3.1. Emissions of regeneration flue gas from three typical FCC units
3.1.1 FTIR results
FTIR method could avoid the effect of CO and water vapor with high concentration, and are applicable for the complex compositions flue gas (Luan et al., 2020). In order to find out the emissions of FCC regenerated pollutants, such as whether the pollutant concentration exceeds the limitation, whether the online monitoring data is accurate and whether there are unknown pollutants, the on-site monitoring of typical FCC unit regenerated flue gas was carried out by using the FTIR method. The emission concentrations of regenerated pollutants in the stack flue gases of three typical FCC units are shown in Table 2.
Table 2 FTIR Monitoring Results of Regenerated Pollutants(mg/m3)
Pollutants
|
NOx
|
SO2
|
NH3
|
C6H6
|
HCN
|
C8H8
|
C2H4
|
CH4
|
CO
|
Projects
|
Hourly mean
|
U1
|
66.71
|
86.41
|
116.99
|
56.41
|
71.94
|
27.79
|
12.50
|
164.14
|
1.48%
|
U2
|
52.22
|
31.65
|
5.50
|
7.13
|
3.46
|
9.95
|
0.98
|
3.91
|
0.78%
|
U3
|
60.93
|
11.96
|
1.14
|
0.04
|
0.01
|
0.06
|
0.55
|
0.00
|
0.08%
|
U1: 1.4 million tons / year partially regenerated FCC unit without CO boiler
U2: 1.2 million tons / year partially regenerated FCC unit with CO boiler
U3: 1.6 million tons / year fully regenerated FCC unit
As seen from Table 2, the NOx and SO2 content of U1 are 66.71 and 86.41, respectively, which are the highest among three units. In addition, the FTIR data of three typical FCC units shows that not only SO2 and NOx, which are already well-known pollutants, are present in the catalytic regeneration flue gas, but also find that the FCC flue gas contains methane, ethylene, benzene, styrene, hydrogen cyanide, ammonia and other pollutants. It is noted that the content of NH3 and HCN are 116.99 mg/m3 and 71.94 mg/m3, respectively, which are much higher than expected. Moreover, some pollutant concentrations have seriously exceeded the limitation. At present, only SO2, NOx and particles are monitored manually or on-linein most enterprises. Obviously, few kinds of pollutants are monitored, so it is difficult to meet the control requirements of national regulations on pollutant emissions from refineries. But this on-site monitoring results found that there were many unpredicted pollutants of FCC flue gas in these three units, which was a quite noteworthy issue. VOCs are organic compounds that participate in atmospheric photochemical reactions or bring about odor complaint. VOCs are important precursors to form PM2.5 and O3, which can lead to the higher atmospheric oxidizing ability and adversely impair the air quality and human health, meanwhile, FCC units were identified as the largest contributor (Wei at al.,2014). In the process of denitration treatment, a large proportion of excessive ammonia spraying in pursuit of NOx emission indicators directly leads to a gas-phase reaction between the gaseous NH3 evaporating from the ammonia water and the SO2 in the flue gas, which form a large amount of aerosol discharging into the atmosphere (Huang et al.,2016; Bao et al., 2017). In addition, HCN is acutely toxic via inhalation. Acute oral doses of cyanide cause cardiovascular, respiratory, and neurophysiological changes, and may even damage the brain (IPCS, 2004). Emission factors were developed from the measured HCN emissions from FCC units undertaken by the US EPA in 2011 and performed by Concawe in 2016, respectively (US EPA, 2015; Concawe, 2019). The results confirm that the FCC unit will emit a certain amount of HCN. In addition, FCC units in other refineries in China have hardly carried out similar on-site monitoring, and have not developed the emission inventory of FCC regenerated pollutants. Therefore, it is not clear whether these pollutants exist in other FCC units and what their emission concentrations are. It is worth pondering whether the "white plume" of FCC unit after flue gas desulfurization which is becoming increasingly public concern is clean or not.
3.1.2 SO2 and NOx results
It is found that the on-site monitoring data of FTIR method differed greatly from the online monitoring data of CEMS. Specific data are shown in Table 3.
Table 3 Monitoring results of SO2 and NOx by FTIR and CEMS(mg/m3)
Project
|
FTIR
|
CEMS
|
SO2
|
U1
|
86.41
|
9.75
|
U2
|
31.65
|
3.35
|
U3
|
11.96
|
1.10
|
NOx
|
U1
|
66.71
|
17.50
|
U2
|
52.22
|
11.39
|
U3
|
60.93
|
14.98
|
U1: 1.4 million tons / year partially regenerated FCC unit without CO boiler
U2: 1.2 million tons / year partially regenerated FCC unit with CO boiler
U3: 1.6 million tons / year fully regenerated FCC unit
From the table, the SO2 and NOx content of U1 from FTIR are 86.41 mg/m3 and 66.71 mg/m3, respectively, which are approximately 9 times and 4 times higher than the CEMS data, respectively. Similar results are found for the other two units. The compositions of the exhaust gas are complex and the humidity is high, which will affect the online measurement results of sulfur dioxide and nitrogen oxides. After investigation, all of these three FCC units used wet desulfurization processes (WGS), and the flue gas humidity was about 10 ~ 30%. The measuring principle of the equipment is all cold-dry straight extraction. During the condensation and water removal pretreatment process, most of the sulfur dioxide was absorbed by the ammonia gas and condensate in the flue gas, which caused the sulfur dioxide online monitoring data to be low, or even undetectable. The Ion chromatographic analysis of the condensate water generated during the condensation process of the flue gas confirmed our conjecture later. Detailed studies have been described in our previous works (Luan et al., 2020). In addition, some catalytic regeneration process uses two stages of incomplete regeneration, and there is no supporting construction of CO boiler, resulting in the presence of a large amount of other pollutants such as HCN, NH3, CO and VOCs in the flue gas. Due to the presence of many components to be measured in the infrared region, the respective characteristic spectra may overlap when using non-dispersive infrared on-line measurement, so these pollutants will cause a certain degree of interference to the online monitoring results of SO2 and NOx, making the measurement results low. According to the survey (Li, 2013), all of the online monitoring equipment used in FCC units in China are cold-dry straight extraction methods, resulting in that the online monitoring data is much lower than the on-site monitoring data. This makes it difficult for on-site monitors to judge the FCC flue gas emissions.
The analysis results obtained through these monitoring processes show that the FCC flue gas has the risk of exceeding the pollutant emission standard; the online monitoring data of the FCC flue gas is not accurate, and there is a risk of being identified as fraud; there is a risk of incomplete monitoring and unclear number of regenerated pollutants. Therefore, it is necessary to further study the emissions of regenerated pollutants from FCC units and establish emission factors for regenerated pollutants from FCC units.
3.2. Emission factors of Regeneration Pollutants
The role of emission factors is to provide relatively accurate estimates of the emissions of pollutants when the online monitoring data is inaccurate and incomplete and there is no on-site monitoring data, especially to provide an informative data basis for the development of regional emission inventories. Therefore, it is critical to accurately calculate the emission factor, which is attributed to the issue of a baseline selection, which needs to be highly correlated with the emissions of pollutants. The emission factors of regeneration pollutants in the stack gases from these FCC units were investigated, which will benefit the regeneration pollutant emission inventories of FCC unit.
3.2.1 Calculation of emission factors
Generally, all regenerated pollutants production is more closely related to coke burn rate than feed rate, and coke yield varies with feed characteristics; however, if FCC unit feed and operating conditions do not vary significantly, throughput-based emissions factors may be used. Therefore, the two types of emission factors of regeneration pollutants in the stack gases from the individual FCC unit could be calculated separately, one is based on the measured emissions rate (kg/h) of regeneration pollutants divided by the coke burn-off rate (t/hr), the other is based on the measured emissions rate (kg/h) of regeneration pollutants divided by the processing rate (throughput/h). The calculation of the coke burn-off rate is based on the main air flow and dry flue gas composition. First, the main air flow is calculated as the flue gas flow using nitrogen balance, and the dry flue gas volume is calculated by considering the molecular humidity of the air. Then, the coke burn-off rate is calculated by using the composition of the regenerator outlet flue gas. Only two elements of coke and hydrocarbon are considered here, and the combustion rate of carbon and hydrogen are calculated, respectively. Sum of carbon burning rate and hydrogen burning rate is the coke burn-off rate. The corresponding calculation results are shown in Table 4 and Table 5.
Table 4 Emission factors based on the coke burn-off rate of regeneration pollutants (kg/t)
Pollutants
|
NOx
|
SO2
|
NH3
|
C6H6
|
HCN
|
C8H8
|
C2H4
|
CH4
|
CO
|
Projects
|
Emission Factors
|
U1
|
1.09
|
1.41
|
1.91
|
0.92
|
1.17
|
0.45
|
0.20
|
2.68
|
301.72
|
U2
|
1.54
|
0.93
|
0.16
|
0.21
|
0.10
|
0.29
|
0.03
|
0.12
|
287.08
|
U3
|
1.74
|
0.34
|
0.03
|
1.1E-3
|
3.0E-4
|
1.7E-3
|
1.6E-2
|
0.00
|
28.49
|
Table 5 Emission factors based on the processing rate of regeneration pollutants (kg/t)
Pollutants
|
NOx
|
SO2
|
NH3
|
C6H6
|
HCN
|
C8H8
|
C2H4
|
CH4
|
CO
|
Projects
|
Emission Factors
|
U1
|
0.070
|
0.090
|
0.122
|
0.059
|
0.075
|
0.029
|
0.013
|
0.171
|
19.3
|
U2
|
0.114
|
0.069
|
0.012
|
0.016
|
7.58E-3
|
0.022
|
2.15E-3
|
8.57E-3
|
21.4
|
U3
|
0.151
|
0.030
|
0.003
|
9.93E-5
|
2.48E-5
|
1.49E-4
|
1.37E-3
|
0
|
2.48
|
U1: 1.4 million tons/ year partially regenerated FCC unit without CO boiler
U2: 1.2 million tons/ year partially regenerated FCC unit with CO boiler
U3: 1.6 million tons/ year fully regenerated FCC unit
The emission factors of different pollutants vary greatly for different units, and the emission factors of each pollutant are affected under different degrees by different factors, or these factors have different effects on the pollutant emission factors. By comparing the differences between these three FCC units, the factors that affect the differences in emission factors can be classified into three categories, namely coke properties, regeneration processes and the presence or absence of CO boilers. The specific impact analysis will be discussed in detail in next section.
According to the US EPA AP-42, fifth edition, Vol. 1, Chapter 5: Petroleum industry, Emissions estimation protocol for petroleum refineries -Version 3.0 and the Report No. 4/19 published by CONCAWE, emission factors of regeneration pollutants in the stack gases from FCC unit in the refinery are shown in Table 6:
Table 6 Emission factors of regeneration pollutants (US EPA, 1995/2008/15; CONCAWE, 2019)
Pollutants
|
NOx
|
SO2
|
NH3
|
C6H6
|
HCN
|
NM
VOC
|
CO
|
Projects
|
Emission Factors
|
K1
|
-
|
-
|
0.57
|
1.10E-3
|
0.43
|
-
|
-
|
K2
|
0.235a
|
1.63b
|
0.043
|
5.96E-5
|
0.023
|
0.729
|
45.4
|
K3
|
0.204c
|
1.41d
|
0.155
|
-
|
-
|
0.630
|
39.2
|
K1:Emission factors calculated based on coke combustion rate,kg/t
K2:Emission factors calculated based on feedstock mass flow rate,(kg/t)
K3:Emission factors calculated based on feedstock volume flow rate,kg/m3
a: The value range is 0.123 ~ 0.481
b: The value range is 0.331 ~ 1.740
c: The value range is 0.107 ~ 1.416
d: The value range is 0.286 ~ 1.505
Among the emission factors for FCC unit regeneration pollutants given in Table 6, the emission factors for NOx and SO2 do not distinguish between the presence and absence of a CO boiler. NH3, C6H6, NMVOC, and CO are considered to be controlled to a negligible level of pollutants after the CO boiler is installed. HCN is relatively special. Europe and the United States have specifically investigated and developed the emission factors of FCCU HCN. For the US EPA, the sets of individual unit emission factors, irrespective of mode of operation, were therefore averaged to provide the final published factors of 0.43 kg HCN/t coke burn and 0.023 kg HCN/kg‧1000 unit feed (US EPA, 2015). According to the European Pollutant Release and Transfer Register (E-PRTR) report,they developed an emission factor for FCCU HCN of 0.58 kg HCN / t coke burn for full regeneration units and 0.042 kg HCN / t coke burn for partial regeneration units (CONCAWE, 2019). These results indicate that reasonable emission factors should be developed for different pollutants under different conditions. This conclusion is consistent with the above research.
Fig. 2 shows the emission factors based on throughput of regeneration pollutants from investigative three FCC units and AP-42. From the Fig. 2,it can be seen that emission factors of three FFC units and AP-42 are different. This is normal because different raw materials, different reaction-regeneration processes, and differences in downstream pollutant emission control devices can significantly affect the final pollutant emissions. Therefore, the next research should investigate more samples and perform corresponding statistical analysis to obtain a reasonable classification and uncertainty range of emission factors.
3.2.2 Effect of regeneration process on emission factors
Fig. 3 shows the emission factors based on the coke burn-off rate and the processing rate of regenerated pollutants from three FCC units, respectively. Different regeneration processes can affect emission factors significantly. The emission factors of the partial regeneration unit without the CO boiler (U1) are generally significantly higher than those of the other two units, and in these two units, the emission factors of the complete regeneration unit (U3) are generally lower than those of the partial regeneration unit with a CO boiler (U2). Generally speaking, a wide range of variables, to varying degrees, affect the regeneration pollutant emissions, such as the refinery crude throughput, coke properties, the process units installed and type of equipment in use, etc. Therefore, this cannot be distinguished only by the presence or absence of CO boilers, but also by the differences in their operation modes and coke properties. The reasons for the different emission levels of each pollutant need to be discussed separately. Here, the typical pollutants SO2, NOx, NH3, HCN, CH4, and CO are discussed separately.
To better understand the relationship between pollutant emissions and pollutant precursors on the spent catalysts, the elements contents on the spent coke are analyzed, which is shown in Table 7. From the table, it can be seen that the content of sulfur and nitrogen of C1 are 0.32wt.% and 0.74wt.%, respectively, which is the highest among three catalysts. This means sulfur-containing and nitrogen-containing pollutants from the U1 flue gas may be much higher than the other two. In addition, the emission level of SO2 has a significant correlation with the sulfur content in coke, this can easily explain the reason for the high SO2 emission factor in U1. As for why the sulfur content of C2 is lower than C3, but the SO2 emission in the flue gas in U2 is greater than that of U3. For this phenomenon, it is related to the form of regeneration of U3. U3 is fully regenerated, and luan et al.(Luan et al., 2020) believes the amount of SO2 generated in fully regenerated process is significantly reduced.
Table 7 Results of spent FCC catalyst elemental analysis
Contents(wt.%)
|
C
|
N
|
H
|
S
|
Sample NO.
|
C1
|
0.84
|
0.32
|
0.71
|
0.74
|
C2
|
1.38
|
0.06
|
0.41
|
0.06
|
C3
|
1.45
|
0.09
|
0.27
|
0.14
|
C1: spent catalyst from U1
C2: spent catalyst from U2
C3: spent catalyst from U3
However, the NOx emission factor for U1 is lower than that of the other two devices. This is because most of the coke nitrogen is converted to molecular nitrogen (N2), even though the presence of coke nitrogen is an important source of FCC NOx (Concawe, 2009). This means changes in the nitrogen content in the feed and coke will not affect NOx emissions significantly. What really affects NOx emissions is the regeneration type and the presence or absence of a CO boiler. Partial regeneration usually exists in the two-stage regeneration process, which is repeatedly performed under low temperature and oxygen-depleted conditions. The oxygen content is usually less than 0.5%. Most of the hydrogen and carbon on the coke react in this zone, which inhibits the formation of NOx and favours the formation of N2 and of more reduced S and N species such as COS, H2S, NH3 and HCN (Babich et al., 2005; Ju et al., 2020). From Table 4 and Table 5, NOx emission factors of U3 is the highest among three modes, which are 1.74 kg NOx/t coke burn and 0.151 kg NOx/t throughput, respectively. This also confirms that full regeneration will produce more NOx. Previous research (Ju et al., 2020) has shown that pyridinic nitrogen (N-6), pyrrolic nitrogen (N-5), and quaternary nitrogen (N-Q) are the main precursors of nitrogen -containing species, including NO, NH3 and HCN etc., during FCC regeneration, different regeneration conditions, such as regeneration temperature and oxygen concentration, could cause these nitrogen-containing compounds to transform into each other. In addition, due to partial regeneration, it produces less NOx, and there is no CO boiler downstream, resulting in that NH3, HCN and other gases cannot be converted to NOx under conditions of high temperature, excess oxygen, and long residence time, and eventually lead to lower NOx emissions in U1. It also explains that the NH3 and HCN emission factors of the device are large.
Table 7 shows that the H/C ratio of the three catalysts is 0.85, 0.30, and 0.19, respectively, which is mostly consistent with the literature (Cerqueira et al., 2008) that the coke component on the catalyst usually has an H/C ratio of 0.3 ~ 1.0. The H/C ratio can represent its coking degree because a substance with low H/C ratio may have more condensed rings than that with high H/C ratio ( Cerqueira et al., 2008; Behera et al., 2013). So it believes that Cat2 and Cat3 may have more condensed rings. Since aromatic carbon is more stable than aliphatic carbon, during the combustion of coke, it is not easy to decompose into small molecules of hydrocarbons. Thus, U2 and U3 emit less hydrocarbons. From the perspective of regeneration process, on the other hand, VOCs、CO and CH4 are formed as the coke is burned in partial burn units. These species leave the regenerator and enter the CO boiler where they are largely converted to CO2. It can be seen from the monitoring results that although there is a CO boiler downstream during partial regeneration, the CO emission concentration in the flue gas may not reach a non-negligible level (< 500 ppm). Under fully regeneration conditions, most of the C and H substances on the coke will be completely burned to form CO2 under sufficient oxygen conditions, resulting in less VOCs, CO and CH4. Thus, the VOCs, CO and CH4 content in U1 is high, while U2 has less VOCs, CO and CH4 and U3 has the least.
To sum up, the emission factors of FCC unit regeneration pollutants need to be distinguished from the types of coke properties, regeneration modes, and pollutant control devices. The emission factors of each pollutant are also affected by each type of impact differently. Reasonable emission factors need to be developed based on the specific circumstances.