Blasting for Improved Autogenous Milling
Blending ores of various grades is an essential function of mine engineers. Iron ore blending focuses on a number of properties, liberation values, weight recovery and concentrate silica being the most important. Other chemical and metallurgical factors are also commonly controlled. Recent research has fueled an emerging interest in blending on physical ore properties as well.
Fully autogenous mills rely on the large, competent fraction of feed to act as grinding media. Research indicated that Hibtac mills require specific amounts of 6 to 10-inch ore. High recirculating loads require a steady influx of large rock. Blast designs, therefore, are presently characterized by wide patterns and low powder factors. Current efforts are aimed at improving mill throughput while reducing overall energy costs.
Based on research and published case studies and the advent of measuring devices; Hibtac has embarked on a blast optimization program to identify optimum mill feed and to design blast fragmentation goals for each mining horizon and each mining area.
This paper describes an ongoing, broad-based, team effort which requires close cooperation of geologists, mine engineers, crushing and milling personnel. Three specific areas of investigation are described:
- 1) Historical relationships between powder factor and mill performance,.
- 2) Blast fragmentation modeling using the Kuz-Ram model and
- 3) Drill core measurements relating bed thickness to mill performance. Early findings indicate that in-situ bed thickness has an effect on mill throughput. A second finding is that even high powder factor blasts still produce a large amount of the coarse feed needed by the autogenous mills.
Physical Ore Specifications
Mine engineers and plant metallurgists recognize the importance of close control of feed to the plant. Every operation has developed key parameters on which the daily blend is based. Mesabi Range iron ore is no exception. Weight recovery, liberation values, silica and ferric/ferrous ratios must be controlled as well as slaty versus cherty rock percentages.
Hibbing Taconite Company (Hibtac) has a long history of also specifying a maximum powder factor (lbs. of powder per long ton of rock) used in blast design. Fully autogenous (AG) mills require feed which includes a percentage of coarse sizes. Early research pointed to an optimum feed of 40% minus 3-inch, 20% 3 to 6-inch and 40% 6 to 10-inch rock. Actual blasted and crushed feed exhibits a spectrum of size fractions reflecting: rock properties, blast design and crusher setting. Given the recent advances in computer modeling and in fragment size measurement; revisiting these size specifications may be helpful. Two obvious questions include: What is the best feed for the AG mills and what blast designs are required to economically produce such feed?
Today, Mesabi Range iron ore producers are challenged, as never before, to reduce costs or face closure. High grade deposits were depleted during the past century of mining. Worldwide competition from high grade producers is intense. Exacerbating the situation is the rising cost of energy. Flint-hard taconite must be ground to 75% minus 325 mesh in order to liberate magnetite from the gangue. Concentrate is subsequently pelletized in another energy-intensive process to facilitate shipping and blast furnace productivity.
In response to this challenge, Hibtac has embarked on efforts to fully understand the relationship between blasting practices and mill performance. Optimizing the throughput and energy efficiency of the AG mills is the object of a three year study which has received funding from the Department of Energy (DOE) plus matching contributions from industrial partners. Following herein is a review of the developing study of blast/mill relationships at Hibtac.
Historical analysis of blasting, crushing, and mill performance.
Hibtac production data from startup in the mid 1970’s to the present is charted in this section. Numerous flow sheet changes, changes in mining areas plus changes in blast design occurred over the past 3 decades, which makes interpretation difficult. However, fundamental efficiencies of each process in fragmentation may be evident.
Blast energy and crusher energy
Figure 2 chart is a plot of annual powder factor in pounds of powder per long ton of ore versus the crusher kw-hr per long ton. Higher powder factors result in increased kw-hr/LT at the crusher. This may reflect an increased amount of finer fragments which tend to draw higher amps.
Crusher energy and total energy
Figure 3 is a similar plot comparing kw-hr/LT for the crusher versus total Hibtac kw-hr/LT. Total energy consumption is dominated by milling. As crushing energy rises, total energy falls. The following observations may explain this phenomenon:
- 1) Crushing is more efficient at producing surface area. Efficiencies in the order of 50% for crushing and of 1% for grinding have been estimated (Hukki, 1975, Morrell et al, 1992)
- 2) Finer blasting causes crusher amps to rise, while mills may drop due to the additional minus 3-inch fraction
In-Situ size affects mill rate
Hibtac blast patterns are characterized by wide burden and spacings and low powder factors. As a result, the fragmentation level of run-of-mine rock may be largely a function of the frequency of joints and bedding planes. In order to begin to quantify in-situ fragment size, drill core was measured to model the percentage of material within different size fractions. Each core was broken into discrete geotechnical intervals that were measured for the cumulative lengths of pieces greater than 2 inches, greater than 4 inches, greater than 8 inches, and greater than 10 inches, and divided by the total length of the interval to determine the percentage of summed lengths of pieces. Care was taken to avoid measuring to obvious man-made fractures in the core box. Also, heavily oxidized intervals with poor core recovery do not reflect taconite ore, and were not measured. From these measurements, additional bins were created that reflect the percentage of core pieces less than 2”, between 2” and 4”, between 4” and 8”, and between 8” and 10”. Through weight averaging, the various geologic units were roughly modeled by mining area. Core length values were then assigned to historical daily mine production through reconciling blast patterns in the daily blend with weight averaged geologic unit determinations from the nearest cluster of diamond drill holes. In this fashion, a daily, weighted average of core length was generated. The attached chart Summarizes core length versus mill throughput.
Core piece length measurements are a recent initiative at Hibtac, and the geotechnical database consists of only 42 diamond drill holes in four clusters marginal to the active mining areas. Therefore, the blasts could be quite distant from the drill hole cluster on which their sizing model was based. Furthermore, all diamond drill holes at Hibtac are vertical, and the mostly steeply dipping joint sets were not consistently intersected. Also, the daily weighted fragment size data do not reflect sporadic contributions to crusher feed derived from active stockpiles to which no sizing model was applied. Despite these limiting factors, preliminary results suggest a possible correlation between core length and mill productivity. Decreasing mill throughput trends with increasing amounts of coarsely bedded material. Likewise, as the less than 2” inch portion rose, so did the mill throughput. These results do not seem to be consistent with the current mill demands for feed a coarse as possible, and may, through further investigation, shed new light on what constitutes optimum feed for autogenous milling. Geotechnical core measurements will continue, and as mining progresses into the areas with measured core holes, enough data may exist to warrant digital modeling of in-situ fragment size.