Emission Factors of NOx, SO2, and PM for Bathing, Heating, Power Generation, Coking, and Cement Industries in Shanxi, China: Based on Field Measurement

Despite rapid development in China, small-scale boilers (SCBs) still occupy a prominent place in industry. Due to the lack of pollutant removal devices (RDs), SCBs emit large quantities of pollutants, which merit increased attention. In this study, various SCBs (operating on coal, gangue, coke oven gas, coal gas, and natural gas) used in bathing, heating, power generation, and coke and cement making were investigated for their SO2, NOx, and PM emission factors (EFs). The EFs were expressed as the emitted pollutant mass associated with fuel consumption (EFI), product yield (EFII), industrial output (EFIII), and power generation (EFIV). Of 17 civil SCBs, 4, 14, and 10 were not equipped with PM, NOx, and SO2 RDs, respectively. Generally, the EFI values for all of the SCBs decreased with increasing coal consumption. The averaged NOx EFI value for the 3 SCBs with installed NOx RDs was 2.00 kg t versus 3.16 kg t for the 17 SCBs. The sulfur content of the coal and the SO2 removal rate were highly influential factors for the SO2 EFI values. The 4 SCBs without PM RDs possessed an average EFI value of 23.9 kg t, which was higher than the corresponding 5.41 kg t for the 13 boilers equipped with PM RDs. The EFI, EFII, and EFIV values for 9 coal-fired power plants (PPs) exhibited the same trends, decreasing as the capacity of the PPs increased from 6 to 330 MW, although slightly higher EFs were found for 600 MW plants compared to 330 MW plants. The gas-fired PPs possessed higher NOx EFs than both the coaland gangue-fired plants, and the gangue-fired PPs displayed significantly higher EFs than coal-fired PPs with the same individual block power capacity. Because flue gas produced in the coking factories was not fully emitted during the combustion process, no correlation existed between the EFs (expressed as EFII and EFIII) and coke production or industrial output. Moreover, due to the lack of NOx RDs, the EFs of NOx were higher than those of SO2 and PM in the coking industry. Among 6 smalland medium-sized cement companies, the factories with lower cement production possessed higher EFI values for PM. A reverse trend was exhibited by the NOx EFI, however, with high combustion temperatures at factories with high production being the possible explanation.


INTRODUCTION
Atmospheric quality deterioration has occurred in Asian, European and North American cities in recent years, and especially in some rapidly developing regions and countries (e.g., China) (Fang et al., 2009;Pascal et al., 2013;Kiros et al., 2016;Li et al., 2017).The focus on Chinese pollutant emissions is still rising because various emission sources, emission intensities, and spatial and temporal emission patterns are contained in this country (Zhao et al., 2017).
Emission inventory is a key factor to the atmospheric science research and policy making on pollution control.The accuracy of emission inventory depends largely on accurate EFs.Although a series of pollution control measures has been conducted to improve air quality and protect human health in China, SO 2 , NO x , and PM are still the primary air pollutants (Zhong et al., 2017).SO 2 has a harmful effect to human health, leads to acid rain and the increase of PM (Annamalai et al., 2016).NO x (NO or NO 2 ) plays a key role in atmospheric chemistry, it can cause depletion of stratospheric O 3 , formation of acid rain and organic aerosols, and results in seriously adverse health effects such as respiratory effects, cardiovascular effects, lung cancer, and mortality (Yao et al., 2015;Xu et al., 2017a;Yan et al., 2017;Wang et al., 2018).PM, especially fine PM as PM 2.5 and PM 10 , not only poses a serious threat to human health but also scatters and adsorbs the incident light, and results in atmospheric opacity and horizontal visibility reduction (Mari et al., 2016;Ma et al., 2017b).
SO 2 , NO x , and PM are always formed and emitted from fossil and biomass burning process (Liu et al., 2017).Industrial boilers for production of the glass, paper, plastic, cement, and chemicals are the main sources of industrial SO 2 , NO x , and PM (Zhang et al., 2015;Wang and Chen, 2016).Compared with industrial boilers, an unknown number of civil boilers have been the other important SO 2 , NO x , and PM sources due to the lack of pollutant removal facilities.Although the SO 2 and NO x emission control has been an important policy in China since 2010 and 2014, the emissions of civil boilers have not been effectively controlled due to the high cost (Zhang et al., 2015).
Residential stationary sources including stoves, masonry heaters, and small-scale boilers for heating and bathing are one of the major emission sources of SO 2 , NO x , and PM due to the lack of pollutant control devices (Horák et al., 2018).The power generation industry fueled with coal, gangue, coke oven gas, natural gas, and coal gas is a major anthropogenic SO 2 , NO x and PM source due to the combustion of fuel containing sulfur and nitrogen (Li et al., 2016(Li et al., , 2017;;Yan et al., 2017).The coal-fired power plants (PPs) contribute approximately half of the total coal consumption in China and emit a large amount of air pollutants regionally and nationwide (Chen et al., 2014).Previous studies carried out were mainly focused on the large-scale coal-fired PPs, few studies were conducted on small-scale gangue-and gas-fired PPs (Li et al., 2016;Dodla et al., 2017;Li et al., 2017).Yan et al. (2017) reported higher NO x , SO 2 , and PM EFs were possessed by the gasfired boilers compared with the coal-fired ones.As the world's largest coke producer, China contributed 68.6% to world's total coke production (447.78Mt) in 2015 (Wang et al., 2018).Due to the low energy utilization rate, complex process flows, abundant pollutant production links, coke manufacturing would be an important source of SO 2 , NO x , and PM (Huo et al., 2012).In China, more than 80% of coking companies are independent small-and medium-scale enterprises distributed loosely across the nation (CCIA, 2016).Previous studies were mainly focused on the POP emissions from the processes such as charging coal, pushing coke, and combustion of coke oven gas based on the data of one or few coke plants (Liu et al., 2009(Liu et al., , 2013;;Mu et al., 2013;Saikia et al., 2015;Mu et al., 2017).China has the largest cement production in world and accounts for 60% the total world production in 2012 due to its fast urbanization (Chen et al., 2015a).The total coal and electricity consumptions of Chinese cement industry in 2009 were 1.87 × 10 8 t and 1.38 × 10 9 kWh, resulted in the emission of 8.9 × 10 5 t of SO 2 , 1.69 × 10 6 t of NO x , and 3.58 × 10 6 t of PM (Mao et al., 2012;Pang et al., 2013).
A variety of expression methods of emission factors (EFs) have been documented elsewhere (Yao et al., 2015;Yang et al., 2017;Hsieh et al., 2018).Hsieh et al. (2018) used the emitted pollutant mass per combusted fuel and generated electricity to calculate the EFs of PCDD/Fs for a municipal waste-fired PP.Yao et al. (2015) reported the NO x EFs as pollutant mass per kilometer (g km -1 ) for China III and IV in-use diesel trucks.Fachinger et al. (2017) discussed the EFs of PM 1 , PM 2.5 , and PM 10 using pollutant mass per calorific value as mg MJ -1 under the influence of burning conditions and fuel types.
To our knowledge, few systematic and integrated EFs for SO 2 , NO x , and PM compiled for boilers with different purposes, fuels, produced outputs, scales, and generated power are currently available.In this study, the establishing of category-specific EFs of SO 2 , NO x , and PM were obtained by field measurements for various companies and the industries involved in small-scale civil heating and bathing, power generation with different fuels and values of individual block power capacity (IBPC), and small-and medium-sized coke and cement manufacturing plants.

METHODS
In this study, 17 civil small-scale boilers (SCBs) for heating, bathing, and production; 13 coking companies; 15 power plants (PPs; 9 coal-fired PPs, 2 gangue-fired PPs, 1 coke oven gas-fired PP, 1 natural gas-fired PP, and 2 coal gas-fired PPs); and 8 cement making companies were field investigated and measured for their SO 2 , NO x , and PM EFs.All the factories were located at Shanxi Province, China.
A flue gas analyzer (Laoying-3012H, Qingdao Laoying Environmental Science and Technology, Co., Ltd.) was used for analysis of the flue gas temperature, concentrations of gaseous pollutants such as SO 2 , NO x , and CO, at the outlet of flue gas after the pollutant control devices in order to investigate the actual pollutants entered into atmospheric environment.At the same time, a pilot tube was applied to the measurement of velocity of flue gas and collection of particle matter (PM) in flue gas.The PM mass divided by sampling flue gas volume was the PM mass concentration.For quality assurance and quality control, the gas analyzers were calibrated with zero gas and targeted standard gases (NO, NO 2 , SO 2 and O 2 ) prior to the first test of the day.
The coal compositions associated with proximate and ultimate analysis were obtained from field measurement or provided by local factories.The proximate and ultimate analysis of coal was conducted based on Chinese standard methods of GB/T-212-2008 and GB/T 476-2001.

Expression Mode of Emission Factor and Calculation Method
The emission amounts of SO 2 , NO x and PM were calculated as their mass concentration multiply by the total flue gas volume.The emission factors (EFs) were expressed in four ways and calculated using Eqs.(1)-(4).EF I (kg t -1 ) was the pollutant mass per fuel mass consumed, EF II was the pollutant mass per product mass (kg t -1 or g t -1 ), EF III reported in kg (10 6 yuan) -1 or kg MY -1 was the pollutant mass per industrial output value, and EF IV (kg MkWh -1 ) was the pollutant mass per generated millionaire kWh by a power plant: where EF I is the EF expressed as pollutant mass per fuel consumption (kg t -1 or g t -1 ), C i is the pollutant mass concentration in flue gas (kg m -3 ), V f is the flue gas volume generated from combustion of 1 kg solid fuel (m 3 kg -1 ), v is the velocity of flue gas (m s -1 ), D is the diameter of chimney (m), S is the feed rate of gaseous fuel (Nm 3 h -1 ), ρ is the mass density of gaseous fuel (t m -3 ).
where EF II is the EF expressed as pollutant mass per industrial output value (kg MY -1 ), m F is the annual amount of coal consumed by a factory (t a -1 ), RMB is the annual total output value of a factory (10 6 yuan a -1 ), T is the actual annual running time of a factory (h a -1 ).
where EF III is the pollutant mass per product mass (kg t -1 ) and designated to access the EFs for coke making industry, m P is the annual product mass (t a -1 ).
where PG is the annual power generation for a power plant (10 6 kWh a -1 ).

Determination of Flue Gas Volume from Combustion of Solid Fuel for Heating, Bathing, and Power
The needed theoretical air volume for complete combustion of 1 kg of coal or gangue was calculated as follows: V a 0 = 0.0889 ω(C ar ) + 0.2567 ω(H ar ) + 0.0333 ω(S ar ) + 0.0762 ω(N ar ) -0.0333 ω(O ar ) where V a 0 is the theoretical air volume for combustion of 1 kg of coal or gangue (m 3 kg -1 ), ω(C ar ) is the carbon content of coal or gangue (as received basis), ω(H ar ) is the hydrogen content of coal or gangue (as received basis), ω(N ar ) is the nitrogen content of coal or gangue (as received basis), and ω(O ar ) is the oxygen content of coal or gangue (as received basis).
The theoretical flue gas volume is calculated by Eq. ( 6): where V f 0 is the theoretical flue gas volume generated from burning of 1 kg coal or gangue (m 3 kg -1 ), V RO 2 is the sum of volume of CO 2 , SO 2 , and NO 2 (m 3 kg -1 ), V is the nitrogen volume in V a 0 (m 3 kg -1 ).
where M ar is the moisture content of coal.
Finally, the actually generated flue gas was calculated by Eq. ( 10): where α is the excess air coefficient, which is obtained from the investigation of factories.

Determination of Flue Gas Volume for Coking and Cement Industries
The flue gas for these two industries is not fully produced by fuel combustion.For coke making, the flue gas is generated from different processes including coal charging, coke pushing, coking and coke quenching, and fuel combustion.The flue gas volume cannot be obtained as aforementioned method and should be calculated by Eq. ( 11): where V f is the generated flue gas volume per annual produced coke or cement mass (m 3 kg -1 ), T is the annual running time (h), CP is the annual production of cement or coke (kg a -1 ).

Emission of Air Pollutants of the Coal-Fired Boilers for Bathing, Heating, and Production
The emission amounts of pollutants from the SCBs were higher than those of large-scale coal-fired boilers due to the absence of pollutant RDs (Ma et al., 2017b).Table 1 listed the parameters for 17 civil coal-fired SCBs for heating, bath, and production.The coal consumptions were ranged from 30 to 70,683 t a -1 for these 17 companies.The companies with coal consumption less than 1100 t a -1 were not equipped with SO 2 RDs.The removal rates for SO 2 RDs for the rest SCBs were ranged from 61% to 85%.The wet flue gas desulfurization (WFGD) technology was not used in these SCBs.The applied PM RDs in these SCBs contained fabric filters, water bath, and ceramic multichannel tubes.The 4 SCBs with the coal consumption as 30-400 t a -1 were not equipped with the PM RDs.Only 3 of 17 companies with high coal consumptions as 18,250, 42,000, and 70,683 t a -1 were equipped with NO x RDs including selective non-catalytic reduction (SNCR) facility, low NO x burner, and oxidation equipment.
Fig. 1 showed the EF I values (reported in kg t -1 ) for the SCBs applied in bath, heating, and production industries.The boilers with higher coal consumptions (Heating 15, 16, and 17) had the lower SO 2 , NO x , and PM EF I values compared with the other 14, possibly resulted from high pollutant removal rate of companies with high coal consumptions.The EF I values for Heating 15, 16, and 17 were ranged from 1.49 to 3.84 kg t -1 for SO 2 , from 1.81 to 2.19 kg t -1 for NO x , and from 0.532 to 3.68 kg t -1 for PM, respectively.
The sulfur content and removal rate of SO 2 RDs were key influencing factors on the SO 2 emissions.The highest SO 2 EF I value of 34.3 kg t -1 occurred at Heating 13, the high sulfur content (1%) of coal and no installed SO 2 RDs were the possible explanations.The following EF I value of 34.0 kg t -1 was possessed by Heating 8 with 2% sulfur content and the lime used as desulfurizer.The rest of the boilers with sulfur contents in the range of 0.29-0.60%had the SO 2 EF I values ranged from 1.49 to 9.67 kg t -1 with the average as 5.72 ± 2.55 kg t -1 .All the 17 civil SCBs were divided into two categories based on with or without SO 2 RDs.Class I was not equipped with SO 2 RDs and the 8 companies Heating 1, 2, 3, 6, 11, 12, 14, and Bath 10 were involved.Class II was equipped with SO 2 RDs and Heating 4, 15, 16, 17, Bath 5, and Production 7 and 9 were involved.The SO 2 EF I values for Class I (3.23-9.67 kg t -1 ; mean ± SD: 6.90 ± 2.31 kg t -1 ) were higher than those of Class II (1.49-7.65 kg t -1 ; mean ± SD: 4.37 ± 2.33 kg t -1 ).
The NO x EF I values for 17 companies were ranged from 1.81 to 9.68 kg t -1 with the average value as 3.16 ± 1.72 kg t -1 .Only 3 of 17 companies were equipped with NO x RDs such as SNCR, oxidation, and low NO x burner.The NO x EF I values for these 3 factories (1.81-2.19kg t -1 ; mean ± SD: 2.00 ± 0.192 kg t -1 ) were lower than those for the rest 14 ones without NO x RDs (2.93-9.68kg t -1 ; mean ± SD: 3.41 ± 1.79 kg t -1 ).The highest NO x EF I occurred at Heating 3 with the coal consumption as 310 t a -1 , its low combustion temperature and the high nitrogen content of coal (2.10%) are likely reasons for the high EF.Compared with SO 2 emission, the influence degree of generated vapor and coal consumptions of companies on NO x was relatively light.
Except for 4 factories with low coal consumptions (30, 80, 280, and 400 t a -1 ), 13 ones were all equipped with PM RDs including water bath, fabric filter, and wet method.The 4 factories without PM RDs possessed PM EF I values as 22.5-27.5 kg t -1 and the mean value as 23.9 ± 2.41 kg t -1 , which were significantly higher than those of 13 ones equipped with PM RDs (0.53-14.4 kg t -1 ; mean ± SD: 5.41

Emission of Air Pollutants from Power Plants
Although some advanced pollutant removal technologies were explored and improved in Chinese power plants, such as increasing generating capacity of individual block, adopt low NO x burner, and adopt advanced air pollution control devices and technologies, the emissions of air pollutants were tremendous (Ma et al., 2017b).The influence of fuel types and individual block power capacity (IBPC) on pollutant EFs should be discussed.
In this study, 15 power plants (PPs)-1 natural gas-fired PP, 1 coke oven gas-fired PP, 2 coal gas-fired PPs with IBPCs of 3 and 6 MW, 2 gangue-fired PPs with IBPCs of 135 and 25 MW, and 9 coal-fired PPs with IBPCs ranging from 6 to 600 MW-were examined.Besides the EF I and EF III , EF IV expressed by pollutant mass per generated power capacity (reported in kg MkWh -1 ) was also used in PPs.
Fig. 2 showed the EF III and EF IV values for all the 15 PPs.For coal-, coal gas-, and gangue-fired PPs, generally the EF III and EF IV values decreased with the increase of their IBPCs.It should be mentioned, the EFs for coal-fired PPs had the decreasing trends from 6 MW to 330 MW and 600 MW PPs possessed slightly higher EFs than 330 MW ones.The coal-fired PPs with IBPCs as 25-135 MW possessed the highest EF III values as 121-187, 141-448, and 36.5-101kg MY -1 for SO 2 , NO x , and PM.The corresponding values were 25.6-148, 50.9-91.9,and 12.9-21.8kg MY -1 for 300-330 MW PPs, while they were 85. 4-100, 111-178, and 18.3-51.6for 600 MW PPs.Gas-fired PPs possessed higher NO x EF III values, and similar SO 2 and PM EF III compared with the coal-and gangue-fired ones due to the lack of SO 2 , NO x , and PM RDs.For gas-fired PPs, the EF III values (in kg MY -1 ) were ranged from 31.4 to 184 (mean ± SD: 120 ± 65.2) for SO 2 , 1090 to 6380 (mean ± SD: 4630 ± 2140) for NO x , and 46.3 to 672 (mean ± SD: 314 ± 225) for PM.Yan et al. (2017) also reported the NO x EFs for gas-fired boiler for Beijing was as high as 1.42-6.86g m -3 .Gangue-fired PPs had the higher SO 2 , NO x , and PM EFs than coal-fired ones with the same IBPCs.For gangue-fired PPs, the 25 MW one possessed higher EF III values than 135 MW one, they were 5900, 333, and 770 of SO 2 , NO x , and PM for 25 MW PP, and 157, 202, and 176 kg MY -1 for 135 MW one.The high ash content in gangue and low removal rate of SO 2 and NO x (0.00%) resulted in higher EFs of gangue-fired PPs.The highest SO 2 EF III of 1520 kg MY -1 occurred at 6 MW coal-fired PP and the lowest value of 25.6 kg MY -1 was possessed by coal-fired 330 MW PP.While for PM, they were ranged from 12.8 kg MY -1 of coal-fired 300 MW PP to 3100 of coal-fired 6 MW PP.
The EF IV and EF I showed the same trends with the EF III (Table 2).The higher NO x EF IV and EF I values (88.4-1640 kg MkWh -1 and 1.29-19.1 kg t -1 ) occurred at the gasfired PPs due to the lack of RDs.The coal-fired 330 MW PP had the lowest NO x EF IV and EF I as 14.7 kg MkWh -1 and 0.265 kg t -1 , respectively.For PM, the coal-fired 6 MW PP possessed the highest EF IV and EF I as 1150 kg MkWh and 9.11 kg t -1 , while the lowest values occurred at 300 MW coalfired PP as 3.03 kg MkWh -1 and 0.061 kg t -1 , respectively.For SO 2 , the gangue-fired 25 MW PP had the highest EF IV and EF I as 3620 kg MkWh -1 and 16.7 kg t -1 , the lowest EF I  and EF IV were possessed by a coal gas-fired 6 MW PP as 0.023 kg t -1 and a coal-fired 330 MW PP as 7.37 kg MkWh -1 , respectively.
Among 9 coal-fired PPs, the 6 MW PP provided the higher pollutant EFs due to its relative low removal rate of pollutants.Except for 600 MW PPs, all the 3 classes of EFs decreased in the order of 6 MW > 25 MW > 100 MW > 135 MW > 300-330 MW.The EFs for SO 2 and PM of 600 MW PP were similar to those of 300-330 MW, while they possessed higher NO x EFs than 135 MW PP due to their high combustion temperature.The coal-fired 6 MW PP possessed the highest EF III and EF IV of SO 2 , NO x , and PM as 1520, 323, and 3100 kg MY -1 , and 561, 120, and 1150 kg MkWh -1 .Also the 6 MW coal-fired PP possessed the highest SO 2 and PM EF I as 4.46 and 9.11 kg t -1 .But for NO x EF I values for the coal-fired PPs, two 600 MW PPs had the highest levels as 1.24 and 2.12 kg t -1 .The significant difference among differently fueled PPs or different IBPC groups of coal-fired PPs was mainly influenced by pollutant removal rate, characteristic of fuel, and combustion conditions (Chen et al., 2014;Li et al., 2016Li et al., , 2017;;Xu et al., 2017b).Chen et al. (2014) indicated that units higher than 300 MW consumed 75% of coal, while the emitted pollutants contributed only 46%, 58%, 55%, and 63.2% to SO 2 , NO x , PM and PM 2.5 .Xu et al. (2017b) suggested the generator factors (GFs) (in g kg -1 ) of pollutants for PPs decreased with the decrease of IBPC values, and reported NO x for different IBPC groups (in MW) as 450-749, 250-449, 150-249, 75-149, 35-74, 20-34, 9-19, and ≤ 8 were 10.11, 9.33, 8.36, 8.13, 6.88, 6.54, 5.14, and 5.04 g m -3 , respectively, while their actual NO x emissions showed the reverse trend (Xu et al., 2017b).Ma et al. (2017b) also put forward that the pollutant EFs decreased with the increase of IBPC values of PPs.Even without any dust removal devices, the PM 2.5 EFs could also decrease from 153 g t -1 for 100 MW PP to 123 g t -1 for 300 MW PP.
Recently, fine particles such as PM 2.5 and PM 10 were more concerning than coarse ones due to their seriously adverse health effects (Ma et al., 2017a, b).In regard to the lack of data from field monitoring of PM 2.5 and PM 10 , the calculated data were given and shown as following: EF PM 2.5 or PM 10 = EFPM × mass contribution of PM 2.5 or PM 10 to total PM (12) Klimont et al. (2002) reported the PM 2.5 and PM 10 fractions as 6% and 23% for the pulverized coal (PC) boiler, while they were 7% and 29% for the circulating fluidized bed (CFB) boiler.Among 9 coal-fired and 2 gangue-fired PPs, 8 coal-fired PPs were equipped with PC boilers and the other 3 PPs possessed the CFB boilers.So the PM 2.5 and PM 10 EFs were calculated by Eq. ( 12).
The PM 2.5 and PM 10 EFs (expressed as EF I , EF III , and EF IV ) had the same trends with PM EFs (Fig. 3 and Table 3).For PM 2.5 EFs, the coal-fired 6 MW PP possessed the highest value as 217 kg MY -1 and 80.2 kg MkWh -1 , the lowest values occurred at 300 MW as 0.767 kg MY -1 and at 330 MW as 0.182 kg MkWh -1 .For PM 10 , the highest EFs were possessed by the 6 MW coal-fired PP as 898 kg MkWh -1 and 323 kg MkWh -1 , while the lowest values occurred at 300 MW coal-fired PP as 2.94 kg MY -1 and 330 MW coalfired PP as 0.698 kg MkWh -1 (Fig. 3).
The gangue-fired PPs possessed higher PM 2.5 and PM 10 EFs (EF III and EF IV ) values than coal-fired PPs with the same IBPC values, possibly resulted the high ash content in fueled gangue.
The highest EF I of 638 g t -1 for PM 2.5 and 3640 g t -1 for PM 10 were possessed by a coal-fired PP with lowest IBPC as 6 MW (Table 3).The lowest EF I occurred at 300 MW PP, which were 3.66 and 14.0 g t -1 for PM 2.5 and PM 10 , which were similar to the documented value (Ma et al., 2017b).They were 14 g t -1 for 300 MW PP with an electrostatic precipitator (ESP) and 3 g t -1 for 300 MW PP with an improved ESP (Ma et al., 2017b).The PM 2.5 EFs for coaland gangue-fired PPs were far less than those of household coal stoves.Chen et al. (2015b) measured the PM 2.5 EFs (reported in g kg -1 ) for 20 Chinese combinations of coal and stoves, and they were 4.25 and 1.44 for bituminous and anthracite coal, respectively.

Emission of Air Pollutants from Coke Making Industries
In this study, 22 small-and medium-sized coke making factories fueled with coke oven gas were investigated to access their pollutant EF II and EF III values.The annual coke outputs were ranged from 2.10 × 10 4 to 2.31 × 10 6 tons.All the 22 factories were equipped with bag filter to collect the dust and the removal rates ranged from 68.9% to 98.4%.Among 22 factories, only 5 ones equipped with SCR devices to eliminated NO x with the removal rate as 40.1-79.8%.The significant fluctuation of the SO 2 removal rate (0.00-98.1%) occurred among these factories.The SO 2 removal measures contained wet and dry FGD methods, and the desulfurization reagents were NaOH, Ca(OH) 2 , NH 3 , and Na 2 CO 3 .
The EF II and EF III values for 22 coking companies were not correlated with their coke productions and industrial outputs because their flue gas was not fully emitted from combustion process (Fig. 4).The heat for raw coal pyrolysis originated from the combustion of cycled coke oven gas, the mass of fuel is difficult to determine, so EF I cannot be used in coking industry.Fig. 4 showed the EFs of SO 2 , NO x , and PM expressed by EF II as kg (t coke) -1 and EF III .
For SO 2 , EF II and EF III values were ranged from 0.098 to 0.170 kg (t coke) -1 and 3.79 to 27.8 kg MY -1 for 22 companies.EF II and EF III values of NO x were in the range of 0.402-1.22kg (t coke) -1 , and 27.8-167 kg MY -1 , respectively.EF II and EF III values of PM were in the range of 0.063-1.08kg (t coke) -1 , and 7.71-148 kg MY -1 , respectively.Compared with SO 2 , NO x and PM have larger fluctuation among 22 companies under the influence of their pollutant removal rates, ash contents of coal and combustion conditions.
Table 4 listed the parameters of 5 companies with byproduct as tar and crude benzene.For these 5 companies, SO 2 EF II values were ranged from 2.38 to 5.14 kg (t tar) -1 and 9.33 to 18.8 kg (t benzene) -1 , NO x EF II values were in the range of 9.82-29.3kg (t tar) -1 and 38.5-111 kg (t benzene) -1 , and those of PM were in the range of 3.96-26.1 kg (t tar) -1 and 14.5-98.7 kg (t benzene) -1 .

Emission Factors for Pollutants from Cement Production Industry
Chen et al. (2015a) used the life cycle assessment (LCA) method to estimate the pollutant emissions of cement industry.The PM emission from cement industry was associated with mineral matter and coal combustion (Mari et al., 2016).The PM associated with raw material grinding and coal combustion entered the bag filter and discharged through induced fan.
In this study, 6 cement making companies including 2 small-and 4 medium-scale ones with the cement production as 16,400-1,464,170 t a -1 were investigated for SO 2 , PM and NO x emissions.The coal consumptions for 6 factories were ranged from 131 to 123,570 t a -1 with the sulfur and ash contents of coal were 0.2-0.8% and 7.6-26.04%,respectively.
All the 6 factories applied bag filter to remove PM with the removal rates ranged from 72.3% to 98.8% due to the fluctuation of running time and running conditions of PM RDs.It should be mentioned, all the 6 cement factories were not equipped with NO x removal devices, and only 2 factories were equipped with SO 2 removal devices.
EF I , EF II , and EF III were used to express the pollutant EFs for these 6 companies.Fig. 5 listed the calculation results of EF I value for SO 2 , NO x , and PM.Generally, the SO 2 EF I values (0.964-10.2 kg t -1 ) were not correlated with the coal consumptions and cement outputs.PM EF I values were negatively correlated with the consumed coal amounts and cement output.The factory with lowest consumed coal (131 t a -1 ) possessed the highest PM EF I as 220 kg t -1 , while the lowest PM EF as 1.26 kg t -1 occurred at factory with more consumed coal (75,000 t a -1 ).The PM EFs were higher than those of coal-, gangue-, and gas-fired power plants.The factories with the higher cement output possessed the higher NO x EFs result from their high combustion temperature and the other combustion conditions.The 2 factories with highest coal consumptions as 75,000 and 123,570 t a -1 had the highest NO x EF I values as 16.9 and 15.2 kg t -1 , while the lowest NO x EF I values of 2.88 kg t -1 was possessed by factory with less consumed coal as 7,167 t a -1 .
Unlike EF I , the EF II and EF III were not correlated with production of cement and value of industrial output, possibly resulted from the differences of cement price, production processes, and investment of environmental protection (Table 5).

CONCLUSIONS
In this study, 17 civil SCBs for heating, bathing, and production; 13 coke making companies; 15 power plants (PPs; 9 coal-fired, 2 gangue-fired, 1 coke oven gas-fired, 1 natural gas-fired, and 2 coal gas-fired); and 8 cement making companies were field investigated and measured for their SO 2 , NO x , and PM EFs.All the investigated factories were located in Shanxi Province, China.
1) Of the 17 civil SCBs, 4, 14, and 10 were not equipped with PM, NO x , and SO 2 RDs, respectively.Generally, the SO 2 , NO x , and PM EFs decreased with increasing coal consumption.The 3 SCBs with the highest coal consumption displayed the lowest EF I values, which ranged from 1.49 to 3.84 kg t -1 for SO 2 , from 1.81 to 2.19 kg t -1 for NO x , and from 0.532 to 3.68 kg t -1 for PM.
The NO x EF I values were lower (1.81-2.19kg t -1 ; mean ± SD: 2.00 ± 0.192 kg t -1 ) for the 3 factories equipped with NO x RDs than for the other 14 (2.93-9.68kg t -1 ; mean ± SD: 3.41 ± 1.79 kg t -1 ).The sulfur content of the coal and the SO 2 removal rate were key influential factors for the SO 2 EF I .The highest SO 2 EF I , 34.3 kg t -1 , was found at a factory without any installed RDs that was burning coal with 1% sulfur.Higher average PM EF I values (23.9 ± 2.41 kg t -1 ) were exhibited by the 4 factories without PM RDs compared to the 13 equipped with dust removal devices (5.41 ± 5.44 kg t -1 ).
2) For the coal-, coal gas-, and gangue-fired power plants (PPs), generally, EF I , EF III , and EF IV decreased with increasing IBPC.Lacking SO 2 , NO x , and PM RDs, gasfired PPs possessed higher NO x EFs but similar SO 2 and PM EFs compared to coal-and gangue-fired plants, and given the same IBPC values, gangue-fired PPs displayed higher EF I , EF III , and EF IV values than coal-fired PPs.The EF I , EF III , and EF IV values for coal-fired PPs decreased as the IBPC increased, in the order of 6 MW > 25 MW > 100 MW > 135 MW > 300 and 330 MW.The SO 2 and PM EFs for 600 MW coal-fired plants, however, were similar to those for 300-330 MW plants, although they exhibited higher NO x EFs than 135 MW coal-fired plants, due to the use of high combustion temperatures at the 600 MW plants.The EF I , EF III , and EF IV values for PM 2.5 and PM 10 followed the same trends as those for the overall PM.
3) The EF II and EF III values for 22 coking companies were not correlated with coke production or industrial output, as the flue gas was not fully emitted during the combustion process.
4) Generally, the SO 2 EF I values (0.964-10.2 kg t -1 ) for the cement factories were not correlated with coal consumption or cement output.However, the PM EF I values were negatively correlated with the aforementioned variables: The factory that consumed the least amount of coal (131 t a -1 ) exhibited the highest PM EF value (220 kg t -1 ), whereas a factory with significantly higher consumption (75,000 t a -1 ) exhibited the lowest value (1.26 kg t -1 ).Additionally, the PM EF I values for the cement factories were higher than those for the coal-, gangue-, and gas-fired power plants.Factories with high cement output possessed high NO x EFs as a result of using a high combustion temperature.EF II and EF III values were not correlated with cement production or coal consumption.
water vapor (the sum of combustion of hydrogen in coal or gangue, vaporization of water in coal or gangue, and vapor in air) (m 3 kg -1

Fig. 1 .
Fig. 1.EF I values for 17 small-scaled boilers for bath, heating, and production industries.

Fig. 2 .
Fig. 2. EF III and EF IV of pollutants for 15 PPs based on power capacity and industrial output.

Fig. 3 .
Fig. 3. EF III and EF IV values of PM 2.5 and PM 10 from coal-and gangue-fired power plants.

Fig. 4 .
Fig. 4. EF II and EF III values of SO 2 , NO x , and PM for 22 coke making companies.

Table 1 .
Emissions of SO 2 , NO x and PM for 17 civil coal-fired small scale heating boilers.

Table 2 .
EF I values of SO 2 , NO x , and PM for 15 power plants, kg (t coal/gas/gangue) -1 .
a Natural gas; b Coke oven gas; c Coal gas.

Table 3 .
EF I values of PM 2.5 and PM 10 for 11 power plants, g (t coal/gangue) -1 .

Table 4 .
Annual raw material consumption and product output for 5 companies with byproduct as crude benzene and tar.
EF I values of SO 2 , NO x , and PM for 6 cement making companies.

Table 5 .
EF II and EF III values for 6 cement making companies.