Emissions of NOx , PM , SO 2 , and VOCs from Coal-Fired Boilers Related to Coal Washing , 1 Iron-Steel Production , and Lime and Gypsum Making in Shanxi , China 2

The accurate pollutant inventories are important for the development of pollution control policies, which further rely on detailed emission factors (EFs) to some extent. However, detailed air pollutant EFs for coal-fired boilers (CFBs) associated with coal washing (CW), iron-steel production (IS), and lime and gypsum manufacturing (LG) are lacking in China at present. CFBs of 91 enterprises involving CW, IS, and LG were investigated to obtain their pollutant EFs associated with coal consumption (EFI, kg t–1), outputs (EFII, kg MY–1), and product yields (EFIII, kg t–1) through field investigation and sampling. The weak correlation between EFs of 4 air pollutants vs. corresponding removal efficiencies (REs), and EFs vs. coal compositions among three industries implied the impact of actual combustion conditions and operating status of removal facilities (RFs). EFs of VOCs from small-scale CW enterprises (SSEs) were much higher than those of large- and medium-scale enterprises (LSEs and MSEs) owning to the incomplete combustion of coal. Also the SO2 and NOx EFs of CW increased with decreasing enterprise scale, while the maximum PM occurred at MSEs. The mean EFI values of LG for the 4 air pollutants was PM > NOx > VOCs > SO2, differed from PM > SO2 > NOx for the IS, VOCs > PM > NOx > SO2 for the CW LSEs and MSEs, and VOCs > NOx > PM > SO2 for the CW SSEs, which suggested the influence of combined factors including coal composition, production processes, combustion conditions, and pollutant removal technologies and removal efficiencies. EFI values for the 8 IS factories followed the order PM > SO2 > NOx, while they were PM > NOx > SO2 for EFII values due to their output fluctuation. For the EFII and EFIII values of SO2, NOx, and PM, LG dominated within the 3 industries, while the corresponding maximum VOCs occurred at the CW industry.


A C C E P T E D M A N U S C R I P T
Shanxi and Inner Mongolia are two predominant coal districts in China, whose coal production

A C C E P T E D M A N U S C R I P T
5 get the information related to pollutants actually released into atmosphere. The additional 1 parameters such as temperature and flow velocity (m s −1 ) of flue gas were also measured and 2 provided by this gas analyzer. PM was also sampled by this analyzer and the collected mass divided 3 by the corresponding sampling volume of flue gas was used to represent PM concentration. The 4 analyzer was calibrated with zero gas and standard gases (NO x , SO 2 and O 2 ) before measurement 5 for the elimination of possible interferences. The mass emissions of SO 2 , NO x , and VOCs were 6 calculated as their concentrations multiplied by sampling volumes of flue gas.

7
Two pathways including request from enterprises and field sampling and subsequent laboratory 8 measurement were used to obtain the information about coal components for all the enterprises. The with contents of ash, sulfur, carbon, hydrogen, nitrogen, oxygen, and water. (EF I , kg t -1 ), output (EF II reported in kg MY -1 ), and product yield (EF III , kg t -1 ) associated EFs were 36 provided and calculated using the Eqs.

A C C E P T E D M A N U S C R I P T
Where C is mass concentrations of gaseous pollutants, V FA is the actual flue gas volume derives 1 from 1 kg coal burning, which can be induced by V AT . 10 V FT is theoretically generated flue gas volume from 1 kg coal combustion and calculated by Eq.

12
Where V CO2 , V SO2 , and V NO2 refer to volumes of CO 2 , SO 2 , and NO 2 derived from burning of C,

15
H, and N in coal, V N2 is the N 2 volume in V AT and equal to 0.79V AT , and V H2O is the water vapor 16 volume sum of coal containing H burning (0.112ω (H ar )), vaporization of coal containing water 17 (0.00124ω(M ar )), and vapor in air (0.0161V AT ).

18
Finally, V AT derived from 1 kg coal combustion is calculated by Eq. (7).
Where α is the excess air coefficient, which is provided by corresponding enterprise.  The RFs applied in coal washing industries should be further improved when all the present RFs 32 were taken into account.
( Fig. 1)  be explained by the fluctuation of RFs. The mean value of SO 2 EF I of F6, and F19 without SO 2 RFs 7 equipped was 2.80 kg t -1 , which was higher than the mean value of 2.73 kg t -1 for the rest 17 In this study, a total of 37 medium scale enterprises associated with coal washing were 37 investigated and field measured. Considering the pollutant RFs, only 2 of 37 MSEs were equipped 38 with NO x RFs with REs as 59.5% and 15.0%, 8 MSEs were not equipped with SO 2 RFs, and all the

A C C E P T E D M A N U S C R I P T
9 37 MSEs were not installed with VOCs RFs and were equipped with PM RFs. Fig. 2  (0.03±0.04) >SO 2 (0.02±0.02) for EF III (in kg t -1 ), respectively (Fig. 4) RFs in all the 37 MSEs, which ranged from 0.21 kg t -1 of F43 to 252 kg t -1 of F39, 3.9 kg MY -1 of 10 F56 to 14300 kg MY -1 of F51, and 1.00×10 -3 kg t -1 of F56 to 1.56 kg t -1 of F46 (Fig. 2) MSEs were possibly attributed to the unstable sale prices resulted from the small enterprise size. 5.31, and 2.97 kg t -1 ), EF II (2280, 867, 370, and 262 kg MY -1 ), and EF III (0.50, 0.14, 0.07, and 0.04 7 kg t -1 ), respectively (Fig. 4) low output and no installation of any RFs (Fig. 3). Due to the limited experimental condition, VOCs field sampling was not conducted for coal-fired 40 boilers applied in 8 iron-steel (IS) enterprises. All the 8 IS factories were equipped with PM RFs

A C C E P T E D M A N U S C R I P T
and not equipped with NO x EFs, and SO 2 RFs were partly installed in 5 of 8 factories ( and actual combustion condition and running status of RFs (Fig. 5). The highest EF I (73.0 kg t -1 ) of 6 the sum of PM, NO x , and SO 2 occurred at I5, while the corresponding values for EF II and EF III were 7 possessed by I4 (22900 kg MY -1 ) and I3 (11.8 kg t -1 ), respectively (Fig. 5).  The SO 2 originated from 5 enterprises with SO 2 RFs installed possessed lower mean EF I of 9.74 37 than 9.90 of the rest 3 ones without SO 2 RFs, while higher EF II and EF III values occurred at 5

A C C E P T E D M A N U S C R I P T
factories owning RFs compared with those of the rest 3 factories without RFs resulted from the 1 fluctuations of outputs and product yields.  37) (Fig. 7). The weak correlation between EFs of air pollutants and 25 corresponding RFs indicated the impact of actual running status of RFs and combustion of boilers.

26
The mean values of EF I for 4 air pollutants followed the order as PM (  should be optimized to obtain the enhanced PM removal efficiencies.   Table 2.

A C C E P T E D M A N U S C R I P T
Statistical values for medium scale coal washing enterprises Table 3.