« Back

An Investigation of Potential for Improved Efficiencies in Phosphate Rougher Flotation Through On-line BPL Measurements



For further information on this project go to the Harrison R. Cooper Systems, Inc. website.

The purpose of this investigation was to make measurements of matrix slurries using a novel technique for on-line phosphorus analysis, and to determine from the measurement information if process economies are possible in phosphate beneficiation. BPL data were collected at closely spaced time intervals in the rougher flotation section of two different operating plants. The data were analyzed statistically to identify and quantify process inefficiencies.

The ultimate objective of this work was to determine how to improve the efficiency of phosphate recovery by identifying and correcting the cause of losses. Significant gains in profit from increased recovery and lower reagent costs would result if this objective were achieved.

A Phospholyzer ™, an on-line phosphorus analyzer, was installed in two Florida phosphate beneficiation plants. The analyzer utilizes magnetic resonance to provide immediate measurement of the percent phosphorus in the solids [expressed as “bone phosphate of lime” or (BPL)] contained in slurry streams flowinq through the analyzer. One BPL is equivalent to 0.46 P2O5. The Phospholyzer ™ operates by measuring phosphorus and hydrogen magnetic resonance voltage signals generated by continuous slurry sample flows through the analyzer. The hydrogen signal indicates the water content of the slurry sample (percent solids), and the phosphorus signal is an index of the BPL content of the solids.

Reliable methods and equipment for automatic sample extraction and presentation were developed to enable on-line slurry measurements. These were installed and proven in the two beneficiation plants. Heretofore, no other practical method for phosphorus on-stream analysis in the Florida phosphate industry has been available. BPL measurements of beneficiation process feed and tails were continuously recorded for time periods of two to three weeks in the two plants.

Time intervals for each analysis were from four to ten minutes on each sample stream, depending on the number of sample streams being tested, the time cycle selected, and the method used to operate the analyzer. Measurements were also taken on a third process stream, rougher concentrate, over portions of the investigation period. Because of the low percent solids in sample flow streams, the accuracy of rougher concentrate BPL measurement was inadequate for practical use in the investigation. This deficiency could have been overcome by further development of sample presentation, but feed and tails data provided sufficient data to demonstrate the utility of the method.

The two beneficiation plants (designated “A” and “B”) in which investigations were carried out differed in one principal respect. Plant A operates by manual settings of control devices such as feed and reagent flows. Plant B employs a central supervisory control system. Plant operation is stabilized and logged with a printer using regulatory control loops. Control at plant B is maintained by the supervisory system from settings provided at a control console by plant operators.

Results of statistical evaluation during continuous routine operation indicated substantially greater loss of product using manual control in plant A as compared to plant B using the central supervisory control system. Inefficiencies were observed in both plants, though to a lesser extent in plant B with its central control system. Analysis of the data indicate that gains in productivity of two to four percent are possible with application of stabilizing and other advanced methods of process control. A higher gain could be attained in the less efficient plant A while less improvement would be seen where a more sophisticated control is employed, as in plant B.

The statistical treatment of data is described in the original proposal for this investigation. The treatment was based on estimating the effects on productivity from influences of known periodic disturbances such as shift breaks, the 24-hour effect of day shift supervision, lunch breaks and the like. Past studies in other similar industries, such as base metal sulfide mineral concentrators, have repeatedly shown significant effects on productivity from these periodic disturbances.

Treatment of operating data confirmed the existence of periodic effects in the case of plant A where manual operation only is employed. Only minor periodic effects on productivity are seen in the data from plant B with centralized control.

Further statistical results show significant positive correlation between tails BPL and plant feed BPL, occurring at plants A and B with a more pronounced effect in plant A. The data suggest substantial productivity (recovery) losses are being experienced in plant A. The data suggest substantial productivity (recovery) losses are being experienced in plant A as feed grades rise, and to a lesser extent in plant B when recovery is above 70 percent. Such an observation was unexpected. It may be possible to suppress recovery losses demonstrated by this effect through more sophisticated controls for reagents and other variables.