Table of Contents
Alternative methods to restrict and eliminate the detrimental effect of critical size material on the performance of an autogenous mill.
Palabora has a unique orebody in that it contains most of the valuable minerals known, but unfortunately also some gangue minerals which have a major influence on the milling rate.
The two gangue minerals having the most significant impact are dolorite and magnetite. Dolorite is defined as a basic rock with the composition of gabbro but finer in grain. Mineralogically the dolorite at Palabora consists mainly of plagioclase and pyroxenite.
Because of the severe interference of dolorite on grinding performance, especially in the autogenous circuits, methods have been investigated to alleviate the problem.
The autogenous mills at Palabora operate in the fully autogenous mode. Dolorite with its high workindex does not grind at the same rate as the rest of the ore with the result that it builds up a high circulating load in the circuit. The mill eventually becomes filled up with fine dolorite which upsets the ratio of coarse grinding media to fines and when this situation is reached, the feed rate is drastically reduced. The dolorite can easily be recognised on the recycle belts by its polished appearance and shape.
Tests performed on the autogenous mills showed that the high circulating load of dolorite can be prevented, and the mill can operate normally. Because of its resistance to crushing and grinding, the dolorite grinds down progressively with the result that most of the dolorite ends up on the screen oversize recycle belt at some point.
Since the ore contains approximately 25% magnetite, it is obvious that a large proportion of the recycle product is magnetic. The separation of this magnetic fraction presents no problem and a magnetic head pulley on the conveyor proved to be adequate.
The third and probably the most viable solution in terms of capital cost and simplicity of operation would be to crush the entire recycle product, including the dolorite, and return it to the mill.
It is well known that crushing is far more energy effective than milling, furthermore, because of the tremendous escalation in power costs of the past few years, crushing has become even more cost effective than grinding.
The most logical step to reduce the adverse effect of dolorite on the autogenous milling performance, is to introduce a crushing stage in the circuit to treat the total recycle product and for this a gyradisc crusher is attractive specifically to achieve fine crushing. The autogenous mill discharge screen oversize would be fed to the crusher and the crusher discharge be returned to the screen in closed circuit.
An Analysis of the Environmental Conditions in Palabora’s grinding circuit and the introduction of High Chrome Balls as Grinding Media.
With the exception of power, one of the major cost elements in the rod and ball grinding circuits at Palabora is the steel consumed in the process. The importance of reducing the steel consumption is especially pertinent because of the increasing cost of this commodity.
It has been stated by recognised authorities that corrosion may be responsible for between 40 and 90% of the steel consumed during wet grinding. In dry grinding the metal loss is attributed to abrasion, but in wet grinding it is difficult to explain the larger increase in metal loss other than by corrosion.
The metal losses in a grinding mill are characterised by three main parameters – ore, steel and mill environment. Since these parameters do not readily lend themselves to adjustment, very little control can be exercised in a production plant without fundamental changes affecting the capacity or other conditions.
It is accepted that the minerals in the ore will have a significant effect on the metal loss rate, and abrasive wear on the grinding media should increase as the hardness of the ore increases. The grinding properties of the ore at Palabora can vary considerably as a result of the variations in magnetite content. Apart from being hard and abrasive, the magnetite also builds up in the milling circuit due to its high specific gravity and variations of between 15 – 40% by weight are experienced. However, on an annual basis the magnetite figures have been fairly constant at about 27% average.
The ball consumption at Palabora was fairly steady for the first 10 years of operation. During 1977 a sharp increase in the consumption was experienced. It is still not clear why the sharp increase took place and a corrosion label has been hung onto the condition, because of the lack of other realistic explanation. No changes have been made to the operating mode.
For corrosion to take place, a galvanic condition must exist between materials of differing potential as occurs in a conducting electrolyte.
Although sulphates generated from the gas scrubbing plant and other sources tend to reduce pH and increase TDS concentration, the harmful effects may result primarily from the contribution to increased electrical conductivity of the water aggravating the reaction of the chlorides. Other dissolved salts, principally magnesium sulphates, calcium sulphates, sulphites and nitrates would also stimulate conductivity and encourage the galvanic process.
The removal of the corrosive environment would necessitate the abandonment of wet grinding with its advantage of power requirements and throughput and to attack the problem through the elimination of the cathodic reaction would necessitate the removal of all reducible species or the use of an expensive cathodic inhibitor.
The most feasible method to reduce or eliminate the corrosion rate of steel balls is to inhibit the anodic reaction.
Tests on the use of high chrome grinding balls have been conducted over a period of eight years. Initially a few marked balls of high chromium content were evaluated in an ordinary forged load. A significant factor emerging from the early tests using a mixed charge of forged and high chrome balls was that extremely low wear was measured on the high chrome media. A reduced consumption of 1/17th of the forged ball was recorded for the high chrome.
As a result of this difference in hardness the wear rate of forged balls increases with time. In a ball mill the situation is eventually reached where a charge can consist of hard and soft balls, which leads to the deformation of the softer balls. As soon as the round shape of the ball is lost the grinding efficiency decreases. Results showed that to maintain a grind under these conditions, the energy requirements can increase by as much as 10%.
Two major qualities of balls (designated 5 and 8) have been tested on a large scale following good results achieved in the Pilot Plant ball mill. These were all essentially high chrome balls having a chrome content greater than 25% and a carbon content of about 2,5% with varying quantities of other elements introduced to effect varying heat treatment response. All the balls tested had a hardness of 600 – 800 Brinell, and superior resistance to corrosion as compared to standard forgings.
The introduction of high chrome media at Palabora, has been extremely successful and profitable. Apart from the direct financial impact achieved,the efficiency of the grinding process has improved because of good spherical media competence. A considerable saving has been achieved on the cost of the ball handling as this has been reduced from a daily operation to a weekly task. The volume of stock holdings has been reduced.
Operationally the series circuit has been easier to control, since the split of the rod mill discharge and the balancing of the mills has been eliminated. The series arrangement has some limitation, but the metallurgical principle sought, of concentrating the grinding energy preferentially into the non-magnetics with avoidance of the magnetics-containing only 20% of the copper, has been partly achieved.
Testwork performed in the plant clearly showed that there was an advantage to be gained by using screens instead of cyclones. The circulating load of magnetite was reduced producing a coarser particle size and allowing more work to be put into the non-magnetics.
Although the series arrangement is better than the original parallel milling circuit, the basic problem of overgrinding magnetite persists and results in operating the cyclones at high density and poor efficiency.
Increasing power and media costs encouraged further investigation into reducing the magnetite circulating load other than by simple screening.
The 500 tons per day pilot plant was modified to include magnetic separation, screening and cycloning to the flow sheet, and the plant was operated to produce some practical information over a period of weeks.
