Time Series Analysis of Coal Data from Preparation Plants

The body of information presented in this paper is directed to those individuals concerned with existing or proposed sulfur emission regulations. The ability of coal-fired boiler operators to comply with emission regulations depends largely on the variability in emission rate. Because of the variability inherent in coal, the mean emission rate (in Ib SO2/million Btu) must be lower than the emission standard to prevent violations. Heretofore, this problem has been analyzed by statistical techniques assuming independent (uncorrelated) data. The results of this paper indicate that this method is inappropriate. Failure to account for correlation structure in coal data results in gross underestimation of emission violations of actual coal. It is concluded that correlation in coal data should not be overlooked, and that the ability of each coal to meet emission regulations must be individually determined by procedures outlined in this paper.