By S. Keles, G. Luttrell and R.H. Yoon; and T. Estes and W. Schultz

The dewatering of fine coal is widely considered to be the most difficult and costly step in coal preparation. The cost of dewatering is a strong function of particle size and increases sharply for particles finer than 28 mesh. In fact, many coal producers often find it to be more economical to discard fine coal provided it constitutes only a small fraction of the overall product stream (Leonard, 1991). As a result, approximately 2 billion tons of fine coal has been discarded in abandoned ponds in the U.S., and 500 to 800 million tons are in active ponds (Orr, 2002). On a yearly basis, U.S. coal producers discard approximately 30 to 40 million tons of fresh fine coal to ponds. This represents a loss of valuable natural resources, a loss of profit for coal producers, and the creation of potential environmental concerns related to waste storage.

In the past, thermal drying was the only practicable method of drying fine coal to below 10% moisture by weight (Osborne, 1988). Unfortunately, thermal dryers are capital intensive and difficult to justify in the current coal market. This is particularly the case with pond fines recovery projects, whose life spans can be only a few years. Furthermore, new air quality standards can make it difficult to obtain permits to install thermal dryers in many states. An attractive alternative to thermal drying would be to use mechanical dewatering systems; however, existing processes are currently incapable of generating very low moisture contents when used to treat fine coals. Therefore, new dewatering technologies are needed to overcome this limitation.

In light of the problems associated with fine coal dewatering, Virginia Tech researchers have teamed with Decanter Machine to develop a hyperbaric filter centrifuge that uses a combination of gas pressure and centrifugal force to increase the driving force for dewatering (Yoon and Asmatulu, 2002). Data collected to date suggest this process can significantly reduce the moisture of fine coal by 30%-50% compared to existing dewatering processes.

Experimental Testing
Substantial improvements in moisture removal can be expected through the use of a hyperbaric (air pressurized) filter centrifuge. To demonstrate the potential of this approach, several test runs were conducted with the batch laboratory-scale test unit. The test unit consisted of a rotating filter drum that was perforated and lined with a suitable filter cloth (or mesh).
During testing, feed slurry was placed within the drum. The slurry formed a compact filter cake along the wall of the filter chamber upon application of the centrifugal force. Compressed air was then injected into the rotating assembly to further increase the pressure drop across the filter cake. The combined action of the centrifugal force and air pressure made it possible to achieve a high rate of dewatering and low equilibrium cake moisture.

The laboratory tests were conducted using fine coal slurry samples that had been thickened to 60%-70% solids by means of sedimentation or prefiltration. All of the tests were conducted at a 15-mm cake thickness using minus 28 mesh coal. As shown in Table 1, the first series of tests were performed under a centrifugal force of 2,000 G without injection of compressed air. The moisture contents obtained using centrifugation alone were in the range of 24.4% to 21% at 30-120 sec of centrifugation time. The next series of tests were carried out at 15 psi of pressure, but without centrifugation. In this case, the cake moistures ranged from 23.8% to 27.5% depending on the drying times employed. In the third series of tests, both air pressure (15 psi) and centrifugal force (2,000 G) were used. The cake moistures obtained by the combined action of these forces were in the range of 14.2% to 10.6%. At a common drying cycle time of 60 sec, the pressure filter, filter centrifuge and hyperbaric centrifuge produced filter cakes with moisture contents of 25.8%, 22.6% and 12.9%, respectively. These results show the addition of compressed air greatly increased the moisture reduction, which can be attributed to the increased pressure applied to the dewatered cake.

Table 2 shows the results of another series of laboratory-scale tests conducted on a very fine (22 micron median) coal sample from a column flotation/pond recovery plant. Because of the extreme fineness of this sample, two series of tests were performed using (i) a sample of the as-received froth product and (ii) a sample of partially deslimed froth product. The deslimed product was prepared by removing half of the minus 325 mesh material from the sample by wet sieving. The tests were conducted at 2,500 G using a constant 0.6 inch cake thickness for the as-received product and a 0.4 inch cake thickness for the partially deslimed product. As expected, the removal of moisture improved when either the spin time or air pressure was increased. For the as-received sample, a high moisture of 39.1% was obtained even after 120 sec of spin time when no air pressure was applied. However, the moisture was substantially reduced to 26.1% in the presence of 37 psi air pressure under the same operating conditions. For the deslimed sample, the moisture contents were reduced down to 25.4% after 120 seconds of spin time without air pressure. After applying just 7 psi of air pressure, it was possible to achieve a substantial reduction in moisture content down to 8.0%. A further increase in air pressure to 37 psi reduced the moisture content of the deslimed sample to as low as 4.6% at the longest spin time (120 sec).

Pilot-scale Testing
Based on the promising results obtained from the laboratory tests, a continuous prototype of the Centribaric technology was designed and constructed by Decanter Machine, Inc. The prototype unit and all ancillary components were trailer-mounted so the equipment could be easily relocated to different industrial sites (Figure 1). The mobile platform included space for a 450 gallon feed storage tank, tank mixer, slurry pump, gas compressor, discharge collection bin, effluent pump/sump and electrical control systems. The prototype was designed to handle a volumetric capacity of approximately 30 to 35 gpm of feed slurry.

Field testing of the prototype unit was performed using fine coal slurry from an industrial preparation plant. The feed to the prototype primarily consisted of ultrafine (minus 325 mesh) clean coal, although small amounts of either screen-drain product (nominally minus 28 mesh) from a screen-bowl centrifuge or clean coal froth (nominally 65- x 325-mesh) from a deslime froth circuit could also be introduced to evaluate the effects of changing feed size distributions. Various combinations of these three clean coal products were collected in the feed sump and thoroughly homogenized prior to feeding to the prototype unit to ensure a consistent/identical feedstock quality was maintained throughout each series of tests. In total, seven different feed mixture combinations were evaluated. In each test, a progressively finer feed stream was used to demonstrate the effects of increased amounts of ultrafines on product moisture content.

During testing, representative timed samples of feed, product, screen effluent and main effluent were taken. After taking the dewatered product sample and two effluent samples, a full timed sample of feed slurry was taken by diverting the stream from the centrifuge to a sample container to estimate the feed flow rate. The feed stream was then redirected back to the centrifuge feed and sufficient time was given to reach steady-state operation with the next set of test conditions before collecting the next set of samples. Since the screen drain stream was not circulated back to the feed stream during testing, no experimental data was collected that included the effect of circulating the screen drain back to the feed stream.

Table 3 shows the moisture and recovery data obtained from the in-plant testing of the pilot-scale Centribaric unit. The first run shows a moisture as low as 13.4% can be obtained when using a relatively coarse feedstock (i.e., 50:50 mixture of minus 28 mesh screenbowl screen drain product and minus 325 mesh column flotation concentrate). From this baseline test run, the moisture climbed to 19.7% as the mass percent of ultrafines increased. The fact the moisture in the final run was below 20% was considered to be very impressive given the extreme fineness of the feed material used in the final test run (96% minus 325 mesh). The test data also demonstrated very high recoveries of ultrafine solids could be achieved while maintaining low moisture contents. The highest recovery of dry solids (i.e., 96.1%) was obtained for the coarsest feed mixture, while only a slightly lower recovery (i.e., 89.4%) was obtained for the finest feed mixture. Iterative simulations were also performed using a mass balance routine to determine the expected solids recovery and product size distribution that would be expected by returning the screen drain back to the feed. The product moistures for each run were also estimated using an empirical correlation between the measured moisture contents and the amount of minus 25 micron particles present in the dewatered products. The simulations showed dry solids recoveries of 98% or greater could be attained by circulating the screen drain, while still maintaining moistures only 2% to 5% higher.

It was originally anticipated that an approximately linear correlation between the moisture values and the amount of ultrafine solids in the product. Such a correlation has been observed for conventional dewatering processes used in the coal industry (Arnold, 1999). However, as shown in Figure 2, the moisture obtained using the Centribaric technology increased linearly until the mass percent of minus 25 micron solids reached 60%. The moisture then reached a plateau at about 19.4% and did not increase further as the amount of minus 25 micron solids increased. This trend was also found to be reproducible in laboratory tests conducted with air injection. No such plateau was observed in the laboratory tests where no air injection was employed (i.e., the moisture continued to increase in a linear fashion as the amount of ultrafines increased). As such, the Centribaric technology appears to be ideally suited for dewatering of ultrafine feeds to low moistures.

Encouraged by the test results from the pilot-scale unit, a full-scale commercial Centribaric unit was sold and installed in an eastern U.S. coal preparation plant. A photograph of this full-scale production machine is provided in Figure 3. Although testing of the unit has been successfully completed, the data from the industrial site cannot be provided at this time due to confidentiality agreements. The unit did, however, meet or exceed all expectations in terms of dewatering performance, dry solids recovery, operating cost and mechanical service. Additional installations of the technology are now moving ahead in the coal industry.

About the Authors
Keles, Luttrell and Yoon work with the Center for Advanced Separation Technologies at Virginia Tech. Estes and Schultz work with Decanter Machine, Inc. This article was adapted from a paper they presented at Coal Prep 2011, which took place during May in Lexington, Ky. For more information visit www.coalprepshow.com.

Acknowledgements
The authors would like to acknowledge the technical and financial support provided by the Center for Advanced Separation Technologies, U.S. Department of Energy Cooperative Agreement No. DE-FC26-01NT41607. The contributions made by the various industrial partners involved in the project are also gratefully acknowledged.

References
Arnold, B. J. (1999). Simulation of dewatering devices for predicting the moisture content of coals. Coal Preparation, 20(1-2), 35-54.
Darcy, H. (1856). Les fontaines publiques de la ville de Dijon. Paris.
Keles, S. (2010). Fine Coal Dewatering Using Hyperbaric Filter Centrifugation. PhD Dissertation, Virginia Tech, Blacksburg.
Leonard, J. W. (1991). Centrifugation Coal Preparation (5 ed., pp. 528-540). Littleton, Colorado: Society for Mining, Metallurgy, and Exploration.
Orr, F. M. (2002). Coal Waste Impoundments: Risks, Responses and Alternatives. Washington, D. C.: National Research Council.
Osborne, D. G. (1988). Solid-Liquid Separation Coal Preparation Technology (Vol. 1, pp. 478-542). London; Boston: Graham & Trotman.
Yoon, R.H., & Asmatulu, R. (2002). USA Patent No. 6440316 B1.
Zeitsch, K. (1990). Centrifugal Filtration. In L. Svarovsky (Ed.), Solid-Liquid Separation (3 ed.). London; Boston: Butterworths.