By Z. Ali, R. Bratton and G. Luttrell, Virginia Tech; M. Mohanty, Southern Illinois University Carbondale;
A. Dynys and L. Watters, Taggart Global; and C. Stanley, Knight Hawk Coal

Modern processing plants incorporate a complex array of solid-solid and solid-liquid separation processes that remove unwanted impurities from run-of-mine coals. The most popular processes used for coal cleaning include dense medium separation, gravity concentration and froth flotation. Unfortunately, the processes traditionally used to upgrade the finest fractions of coal are generally inferior in terms of separation performance. In particular, these circuits often perform very poorly in reducing sulfur levels in the fine coal products.

Although coal pyrite particles are often well liberated at particle sizes down to nominal flotation feed sizes (minus 0.15 mm), this unwanted mineral often reports to the froth product due to entrainment and/or the inherent hydrophobic nature of coal pyrite. Thus, froth flotation, which otherwise provides excellent ash separation performance, often performs poorly in terms of sulfur rejection. In fact, sulfur contents from industrial froth flotation circuits are often found to be higher in sulfur than the original feed due to the flotability of pyrite and the removal of other low-sulfur minerals from the floated product.

One of the most interesting approaches for enhancing the removals of both ash- and sulfur-bearing mineral matter from coal is the multiproperty processing strategy developed and patented at Virginia Tech in the 1990s. In this approach, particles containing both valuable components and undesirable impurities are passed through two or more processes, each of which separate based on a different material property. Common properties that may be exploited include size, shape, density or surface wettability/hydrophobicity. By using more than one property, a significantly broader range of impurities can be removed at a higher overall separation efficiency.

In the 1990s, this unusual concept was demonstrated using a pilot-scale coal cleaning circuit that combined column flotation with an enhanced gravity concentrator. Flotation is a surface-based process that is effective in removing ash-forming minerals such as clay, while enhanced gravity concentration is a density-based process that is more efficient in rejecting high-density minerals and middlings that contain pyrite. By combining two processes into a single circuit, high rejections of both ash-forming minerals and pyrite were obtained as compared to conventional single-stage coal cleaning processes. For some eastern U.S. coals, the two-stage circuitry was shown to be capable of doubling the rejection of mineral matter and pyritic sulfur with minimal loss of heating value.

Unfortunately, the multiproperty processing concept failed to be commercially accepted due to a variety of operational issues. One of the greatest barriers to industrial implementation was the inability of commercially available density separators to handle the large volumetric flows of flotation-feed slurry. This issue was particularly serious for enhanced gravity separators that typically have low throughput capacities. Reversing the cleaning stages, i.e., placing the density separator after flotation, was also unsuccessful due to the presence of residual air-bubble aggregates and uncontrollable froth handling problems.

In addition, the conservative nature of the highly competitive coal industry made it difficult for operators to invest in the new and unproven enhanced gravity separation technologies. Consequently, the multiproperty processing strategy was never implemented commercially and the associated patent was allowed to expire.

Figure 1: Simplified schematic of the original flowsheet layout used at the Prairie Eagle Prep Plant.

During the past decade, research and development programs instigated by Taggart Global have created re-newed interest in developing and designing high-efficiency multiproperty processing circuits for the coal industry. The specific goals of this initiative have been to extensively evaluate the ash and sulfur partitioning performance of different fine coal sizing and cleaning operations, to use this fundamental insight in the design of fine coal circuitry that provides the highest coal recoveries and mineral rejections, and to demonstrate the improved circuitry at a full-scale level at an industrial coal preparation plant.

In 2009, Taggart Global commissioned its academic partners to investigate improved processing strategies for fine coal processing at the Prairie Eagle preparation plant. The university facilities were used to conduct laboratory and pilot-scale test work necessary to fully delineate the extent of improvements that could be realistically achieved at this site in a cost-effective manner. This article presents some of the important data and key findings obtained from this investigation.

The fine coal optimization tests commissioned by Taggart Global were conducted using samples of fine coal (<1 mm) collected from the Prairie Eagle prep plant. The plant complex is located in Perry County near Cutler, Ill. The facility, which is operated by Knight Hawk Coal, was built in 2005 to upgrade coal from surface, highwall and underground mining operations. It currently produces thermal coals containing 10% ash, 2.8%-3.2% sulfur and 11,200 Btu/lb. The plant has a nameplate feed capacity of 500 tons per hour (tph) and produces approximately 2.7 million tons of clean coal annually.

The coarse coal circuitry at the Prairie Eagle prep plant incorporates a dense medium cyclone to upgrade particles larger than 1 mm in diameter. The original fine coal circuitry, which is shown in Figure 1, was used to clean raw feed coals in the 1 x 0.15 mm size class using a combination of water-only cyclones (380 mm diameter) and secondary scavenging spirals. The coal products from the cleaning units are initially dewatered using clean coal classifying cyclones (380 mm diameter) and fine wire sieves (0.35 mm aperture) and then fully dewatered using a centrifugal dryer (EBW-40). Particles finer than 0.15 mm from the clean coal classifying cyclones are discarded as waste along with the underflow streams from the fine wire sieves and the refuse high-frequency screen.

The plant flowsheet also incorporated two recycle streams, i.e., the effluent from the centrifugal dryer is recycled back to the clean coal classifying cyclones and the spiral middlings are recycled back to the water-only cyclones. These recycle streams are needed to avoid excessive coal losses that would otherwise occur if these streams were discarded.

For the Prairie Eagle operations to respond to increased market demands, Taggart Global was contracted to design a plant expansion that would increase capacity up to 750 tph. The plant expansion also provided an opportunity for the engineering firm to work with company management in developing more efficient circuitry for processing fine coal. Due to market demands, there was considerable interest in identifying new equipment solutions that would allow the operation to recover additional fine coal without adversely impacting the clean coal ash and sulfur levels. In particular, considerable interest was expressed in adding a froth flotation circuit to the plant flowsheet.

Classification Testing
The use of a conventional “by-zero” froth flotation circuit was not considered to be a viable option for the Prairie Eagle prep plant because of the extreme fineness (>85% minus 325 mesh) and poor quality (>60% ash) of the minus 100 mesh feed. The best alternative was considered to be a flotation circuit incorporating deslime classifying cyclones to discard the minus 325 mesh fraction followed by column flotation to upgrade the nominal 100- x 325-mesh cyclone underflow. To examine the potential benefits of this alternative, a sample of minus 100 mesh overflow was acquired from the 15-in. diameter clean coal classifying cyclones currently operating in the plant. The sample was shipped to the laboratory and passed through a 6-in. diameter deslime cylone to determine the effects of classification on ash and sulfur partitioning.

The data from the deslime cyclone test results was promising in that the overflow represented 63.4% of the feed weight and contained 71.8% ash. Very little (roughly 2.6%) material in the 100- x 325-mesh size fraction reported to the overflow. The data also indicated very little difference in sulfur contents of the overflow and underflow products for sizes larger than 325 mesh. This result was unexpected since it was assumed that sulfur levels in the underflow would be higher due to preferential partitioning of high-density pyrite to the underflow stream. Surprisingly, the plus 325 mesh sulfur contents ranged from 2.20%-2.45% sulfur, which were well below the plant contract specification of 2.8%-3.2% sulfur.

Upgrading of Clean Coal Classifying Cyclone Overflow
The feed, overflow and underflow products from the raw coal deslime cyclone were subjected to laboratory froth flotation tests. The test work included both kinetic tests and release analysis tests to evaluate the practical and ultimate cleanability of each sample. The test results from the release analysis tests showed that products containing less than 10% ash could be theoretically obtained from any of the three process streams. However, high recoveries were difficult to obtain for either the feed or overflow samples due to the very high feed ash contents of these samples. In contrast, combustible recoveries approaching 90% were attainable for the deslimed underflow. The only apparent disadvantage of processing the underflow sample was that slightly higher sulfur values were obtained, perhaps due to the comparatively high sulfur content of the underflow feed sample.

Data from the kinetics tests were found to be substantially inferior to those obtained from release analysis testing. This large difference suggests that column flotation with froth washing to minimize entrainment would be the preferred option for treating any of these process streams. Based on this data, column flotation would be expected to produce a froth product containing about 7%-8% ash, while conventional flotation would be expected to result in a froth product containing 17%-18% ash. Both methods produced clean coal products containing about, 2.5% sulfur, which as indicated previously well below the sulfur was recovered by flotation as a result of either poor liberation or pyrite floatability and not due to hydraulic entrainment in the froth water.

Upgrading of Clean Coal Sieve Underflow
The experimental results obtained from the initial round of classification and flotation tests indicated that deslime column flotation would be an ideal approach for upgrading the overflow from the clean coal classifying cyclones currently installed in the Prairie Eagle prep plant. Further investigations indicated that flotation performed well since most of the high-density pyritic sulfur in the minus 100 mesh-feed slurry had already been captured in the underflow of the plant’s clean coal classifyng cyclones. The cyclone underflow then reports to fine wire sieves where the sulfur-enriched fines eventually passed through as effluent. Sampling and size-by-size analysis of the sieve underflow showed that the effluent contained 39.8% ash and 5.6% sulfur, with the greatest proportions of the sulfur occuring in the sizes less than 100 mesh. This created a dilemma for plant designers since the sieve effluent contained too much valuable coal to discard despite being highly enriched in both ash and sulfur.

Figure 2
Figure 2: Simplified schematic of the upgraded flowsheet layout used at the Prairie Eagle prep plant.

Several additional series of coal cleaning tests were conducted to determine how to best process the material present in the clean coal sieve underflow. In this round of tests, two different processes were considered, i.e., column flotation and fine spirals. In addition, a multiproperty processing circuit consisting of density-based spirals followed by surface-based flotation was also evaluated for this unique application. This circuitry was considered to be a viable option since the upstream clean coal classifying cyclones and fine wire sieves had created a manageable low-volume high-impurity stream that could be easily handled by a multiproperty processing circuit.

The flotation test work was conducted using both laboratory and pilot-scale units, while the spiral tests were conducted using a full-scale single-start compound spiral rig. The spiral rig was equipped with a partitioned product launder that allowed six products to be collected across the spiral profile in addition to a primary reject. Six separate sets of fine spiral tests were conducted to identify the optimum conditions (i.e., dry solids mass rate and slurry volume flow rate) for separating the nominal 100- x 325-mesh material from the sieve underflow.

The data (combustible recovery vs. ash and sulfur contents) obtained from the flotation, spiral and flotation/spiral test runs effectively demonstrates that froth flotation was much superior to the fine spirals in removing ash-forming minerals. Flotation effectively reduced the ash of the feed from about 43% to about 15%-18% ash, while the fine spirals never achieved ash levels lower than about 34% ash. While much of this problem can be attributed to the inability of the spiral to deal with misplaced ultrafines (minus 325 mesh solids), size-by-size analyses of the test data showed that froth flotation was superior to the fine spirals even for the coarsest size fraction of 100- x 200-mesh present in the feed.

On the other hand, fine spirals were much superior to flotation for reducing sulfur levels, often rejecting solids containing double-digit sulfur contents. The spiral unit lowered the feed sulfur from about 5.6% to about 4.2% sulfur. Flotation was not able to match this level of performance and, in fact, actually concentrated the sulfur-bearing components into the clean coal product. Consequently, most of the froth products from flotation testing had higher sulfur values than the original feed.

The most interesting and most promising results were obtained from the testing of the combined fine spiral and flotation circuit. The combined circuit was able to produce combustible recoveries in the range of 80%-90% while reducing the total ash content into the 10%-11% range. Moreover, the combined circuit also attained the best reductions in sulfur. In this particular case, sulfurs under 4% were obtained while keeping combustible recoveries in the 85%-90% range. The improvement in performance achieved by the combined circuit is best illustrated by plotting the product ash and sulfur values obtained from the flotation, spiral and combined test runs on a single graph. The format clearly illustrates the inherent capabilities of the different processes in dealing with ash- and sulfur-forming minerals and the synergistic effect of the two-stages combined in maximizing the purity of the final clean coal product.

Industrial Demonstration
In light of the promising data obtained from test programs, a new plant flowsheet was designed by engineers at Taggart Global. A simplified schematic of the upgraded flowsheet is provided in Figure 2. The new circuitry continued to use the existing water-only cyclones and clean coal classifying cyclones for their primary functions of cleaning and sizing coal. In addition, the sulfur partitioning capabilities of these units were exploited to ensure that nearly all of the high density pyritic sulfur was captured in the cyclone underflows.

The overflow from the clean coal cyclones, which was now depleted of pyritic sulfur, was then passed to a bank of deslime cylones where high-ash ultrafines were rejected as overflow. The low-sulfur underflow was then upgraded by column flotation cells, which used froth washing to minimize the recovery of ultrafine mineral impurities, such as clays, misplaced in the water reporting to the underflow of the deslime cyclones.

The pyrite-enriched underflow from the clean coal classifying cyclones was permitted to flow to the fine wire sieves and pass as an underflow effluent into a newly installed bank of fine desulfurization spirals. The high-sulfur reject from the spirals was allowed to pass to the thickener for disposal, while the pyrite-depleted fines passed to column flotation along with the overflow from the clean coal classifying cyclones.

One additional upgrade to the plant was the installation of a screenbowl centrifuge to maximize the amount of moisture removed from the fine clean coal products. Simulations indicated that the improved flowsheet would provide an additional 10 tons of clean coal from the flotation circuit at ash and sulfur levels below 10% and 3%, respectively.

Plant sampling campaigns conducted by Taggart Global shortly after plant commissioning indicated the new flowsheet did meet expectations. The new fine coal flotation column provided approximately 9 tph of additional clean coal at ash and sulfur contents of 10.8% and 3.1%, respectively. A much more extensive sampling of the new circuitry was undertaken now that the plant operations have been fully optimized; however, the laboratory analysis of the samples was not completed at the time this article was written. The results obtained to date indicate that the two-stage circuit has performed extremely well and that all projected targets have been successfully met in terms of both technical performance and financial returns.

This article was adapted from a paper the authors presented at Coal Prep 2013, which took place during late April in Lexington, Ky. To see the original work, visit: