| Abstract: | This paper presents results of multivariate regression models developed to estimate the properties and cost of U.S. coals washed for varying degrees of sulfur removal using commercially available physical coal preparation processes. The models allow washed coal characteristics to be predicted from information on coal origin, heating value, ash, and sulfur content. The models were developed by first "processing" each of the 710 coals in the U.S. Bureau of Mines (USBM) coal washability data base through a coal preparation plant computer model which optimizes plant performance to achieve a desired washed coal quality. Washability data are adjusted to account for the inefficiencies of coal washing equipment, and the actual coal sizes treated by various plant wash streams. Since different plant designs may be capable of achieving a given level of sulfur removal, three nominal levels of plant complexity (Levels 2, 3, 4) were included to identify the most economical alternative. The washed coal characteristics thus derived were then analyzed using standard statistical techniques to develop regression equations linking washed coal properties and cost to raw coal properties for each of 18 geographical regions encompassing the entire U.S. These regression models are incorporated in the Advanced Utility Simulation Model (AUSM) to estimate the economic potential of coal washing as a sulfur abatement strategy, in conjunction with other options available to coal-fired power plants. Modeling results for Pennsylvania showed that washed coals frequently were selected as part of a cost-effective control strategy, accounting for 10 to 30 percent of the total emissions reduction, and that "local coal" restrictions significantly increase the use of washed coal as an SO2 control strategy. Hypothetical requirements for mandatory coal cleaning, however, were found to be costly and ineffective. |
