Unit-based emission inventory and uncertainty assessment of coal-fired power plants

A unit-based emission inventory of coal-fired power plants in China was developed which contains unit capacity, coal consumption, emission control technology and geographical location. Estimated total emissions of SO2, NOx, particulate matter (PM) and PM2.5 in 2011 were 7251 kt, 8067 kt, 1433 kt and 622 kt, respectively. Units larger than 300 MW consumed 75% coal, while emitting 46% SO2, 58% NOx, 55% PM and 63.2% PM2.5. Emission comparisons between key regions such as the Yangtze River Delta, the Pearl River Delta and Shandong Province showed a general downward trend from 2005 to 2011, mainly because of the growing application ratio of desulphurisation, LNBs, denitration and dust-removal facilities. The uncertainties at unit level of SO2, NOx, PM and PM2.5 were estimated to be -10.1% similar to +5.4%, -2.1% similar to +4.6%, -5.7% similar to +6.9% and -4.3% similar to +6.5%, respectively. Meanwhile sector-based Monte Carlo simulation was conducted for better understanding of the uncertainties. Unit-based simulation yielded narrowed estimates of uncertainties, possibly caused by the neglected diversity of emission characteristics in sector-based simulation. The large number of plants narrowed unit-based uncertainties as large uncertainties were found in provinces with a small number of power plants, such as Qinghai. However, sector-based uncertainty analysis well depends on detailed source classification, because small NO uncertainties were found in Shandong due to the detailed classification of NO emission factors. The main uncertainty sources are discussed in the sensitivity analysis, which identifies specific needs in data investigation and field measures to improve them. Though unit-based Monte Carlo greatly narrowed uncertainties, the possibility of underestimated uncertainties at unit level cannot be ignored as the correlation of emission factors between units in the same source category was neglected. (C) 2014 Published by Elsevier Ltd.