The results of proximate, ultimate, and petrographic analysis for a wide range of Kentucky coal samples were used to predict Free Swelling Index (FSI) using multivariable regression and Adaptive Neuro Fuzzy Inference System (ANFIS). Three different input sets: (a) moisture, ash, and volatile matter; (b) carbon, hydrogen, nitrogen, oxygen, sulfur, and mineral matter; and (c) group-maceral analysis, mineral matter, moisture, sulfur, and R max were applied for both methods. Non-linear regression achieved the correlation coefficients (R 2 ) of 0.38, 0.49, and 0.70 for input sets (a), (b), and (c), respectively. By using the same input sets, ANFIS predicted FSI with higher R 2 of 0.46, 0.82 and 0.95, respectively. Results show that input set (c) is the best predictor of FSI in both prediction methods, and ANFIS significantly can be used to predict FSI when regression results do not have appropriate accuracy.

This paper presents results of an interlaboratory exercise on organic matter optical maturity parameters using a natural maturation series comprised by three Devonian shale samples (Huron Member, Ohio Shale) from the Appalachian Basin, USA. This work was conducted by the Thermal Indices Working Group of the International Committee for Coal and Organic Petrology (ICCP) Commission II (Geological Applications of Organic Petrology). This study aimed to compare: 1. maturation predicted by different types of petrographic parameters (vitrinite reflectance and spectral fluorescence of telalginite), 2. reproducibility of the results for these maturation parameters obtained by different laboratories, and 3. improvements in the spectral fluorescence measurement obtained using modern detection systems in comparison with the results from historical round robin exercises. Mean random vitrinite reflectance measurements presented the highest level of reproducibility (group standard deviation 0.05) for low maturity and reproducibility diminished with increasing maturation (group standard deviation 0.12). Corrected fluorescence spectra, provided by 14 participants, showed a fair to good correspondence. Standard deviation of the mean values for spectral parameters was lowest for the low maturity sample but was also fairly low for higher maturity samples. A significant improvement in the reproducibility of corrected spectral fluorescence curves was obtained in the current exercise compared to a previous investigation of Toarcian organic matter spectra in a maturation series from the Paris Basin. This improvement is demonstrated by lower values of standard deviation and is interpreted to reflect better performance of newer photo-optical measuring systems. Fluorescence parameters measured here are in good agreement with vitrinite reflectance values for the least mature shale but indicate higher maturity than shown by vitrinite reflectance for the two more mature shales. This red shift in λ max beyond 0.65% vitrinite reflectance was also observed in studies of Devonian shale in other basins, suggesting that the accepted correlation for these two petrographic thermal maturity parameters needs to be re-evaluated. A good linear correlation between λ max and T max for this maturation series was observed and λ max 600 nm corresponds to T max of 440°C. Nevertheless if a larger set of Devonian samples is included, the correlation is polynomial with a jump in λ max ranging from 540 to 570 nm. Up to 440°C of T max , the λ max , mostly, reaches up to 500 nm; beyond a T max of 440°C, λ max is in the range of 580-600 nm. This relationship places the "red shift" when the onset of the oil window is reached at T max of 440°C. Moreover, the correlation between HI and λ max (r 2 = 0.70) shows a striking inflection and decrease in HI above a λ max of 600 nm, coincident with the approximate onset of hydrocarbon generation in these rocks.

Kentucky produced over 8 Mt of coal combustion products (CCPs) in 2006, with 30% of the CCPs being utilized, a significant increase from our 1996 and 2001 surveys. As much of the increase is related to increased utilization of flue-gas desulfurization (FGD) gypsum, the increased production of FGD gypsum coincident with the commissioning of new FGD units and the saturation of the (currently) weak market for new construction, the percentage of utilization may decrease by the time of the next planned survey .

Seven feed coals used in the Brazilian power generation industry were obtained and subsequently analysed together with fly ash and bottom ash from a major Brazilian power plant. The samples were investigated by means of room temperature Fe-Mö ssbauer analyses, X-ray diffraction, Raman spectroscopy, scanning electron microscope and petrographic analysis. In addition, nanometer-sized crystalline phases in coals and ashes were characterised using an energy-dispersive X-ray spectrometer and a high-resolution transmission electron microscope. The major identified Fe-bearing minerals in the coals were found to be actinolite, ankerite, chalcopyrite, chlorite, goethite, illite, ilmenite, magnesioferrite, natrojarosite, pyrite, pyrrhotite, and siderite; whilst in the fly ash and bottom ash, ankerite, chlorite, chromite, goethite, hematite, hercynite, jarosite, maghemite, magnesioferrite, and magnetite were identified. Most of the Fe in the ash samples was present as Fe 3+ resulting from the melting of Fe and silicates during combustion. The fraction of glassy Fe in those particles is high because of the high contact probability between Fe melt and silicates. The combination of the various methods offers a powerful analytical technique in the study of coal and coal ashes. This investigation can be regarded as an introductory and prospective study prior to further quantification.

Coal fires in natural outcrops, in abandoned and active coal mines, and in coal and coal-refuse piles are responsible for the uncontrolled emissions of gases, including CO, CO 2 , H 2 S, hydrocarbons, and volatile aromatics. Typically, measurements of gases at a mine vent are made over a short time interval, perhaps no more than 10 min, including the time for replicate measurements. Such timing provides little information on longer-term variations in emissions, although comparisons of seasonal measurements suggest such changes do occur. To address this problem, we placed temperature and CO data loggers in coal fire vents to collect time series measurements for a period of up to three weeks. For one experiment, 11 days of data at one-minute intervals indicated that the CO emissions were generally in the 400-550 ppmv range. However near daily depressions in CO concentrations occurred and in some cases fell below 50 ppmv; followed by an increase to~700-800 ppmv; in turn followed by a return to the ambient conditions. Data for a separate 21-day collection period in a different vent of the same fire exhibited similar trends, albeit at a higher CO concentration. The drop in CO concentration may be associated with a meteorologically-driven inhalation cycle of the fire, whereby air diluted the combustion generated CO. We propose this was followed by an increase in the intensity of the fire due to increased O 2 from the inhaled air, producing increased CO concentrations, before settling back to the ambient conditions.

Coal fires in eastern Kentucky are usually in abandoned mines. In many cases, the fires can be in close proximity to populated areas and can encroach upon the infrastructure. In addition, the gas emissions can be trapped in valleys at times of temperature inversions. The Truman Shepherd coal fire, Floyd County, Kentucky, is located in an abandoned mine adjacent to highway KY 80 and to the town of Howard. Building upon our previous studies of this fire from 2009 to 2012, in October 2013 we measured carbon monoxide emissions from a vent. Two methods of exponential smoothing and seasonal autoregressive integrated moving averages (SARIMA) were used to model CO emissions in the context of time series analysis. Due to the seasonality of the observations, the SARIMA model showed superior in both modeling and forecasting CO emissions with mean absolute percentage errors of 3.2% and 11.9%, respectively. Carbon monoxide emissions, on average, showed a gradual increase throughout the study period. Superimposed on the gradual CO increase is a daily sharp rise in CO, with the maximum peak exceeding 300 ppm CO. The increase in CO starts between 10:00 and 11:00 AM and generally peaks in the early afternoon, then declining until the next day's increase.

This paper presents data on the properties of coal and fly ash from two coal mines and two power plants that burn singlesource coal from two mines in Indiana. One mine is in the low-sulfur ( < 1%) Danville Coal Member of the Dugger Formation (Pennsylvanian) and the other mines the high-sulfur (>5%) Springfield Coal Member of the Petersburg Formation (Pennsylvanian). Both seams have comparable ash contents ( f 11%). Coals sampled at the mines (both raw and washed fractions) were analyzed for proximate/ultimate/sulfur forms/heating value, major oxides, trace elements and petrographic composition. The properties of fly ash from these coals reflect the properties of the feed coal, as well as local combustion and post-combustion conditions. Sulfur and spinel content, and As, Pb and Zn concentrations of the fly ash are the parameters that most closely reflect the properties of the source coal. D

Coal properties impact the quality of coal combustion by-products (CCBs). Tracking impacts can often be difficult, particularly in the eastern United States, because utilities use blended coal feeds to meet their quality specifications. To circumvent this problem, we made arrangements for a single seam/single mine coal to be burned at a 220-MW wall-fired boiler.

Markov Chain analysis was applied to the description of the megascopic lithologic transitions in Pennsylvanianage
eastern Kentucky coals. Coal lithology modeling can be problematic as individual lithotypes can represent
near-instantaneous events (vitrain), prolonged degradation (durain), or fire-induced loss of previously deposited
lithologies (fusain). Each of the latter lithotypes, potentially representing vastly different amounts of time, could
be of the same thickness. Therefore, equal thickness does not necessarily imply equal time. Probability transform
matrices that employ uniform lithotype thicknesses were used, allowing transitions between like lithotypes; embedded
Markov Chains, thereby only considering transitions between different lithotypes; and continuous-time
Markov Chains were employed in the assessment of a section of the No. 5 Block coal (Pennsylvanian Breathitt
Group, Martin County, Kentucky). Embedded Markov Chains could successfully simulate the lithologic transitions.
A Monte Carlo random process was programmed to simulate thickness variations of lithotypes between
the transitions. The proposed hybridmodel ofMonte Carlo–Markov Chain was able to predict the randompattern
that underlies lithotypes transitions and thickness. The hybrid Monte Carlo–Markov Chain technique proved to
be effective in the case study in simulating both the lithologic thickness variations and transitions.