The Sulfur Balance
One aspect that has become very important is the heterocompound presence in fuels. Heterocompounds are molecules that contain sulfur, nitrogen, oxygen, or metallic porfirins. The combustion of these heterocompounds produces pollutants that contain sulfur oxides (SO*, x = 2, 3), nitrogen oxides ( NOx, x = -, 1, 2, 3), and so on. Currently, the most regulated of them are the SOx; therefore, it is necessary to consider their generation during cracking reactions. As described, at the refinery the FCC process produces about 40% of the total gasoline; however, it contributes more than the 90% of the sulfur content of commercial gasolines. The sulfur content of FCC products depends on catalyst, feedstock, and conversion as well as the reactors operating conditions. The FCC feedstock contains sulfur linked to organic compounds of high molecular weight; these heterocompounds are concentrated at the heavy end. Cracking these molecules produces either sour gas (hydrogen sulfide, which is desirable) or sulfided fuels (undesirable). Sour gas can be recovered and treated downstream in order to produce solid sulfur or sulfuric acid. In contrast, sulfur contained by fuels will produce SOx emissions at internal combustion engines.
In the case of sulfur in coke, stack gas emissions from an FCC regenerator, including SOx, NOx, and catalyst particulates, constitute a major environmental pollution concern. Some other strategies can be implemented after the stack of the regenerator, such as gas desulfurisation and scrubbing; however, these solutions are noneconomic. For fuels produced by FCC units, strategies to reduce sulfur in FCC gasoline include naphtha hydrofinishing and lowering the gasoline endpoint. Hydrofinishing significantly lowers the octane of FCC gasoline, which depends on the presence of unsaturated compounds; meanwhile, lowering the gasoline endpoint can significantly diminish yield to gasoline. Therefore, it is very important to separate sulfur in catalytic cracking processes as sour gas at the riser outlet.
Kinetic schemes and mathematical models presented previously do not consider sour gas generation and sulfur distribution into cracking products; however, there have been some attempts at explicit prediction of sulfur content and distribution into catalytic cracking products. For example, Villafuerte . Madas et al. (2003, 2004) proposed a seven-lump kinetic scheme that considers the individual formation of H2S. The seven lumps are selected as follows: for liquid products, according their boiling points: feedstock (343 to 560°C), cyclic oils (HCO + LCO, 223 to 342°C), and gasoline (C5+, 37 to 222°C); and for light products, according to environmental and trade requirements: LPG (C3-C4), dry gas ( C1-C2,H2), sour gas ( H2S), and finally, solid coke (C). Each product is able to be cracked into lighter products. The feedstock cracking reaction is of second order, and all other cracking reactions are of first order, as has generally been assumed. Rate constants for cracking reactions follow the Arrhenius dependence on temperature. Initial numerical values of kinetic parameters were selected from literature data and were adjusted as well as validated, utilizing a number of sets of industrial refinery product results. Figure 5.64 illustrates the kinetic scheme proposed, and Table 5.22 provides the kinetic parameters used.
Authors also developed empirical functions related to feedstock conversion as well as reactor temperature, to represent not only the sulfur content and sulfur distribution in CO, gasoline, and coke, but also sulfur as sour gas independent of light lumps. Following industrial practice (MAT laboratory evaluation scaled to industrial data), the parameters of these functions were obtained for a particular feedstock and a "type of catalyst" and are supported with industrial (actual) data. These parameters should be fitted whenever different feedstock or "another type of catalyst" is used. The operating conditions of a riser-regenerator system do not modify parameter values.
Figure 5.64. Seven-lump kinetic scheme.
TABLE 5.22. Kinetic Parameters Used in the Model
|
Cracking Reaction |
k a |
E (kJ/mol) |
 |
240.0 |
70.0 |
 |
380.0 |
70.0 |
 |
70.5 |
70.0 |
 |
217.5 |
80.0 |
 |
2400.0 |
70.0 |
 |
0.40 |
50.0 |
 |
24.0 |
60.0 |
 |
30.0 |
60.0 |
 |
217.5 |
60.0 |
 |
600.0 |
70.0 |
 |
0.60 |
50.0 |
 |
1.0 |
50.0 |
 |
145.0 |
70.0 |
 |
300.0 |
70.0 |
 |
0.50 |
50.0 |
 |
261.0 |
40.0 |
 |
0.40 |
40.0 |
 |
1.30 |
40.0 |
TABLE 5.23. Characteristics of the FCC Unit
|
Type |
Riser reactor/adiabatic regenerator |
|
Operating mode |
Complete combustion |
|
Feedstock capacity (bbl/day) |
25,000 |
|
Average coke production (tons/day) |
160 |
|
Average airflow rate (m3/h) |
75,000 |
Figure 5.65. Predicted vs. observed values for product yields.
A complete combustion regenerator will be analyzed. Its main characteristics are listed in Table 5.23. The vector of state variables consists of oxygen and sulfur concentrations, coke on regenerated catalyst, CO concentration, and dense bed temperature. Predicted yield values of cracking products and actual values are compared in Figure 5.65. It is important to observe that prediction points fall in the neighborhood of the 45° line; therefore, the values predicted are close enough to the values observed.
Product yield profiles and actual yield values have been modeled at the riser exit (Figure 5.66). It is possible to observe that the main feedstock cracking occurs before the first third of the riser length. The cyclic oil yield reaches a maximum value before the first half of the riser length, following a predominant soft decreasing yield due to cracking. This last result is in agreement with those intermediate-weight mass products that might be converted to minor molecular weight compounds. Most (about 90%) of the total gasoline final yield is obtained before the first half-length of riser. Under industrial conditions, it has been observed that once gasoline is produced, it is not easily cracked and also that LPG and dry gas yields are increased continually (Figure 5.67. . The greatest sour gas yield is obtained before three -fourths of riser length, indicating an initial easy link breaking sulfur hydrocarbons. Coke yield is increased as a result of the condensation of cyclic, heterocyclic, and alkyl compounds on catalyst particles.
Figure 5.66. Axial profiles of feedstock and products in the riser.
Figure 5.67. Sulfur content of FCC products.
The predicted sulfur content of cyclic oil, gasoline, and coke obtained as a function of ROT is shown in Figure 5.67 . It is important to note that sulfur content in cyclic oils increases as ROT is increased; meanwhile, the sulfur content of gasoline, as well as of coke, decreases. Unstable sulfur-linked hydrocarbon compounds are cracked into sour gas and more light hydrocarbons; meanwhile, noncracked sulfur compounds go into cycle oils and only a little into gasoline. The sulfur content predicted for cyclic oils, gasoline, and coke is in agreement with actual data. It should be noted that to obtain gasoline with a lower sulfur content and a higher coke yield with a lower sulfur content, the unit must be operated at higher ROTs.
Figure 5.68. Product profiles as a function of ROT.
Profiles predicted for cyclic oil, gasoline, LPG, dry gas, sour gas, and coke yields obtained when simulating the operation between 510 and 550°C of the ROT are shown in Figure 5.68. Actual data are also included. The values predicted are depicted by a line crossing a neighborhood of actual data. It is observed that the cyclic oil yield predicted decreases as the ROT is increased, whereas the gasoline, LPG, sour gas, dry gas, and coke yields predicted increase. The decrement in cyclic oil yield is a result of cracking to LPG, dry gas, and some gasoline. At the highest temperature there is only a little or no increase predicted for gasoline yield. The industrial practice suggests gasoline cracking at an ROT higher than 550°C. It is important to note that an increase in coke yield is a possible advantage because of the relationship between necessary energy in the regenerator and the heat balance of the riser-regenerator-stripper system. The high sour gas yield predicted induces a lower sulfhur content in gasoline and coke, basically an advantage.
There is a type of synergy between an increase in gasoline yield and a decrease in the sulfur content in this fuel as the ROT increases. Therefore, to preserve the profitability of the operation, FCC units should be operated at the highest possible ROT. At the same time, there is an increase in sour gas yield, which is also an advantage from an environmental point of view. Both enhancements are made at the cost of a higher sulfur content in cyclic oils. This situation has to be balanced because of the cost of desulfurization downstream. However, the yield of cyclic oils is also decreased, which could also be an advantage.
By using a seven-lump kinetic scheme, which clearly specifies H2S generation, it is possible to account for contributions of H2S formation, sulfur content in cracking final products, and sulfur distribution in cracking products. In addition, but in common with other models, it helps to predict cracking product distribution. The results predicted are referred to the feedstock volume conversion and the riser outlet temperature range in which FCC units are commonly operated. This information helps to manage the sulfur content of fuels during fuel production. This model is, furthermore, a helpful tool for modeling steady-state FCC operation, taking into account valuable cleaner fuel production and satisfactory environmental control.
CONCLUSIONS
Fluidized-bed catalytic cracking (FCC) is one of the main processes in petroleum refining. The heart of this process is the converter, which consists of a riser (transported- bed reactor, where the principal reactions take place), a stripper (a fluidized-bed reactor used to desorb gaseous hydrocarbons from the catalyst surface), and a regenerator (a fluidized-bed reactor used to burn off coke produced, recovering catalyst activity and energy to sustain the converter). This converter is a very complex system, due to the variety of compounds used as feedstock and the highly interacting nature of the system as a consequence of the energy balance.
During the analysis and design of the riser, it is necessary to evaluate (properly) the cracking reaction kinetics; nevertheless, this kinetics involves too many compounds and requires intensive feedstock analysis in order to be characterized in an acceptable way. On the other hand, due to the catalyst and reacting fluid process, the entire cracking kinetic process is rated by mass transfer, either at the fluid-particle interface and/or as intraparticle diffusion. As a consequence, evaluation of effective (apparent) kinetic parameters in laboratory devices (such as MAT, ACE, CREC-riser-simulator, and pilot plants) are linearly scalable to industrial riser parameters. This result has been used, empirically, for many years to evaluate the performance of catalyst-feedstock couples at the laboratory scale, prior to use at industrial FCC units.
The regenerator of the FCC manages the energy balance. This reactor is a very complex system because it is able to exhibit a variety of dynamic responses to the many disturbances that commonly occur. There are two principal types of regenerator modes: full combustion (no CO in the flue gases) and partial combustion (CO in the flue gases, followed by a CO combustor). Regenerator study is not finished; therefore, there are several recent papers regarding the dynamics and control of FCC regenerators.
Finally, FCC is able to utilize many different types of feedstock; additionally, it is possible to alter the conversion to products by managing the riser outlet temperature. Hence, these units could help to transform solid pollutants (such as plastics) into gasoline. Meanwhile, the use of gas oils containing heterocom-pounds that contain sulfur or nitrogen causes the oxides of those heteroatoms to be emitted. This problem is under research in order to improve process operation, to pretreat feedstocks, and to find options that are more environmentally friendly.
Some Perspective on Present and Future Opportunities Everything (i.e. scientific reasearch) can be studied taking as a target the FCC unit. The question is: Why is it that way?
FCC units were developed during World War II; gasoline for airplanes was critical and there was no unit to produce it. Therefore, in a very intrepid and rapid development, some inspired researchers designed the first TCC unit, the precursor of the current FCC. This new unit was very complex, exhibiting very innovative features, including moving-bed reactors and the chance to spend and regenerate the catalyst inside the converter during current operation. However, even though the unit eventually worked successfully, the development brought with it a lot of empirism. Even when the economists running refineries want "explicitly sure results," this is one of the best examples of engineering design.
As market demands had changed, feedstock had to adapt to many other streams inside the refinery, there is much new environmental legislation, and so on. However, the FCC unit is still the heart of the refinery and continues to evolve. Furthermore, whatever the future outlook for petroleum, FCC units will continue working in the production of raw materials for specialty plastics. Therefore, it is necessary to continue to develop better kinetic models. Lumping models have sufficed to date, but they lack many beneficial feedstock and product properties. Use of single-event models could be a good future strategy, as they require detailed analyses of feedstock, products, and byproducts such as coke (the second- most- i mportant FCC product). Catalyst development is one targets of focus around the world, but descriptions of the loss of activity (deactivation) are still very empirical. The most advanced part of this study explains the deactivation phenomenon in terms of the specific rate of formation of coke. This is not enough to characterize the phenomenon; there are no explanations regarding coke deposition on a catalyst surface causing a decrease in the pore-mouth diameter (e.g., Jimenez t Garcta et al., 2009). If this phenomenon were not important, we would not see very dark spent catalyst.
Finally, with regard to dynamics modeling and control, do we understand the FCC unit? Are we able to build the next generation of FCC units? Well— work for the future!
