From Coal Consumption to Ash Composition: Predictive Modeling for Sustainable Coal Fly Ash Management

Abstract

Coal combustion residuals (i.e., coal ash) are the solid wastes generated from burning coal for electricity and primarily comprise bottom ash, fly ash, and flue gas desulfurization sludges. Coal ash represents one of the largest solid waste streams in the United States. While approximately half of the coal ash generated in the U.S. is re-purposed for beneficial applications such as construction materials, the rest is discarded, mostly in unlined landfilled and ponds adjacent to coal fired power plants. Collectively disposal sites in the U.S. host more than 2 billion tons of coal ash that have accumulated over multiple decades of coal power generation. Coal ash disposal sites pose a threat to local water quality due to leaching of toxins and uncontrolled releases of ash residual resulting from impoundment failures. Despite their environmental risks, these disposal sites also represent an untapped resource for raw materials for supplementary cements and other infrastructure material applications. The chemical composition of coal ash (particularly fly ash and bottom ash) has not been routinely monitored in landfills and surface impoundments, resulting in insufficient knowledge to assess the reuse potential and toxicity of coal ash at these disposal sites. This dissertation addresses this knowledge gap by investigating the complex chemical composition of coal ash and evaluating its reuse potential. The research is structured into three chapters, each focusing on different aspects of the coal ash challenge, from the evaluation of the resource availability to the toxicity of this waste product. As a first step towards understanding the scope of legacy coal ash at disposal sites managed by power plants, one can first evaluate the various sources of coal that has been burned at power stations over the last decades and to use this information to assess the quantity and quality coal ash generated. To this end, Chapter 2 describes the compilation of a 50-year coal supply chain dataset for major U.S. power stations, integrating data on coal purchases, consumption, and regional supplies of coal described in a variety of databases hosted by U.S. federal agencies. Detailed records of coal purchases by power stations across the United States from 1973 to 2022 are included. Geographic inconsistencies of power stations were corrected by comparing coal purchase patterns across different energy databases. Discrepancies due to missing reports from power stations were addressed using a Compound Annual Growth Rate (CAGR) algorithm, resulting in a valuable database for future research on coal usage and coal combustion residuals reuse in the U.S. With this historical coal supply chain record, Chapter 3 details the development, testing and application of a model to predict the concentrations of major and minor elements in coal fly ash at each U.S. power station based on annual coal purchase records from power stations. This model, employed multivariate linear regression and Bayesian regression methods and was trained with chemical compositional data from the U.S. Geological Survey’s COALQUAL database and elemental contents that were measured in 696 coal ash samples from previous studies. The models indicate that the chemical composition of coal fly ash is significantly impacted by the coal source, rather than the year trend and other combustion parameters. The Bayesian model, utilizing prior knowledge of data from COALQUAL, effectively predicts fly ash elemental contents with relatively small sample sizes (<40 samples) and provides a a method to quantify uncertainties for the prediction historical fly ash composition produced annually at each major power station. This approach leverages prior knowledge of coal element distributions to estimate element concentrations in coal ash, providing guidelines for ash quality at disposal sites across the U.S. and the market potential for recycling efforts. Chapter 4 evaluates the toxicity of coal ash, focusing on the leachability of trace toxic elements like arsenic and selenium, which influence reuse potential. Leaching experiments with 52 fly ash samples from 15 different U.S. power plants representing coal feedstocks from three major domestic regions were conducted. The mobilization potential of As and Se in fly ash was assessed based on standardized leaching protocols. Multivariate and lasso regression analyses explored correlations between leachable As and Se contents and characteristics such as major element contents, loss on ignition, and pH. The results indicate that major elements (Fe, Ca, Al) can serve as predictor variables for the leaching potential of As, but not Se. Loss on ignition and pH were not significant predictive variables. The regression models showed relatively strong fits for leachable As (R2 = 0.78 for both models) compared to leachable Se (R2 = 0.49). These findings suggest that correlation models combined with on-site elemental analysis with portable analyzers could enable a screening method for leachable As in coal ash. Overall, this dissertation examined the recyclability, toxicity, and leachability of coal ash through the development of predictive models and protocols for assessing coal ash quality and leaching potential of toxins at disposal sites. The research contributes to a national dataset on the coal supply chain of power plants and coal ash residuals at disposal sites, thereby promoting sustainable management practices.