By R. Bratton, Z. Ali and G. Luttrell, and R. Bland and B. McDaniel

Historically, thermal dryers have been used in the coal preparation industry to reduce clean coal moisture to single-digit values whenever mechanical dewatering processes were incapable of meeting contract specifications. When operating correctly, thermal dryers can reduce the clean coal moisture to less than 6% by weight. Unfortunately, thermal dryers involve a substantial investment when installed and large annual costs for equipment maintenance and repair throughout their life. Operating costs for thermal dryers have increased and they can be difficult to permit due to more stringent emissions standards.

Realizing the urgent need to develop an innovative, efficient and low cost technology for removing moisture from fine coal, entrepreneurs based in Beckley, W.Va., developed a novel thermal-mechanical dewatering process, Nano Drying Technology (NDT). An experimental test program at Virginia Tech recently evaluated the dewatering performance of the NDT process. And, the results look very promising.

Nano Drying Technology
The NDT drying system uses molecular sieves (think little sponges) to wick water away from wet fine coal particles and does not require crushing or additional finer sizing of the wet coal to dry it. These molecular sieves are a form of nano-technology based particles, which are typically used for extracting moisture from airborne, aerosol and liquid environments. Molecular sieves contain pores of a precise and uniform size, typically in the range of 3 to 10 angstroms. These pores are large enough to draw in and adsorb water molecules, but small enough to prevent any of the fine coal particles from entering the sieves. Some can adsorb up to 42% of their weight in water. They can be re-used after the absorbed water is removed by heating.

Molecular sieves often consist of alumino silicate minerals, clays, porous glasses, micro-porous charcoals, zeolites, active carbon or synthetic compounds that have open structures through or into which small molecules such as nitrogen and water can diffuse. When the molecular sieves are mixed with wet coal fines, these sieves quickly draw water away from the wet solids. To maximize surface contact between molecular sieves and coal particles, the mixture is contacted/ mixed/agitated for a short period of time. After contacting, the molecular sieves are recovered from the dry coal by simple screening since the sieves are substantially larger in size than the top-size of the dried coal particles.

Once the separation occurs, the remaining coal particles have a substantially reduced moisture content, which can reach low single-digit values regardless of coal particle size. The molecular sieves are then regenerated by removing the trapped moisture and are recycled back through the process. It is important to note that the regeneration occurs after the deeply dewatered coal particles have been removed (i.e., no portion of the coal is ever subjected to heating). Consequently, this process is considered to be an advanced dewatering process and not a thermal drying process, which offers many advantages in terms of operational cost and environmental compliance.

Bench-scale Testing
A bench-scale experimental test program evaluated the performance of the process of the NDT system in removing water from fine coal. For all experimental tests, the wet feed sample consisted of either 0.6 mm or 0.15 mm top-size clean metallurgical coal (filter cake) collected from an industrial plant. During testing, a weighted sample of as-received fine feed coal was mixed with a predetermined weight of molecular sieves. The mixture was then contacted together in a small bench-scale rotary mixer for a defined period of time. After contacting, the mixture of molecular sieves and coal fines was separated by using a laboratory sieve. The dewatered coal particles passed through the sieve and were collected as an underflow product, whereas the molecular sieves were retained on top of the sieve and were collected as an overflow product. Once separated, the coal particles and molecular sieves were individually weighed and the reduction in the percentage moisture of the coal sample was calculated. The last step in the experimental procedure was drying the molecular sieves.

To speed the regeneration process, a microwave oven was used to evaporate the moisture held in the pores of the molecular sieves. The regenerated molecular sieves were then reused in the testing program. No significant difference was observed in the effectiveness of the moisture removal using either newly manufactured or regenerated molecular sieves.

Five independent “groups” of statistically designed bench-scale experiments were performed using the patent-pending process developed by NDT. The type (size) of molecular sieves and weight of coal sample was kept constant for each experimental group, while the weight of molecular sieves and time of contact were varied over a range of predetermined values as dictated by the statistical parametric test matrix. Duplicate test runs (a minimum of three to four) were conducted at each test point to assess the degree of variability and level of reproducibility in the test data.

A target moisture of 9% was selected with a range of 8% to 10% as the operating parameter for the NDT system. When 100 mesh x 0 product coal gets below 8% moisture, dust problems become a concern and, if dried further, explosion hazards must be considered. If the moisture is more than 10%, the potential blending benefits are reduced. Therefore, the tests were designed to determine whether the NDT process could produce a 9% moisture product with a 95% confidence level. It should be noted that maximum drying tests conducted during the bench- and pilot-scale testing showed that moisture levels in the 1.5% to 2.5% range could be easily produced if desired.

The data indicates that the technology can readily provide single-digit moistures over a wide range of operating conditions. In fact, moisture values in excess of 10% were only obtained when using very short contact times or when low weight fractions of molecular sieves were used.

Figure 1 shows the central composite text matrix used in the Group B test program on the 0.6 mm x 0 feed. A total of 39 individual test runs were performed in this group using type I media. The tests included nine different combinations of experiments based on contact time and media factor. The media factor is a dimensionless number representing the relative amounts of coal and molecular sieves used in the test run (i.e., a larger media factor represents a greater addition of feed coal relative to sieve weight, while a smaller number represents less coal relative to sieve weight). The central test conducted at 4 minutes of contact time and media factor of approximately 0.3 was randomly repeated 15 times throughout the test matrix to evaluate the statistical reproducibility of the process.

All but one of the test runs conducted for the Group B test matrix gave single-digit moistures in the final 0.6 mm x 0 product (See Figure 1). The product moistures decreased with either an increase in contact time or a decrease in media factor (i.e., less coal per unit weight of sieve media). The standard deviations for each set of conditions varied from a low of 0.01 to a high of 1.05, which indicated that the data was generally reproducible. In fact, the 15 replicate tests conducted at the central point of the test matrix (i.e., 4 minute contact time and 0.3 media factor) showed little variability in the product moisture despite significant variations in the feed moisture. The average moisture content for the feed sample used in the 15 central point tests was 21.8+0.16% with a standard deviation of 0.90. After contacting with the molecular sieves, the product moisture dropped to an average value of 8.90+0.02% with a standard deviation of 0.14. The very small confidence interval and low standard deviation values associated with the data obtained for the dewatered product indicates that a high degree of reproducibility can be achieved using the bench-scale version of the process of the NDT system.

A similar trend in moisture removal was observed for the tests conducted for Group E having a top-size of 0.15 mm. These experiments were conducted using type II molecular sieves over a similar range of contact times and a lower range of media factors. Each satellite test conducted around the central test point was repeated four times to assist in identifying outliers and evaluating reproducibility. The central test point, which involved a contact time of 3.5 minutes and media factor of -0.63, was repeated 20 times in random order throughout the test matrix.

For this particular group of tests, the average moisture contents of the as-received 0.15 mm x 0 feed was 26.2+0.10%. After contacting with the molecular sieves, the 0.15 mm x 0 product moistures were reduced to single-digit values for all tests conducted at contact times of 3.5 minutes or longer (See Figure 2). The lowest product moisture content of 6.38% was achieved for the longest contact time of 4.9 minutes. Tests conducted with contact times less than 3.5 minutes did not achieve single-digit moistures, but at 10.2% to 10.9% moisture were not far from breaking this meaningful barrier.

Pilot-scale Demo
In light of promising bench-scale data, a decision was made to construct a pilot-scale NDT plant to demonstrate the capabilities of this new patent-pending technology in continuous mode. While the small scale testing validated the basic system, numerous additional proprietary refinements were developed by NDT for operating on a larger scale.

The completed facility, which was largely assembled using off-the-shelf components, was designed with an effective throughput capacity of 0.5 tons per hour (tph). The self-contained facility included unit operations for handling, contacting and separating the coal and media. An advanced gas-fired dryer was used to regenerate the molecular sieves such that the entire process operated in a closed-circuit loop. The prototype facility was designed, constructed and successfully commissioned over a period of approximately 10 months. During this time, shakedown tests were completed and the process circuit was refined, modified and optimized using proprietary optimization techniques to provide a demonstration facility that operated smoothly and efficiently.

The prototype facility successfully achieved single-digit product moistures for a wide range of feed coal applications (See Table 2). Engineering criteria developed from benchscale testing, such as contacting (retention) times and coal-to-sieve loadings, were also validated using the pilot-scale plant. More importantly, the pilot-scale test runs successfully demonstrated that the molecular sieves could be regenerated and recycled back through the process without incurring significant losses due to media degradation and at a lower heating/evaporation cost than traditional thermal drying.

Removing unwanted moisture from fine coal has historically been considered one of the most challenging technical problems in the coal preparation industry. Experimental data collected from both bench- and pilot-scale operations show that the NDT process is capable of achieving single-digit moisture values from fine coal feeds containing 30% moisture or more. The process is highly flexible in that the product moisture can be “dialed in” by varying contacting time and coal-to-sieve media loadings. Also, unlike existing mechanical processes, the product moisture from the NDT system is largely independent of the particle size distribution of the feed stream.

The NDT process takes place at ambient temperature. As such, coal particles are never subjected to high temperatures, which greatly reduces the emissions concerns normally associated with conventional coal drying systems. Recent estimates by an environmental consulting group indicate that emission reductions as large as 90% or more when compared to a thermal dryer are possible using the NDT system.

Since the process treats only the fine coal fraction, which is generally between 10% to 15% of the total clean coal product (and not the entire clean coal product treated by conventional thermal dryers), the required footprint for the facility is only a fraction of that demanded by a large-scale coal thermal dryer. Also, due to fewer operational complexities, significant cost savings are also expected for ancillary items such as electricity, chemicals, maintenance and labor. Although such economic calculations tend to be site specific (See Figure 3), the costing figures for this site do suggest a relative operating cost of less than half of that required to operate a conventional thermal dryer.

References
Le Roux, M, Campbell, QP, Watermeyer MS and de Oliveira, S (2005). The Optimization of an Improved Method of Fine Coal Dewatering. Minerals Engineering, 18(9):931––934.

Osborne, D. G. (1988). Solid-Liquid Separation Coal Preparation Technology (Vol. 1, pp. 478-542). London; Boston: Graham & Trotman.

Orr, F. M. (2002). Coal Waste Impoundments: Risks, Responses and Alternatives. Washington, D. C.: National Research Council.

Wills, B.A and T.J. Napier-Munn. (2006) Wills Mineral Processing Technology, seventh edition, chapter 15

Orr, F. M. (2002). Coal Waste Impoundments: Risks, Responses and Alternatives. Washington, D. C.: National Research Council.

Ramakrishna, S. Ma, Z. and Matsuura, T. (2011) Polymer Membranes in Biotechnology, Imperial College Press, London. Chapter 1.

Breck, D.W. (1964) Crystalline Molecular Sieves, Journal of Education, Vol. 41, NO. 12 Pp. 678-689.

Bland, R.W. Harsh, P. Hurley, M. Jones, A.K. Vinod, K and Sikka. (2011) US Patent Application Publication, Pub No. US 2011/0078917 A1 Pub. Date Apr. 7, 2011.

Keles, S. (2010). Fine Coal Dewatering Using Hyperbaric Filer Centrifugtaion. PhD Dissertation submitted to Virginia Polytechnic Institute & State University Blacksburg, VA. Pp. 4-18.


Bratton, Ali and Luttrell are Virginia Tech professors who conducted the research. Bland and McDaniel are executives with Nano Drying Technologies, a company based in Beckley, W.Va., that developed the system. This article is based on a presentation made by Professor Luttrell at Coal Prep 2012, which was held during May in Lexington, Ky. For the full paper, visit www.coalprepshow.com.