 200,000 to 550,000 m 3 during the first 6 years of TSF operation. Figure 4: Reclaim water barge channel at the completion of the capital construction TSF Operation The construction of the Enlozada TSF began in May 2005 and was completed in August During the design, it was estimated that a start-up water pond would be required to offset the larger water losses expected at the beginning of the operations due to wetting the foundation and time required to create a tailing beach and a clear water pond. Filling the startup water pond began in August 2006 and was completed in October The concentrator plant was commissioned on November 20, Mine Water Solutions 2013, Lima, Peru 7

8 Tailing generated in the flotation process is sent to two high capacity thickeners that thicken the tailing slurry from about 27% solids to about 55% solids by weight. The thickened slurry is conveyed by gravity through a 48-inch diameter HDPE SDR 21 pipeline to a cyclone station located on the right (East) abutment of the tailing embankment. At the cyclone station, the slurry undergoes two-stage cycloning to separate the sand fraction from the fine fraction. Due to delays in completing the cyclone station, deposition of tailing was initiated with whole tailing deposition at a single point located at the right abutment, followed with whole tailing deposition from spigots along the embankment crest. Deposition of underflow began by stockpiling sand on the blanket drain downstream of the Starter Dam toe on December 27, Deposition of underflow on the embankment began on January 3, Pumping of water from the Reclaim water pond back to the concentrator began immediately after the start of the operations, and pumping of water from the Seepage Collection Sump began a few days after deposition of underflow was initiated. Make-up water to the plant to compensate for the losses at the TSF is conveyed from Rio Chili via an 11.5 kilometers long freshwater delivery pipeline. This line was used prior to the start of the operations for filling the startup water pond. Figure 5 shows a recent photo of the Quebrada Enlozada TSF. The embankment height is approximately 150 m under the centerline as of October Figure 5: Enlozada TSF, October 2012 Mine Water Solutions 2013, Lima, Peru 8

9 Water Balance Model The main objective of the water balance model developed for the TSF during the detailed design stage was to estimate the water losses and the required make-up water. The main inputs to the water balance were deposition rate and method, percent solids of the deposited underflow and overflow tailing materials, elevation/area/capacity curves, evaporation and precipitation data, density of the in-situ tailing, pond size, wet and dry beach area ratio, moisture retained in the underflow and bank storage in the foundation materials. The results of the water balance include estimates of water loss due to evaporation, entrained water in the voids, seepage pump back rates, and estimates of required make-up water. Figure 6 presents a schematic diagram of the primary components of the water balance model. Figure 6: Water balance model schematic Based on an assessment of the available data and of the complexity of the model, it was decided that developing a spreadsheet-based deterministic model would be appropriate. The model was based on a monthly time-step and covered the entire operational life of the TSF. The water balance model was directly connected to the results of material balance model providing estimates of the embankment and impoundment raise with time. The density curve illustrating the change in average density of the impounded tailing used in the material balance analysis was also used in the water balance model. The density curve was created based on the results of laboratory testing of a pilot plant tailing sample, followed by consolidation modeling using the computer code FSConsol. It was assumed that the reclaim water pond size will remain constant throughout the operational life of the TSF, while the amount of reclaim water pumped back to the plant will vary based on availability. The climate of the area is mild and arid with temperatures fluctuating between 10 and 24 C and average annual precipitation of approximately 35 mm. The rainstorms occur seasonally and are typically of short Mine Water Solutions 2013, Lima, Peru 9

10 duration and high intensity. Over 90% of the annual rainfall is recorded during the months of January, February and March. The recorded evaporation at the project site exceeds the precipitation over 60 times. The estimated average annual evaporation rate is about 6.1 mm/day. The humidity ranges from about 30% in July to about 70% in February. It was decided that the climate data from the South Zone meteorological station located at the mine should be used despite the short period of record (about 9 years). Average monthly evaporation and precipitation values were used in the model. The sensitivity of using the wettest and driest year on record was also evaluated. The results of the water balance analysis performed during the design phase indicated that the required make-up water would range from over 700 liters per second (l/s) in the first year of operation to about 500 l/s during the later years of operation. The higher water consumption in the initial years of operation was related mostly to the rapid rate of rise and lower achieved average density of the impounded tailing. As the tailing materials consolidate with time, the recovery of water increases. Since the start of the operations, the model has been regularly updated with actual input information, including production rate, cyclone operating time and underflow sand deposition. The results of the updated analysis are illustrated in Figure 7 below. The actual make-up water consumption is plotted on the figure with circles representing average monthly make-up water consumption. As was mentioned above, it was assumed that the reclaim water pond would be maintained at a minimum constant value. In reality, the size of the pond has fluctuated (see Figure 7). It is interesting to note that the periods of recorded higher make-up water consumption have coincided with periods of accumulation of a larger pond and vice-versa. Maintaining a constant pond at about 250,000 m 3, as assumed for the water balance analysis, would have resulted in an even closer match between the estimated and actual results. Lower make-up water consumption was also observed during the unusually wet periods of 2008, 2011 and The variations in the estimated make-up water requirements represent the seasonal variation in the precipitation and evaporation. Peaks in the estimated make-up requirements represent the impact of the driest year on record, while the lowest estimated make-up requirements represent the impact of a wet year. Make-up Water Requirements, L/s Measured Data Design Adjusted with Actual Production Reclaim Pond Volume Year 2026 Figure 7: Make-up water requirements Design Assumed Dry Years Assumed Wet Year Pond Volume, Million m 3 Figure 8 presents the estimated and actual make-up water requirements at the Cerro Verde Concentrator plant in cubic meters of water per tonne of processed ore. The figure confirms the higher expected water Mine Water Solutions 2013, Lima, Peru 10

11 consumption during the first year of operation. As was described above, some variation of the water consumption is observed that is mostly due to fluctuations in the pond size and experiencing higher than usual precipitation. A comparison of these results with results from mines located in similar climatic regions (Arizona, Southern Peru and Chile) show that the Cerro Verde Mine has the lowest water consumption with an average actual make-up water of less than 0.38 m3/t, compared to reported make-up water from about 0.4 to over 0.7 m 3 /t for the other mines Measured Data Design Adjusted with Actual Production t water/t ore Design Figure 8: Make-up water requirements per tonne processed ore Factors for the Low Water Consumption Year Achieving low water consumption and minimizing potential impacts to the environment were some of the main objectives of the Quebrada Enlozada TSF design. Comparison of the design and actual water balance results indicate good correlation and provide confidence that the TSF water balance is about as expected. Some of the main factors influencing the low water consumption of the Concentrator operations are presented below: 1. Embankment Construction Method: As was described above, the TSF embankment is raised using tailing cyclone underflow sand. The hydraulically deposited sand drains within a few hours and remains at a moisture content of about 13%. This results in the recovery of a large portion of the water. Because the impoundment to embankment volume ratio is relatively low (about 3.3), a large amount of tailing is deposited on the embankment rather than in the impoundment. The high density and low moisture content of the embankment result in much higher water recovery from the embankment than from the impounded finer tailing, which Mine Water Solutions 2013, Lima, Peru 11

12 Closing remain saturated and at a lower density. Cycloning a significant portion of the tailing reduces the losses due to water trapped in the voids and increases the water efficiency of the TSF. In the case of the Quebrada Enlozada TSF, an average of about 25% to 30% of the tailing materials by weight are placed in the compacted underflow embankment. 2. Efficient Seepage Collection System: The network of embankment underdrains facilitates fast harvesting and reuse of water from the underflow. The low permeability of the impoundment foundation limits the amount of water lost due to foundation wetting and minimizes seepage from the impoundment. Seepage from the impoundment makes its way to the Seepage Collection Sump where it is collected and returned for reuse. 3. Small Impoundment and Reclaim Water Pond Surface Area: The TSF is located in a steep, bowl-shaped valley with small surface area, which results in low evaporation losses. Maintaining a small pond size, of about 40 Ha results in reduced evaporation losses, and is one of the priorities of the TSF operators. 4. Optimized Deposition Practice: The deposition plan was designed to minimize the development of clear water ponds in the side drainages by depositing from around the impoundment perimeter. Another important function of the deposition practice is maintaining a reclaim water pond in a confined area in the southeast corner of the impoundment. Minimizing the number and size of inaccessible water ponds results in an increase in the water reporting to the reclaim water pond from where it is pumped back for reuse. 5. Reclaim Water Pond Size: One of the keys to maintaining a small reclaim water pond on the TSF, in spite of the relatively flat bench slope within the impoundment, is the ability to pump water of suitable quality for the concentrator from a pond with a depth of only 2.5 meters. This is the result of a barge design for the pumping system that skims clear water from the surface of the pond. Operating the barge pumps in a pre-excavated channel along which the barge pumps move as the TSF grows is also a factor contributing to being able to maintain a small pond. 6. Tailing Thickeners: Two high capacity tailing thickeners are used to thicken the slurry exiting the concentrator plant from about 27% to about 55% solids by weight. Thickening the slurry prior to conveying it to the TSF results in the recovery of a significant amount of water close to the plant, and thus contributes to reduced pumping costs and water losses. 7. Closed-Circuit Water Balance: Performing a detailed water balance analysis early in the design of the TSF contributed to gaining a good understanding of the factors resulting in water losses and provided an opportunity for optimizing the design to achieve a better water efficiency. The water balance analysis is updated regularly to verify its accuracy, and adjust the operations to improve water recovery. The water balance, together with the results of the monitoring program, confirms that the facility is operating consistent with the zero discharge concept adopted for the design, and that the TSF water balance is as predicted during design of the facility. Water is a precious commodity, and in many cases, the ability of a mine to operate is contingent on having sufficient make-up water to compensate for the losses incurred during the operations, which are mostly accounted by the TSF. For this reason, considering water saving measures in the design of the TSF, and developing an accurate water balance model are important factors for the success of the project. The experience at the Cerro Verde Mine demonstrates that careful planning and attention to the established operation of the TSF has resulted in low water losses and having one of the most water efficient TSFs in the world. Mine Water Solutions 2013, Lima, Peru 12