当前位置:首页 >> 能源/化工 >>

Modeling of Naphtha Pyrolysis


774

Ind. Eng. Chem. Process Des. Dev. 1985, 2 4 , 774-702

Modeling of Naphtha Pyrolysis
Pramod Kumart and Deepak Kunzru'
Department of Chemical Engineering

, Ind&n Institute of Technology, Kanpur, India

The effect of temperature, inlet steam to naphtha ratio, and space time on the yields of the major products in naphtha pyrolysis has been investigated. The temperature, steamlnaphtha, and space time were varied in the range 993-1073 K, 0.71-1.43 kg/kg, and 0.024-0.76 s, respectively. The overall naphtha decomposition could be represented by a first-order reaction. Based on the experimental results, naphtha pyrolysis was modeled by use of a set of molecular reactions together with an overall fistorder primary reaction with constant selectivities. The secondary reactions used in the model were found to be vktualry independent of the feed. Very good agreement was obtained between the calculated and experimental results. Moreover, the same model could also satisfactorily predict the product distribution data available in the literature for another naphtha, kerosene, and gas oil.

Naphtha pyrolysis is the major source of ethylene, propylene, and butadiene, which are the basic feedstocks for the petrochemical industry. Depending upon the naphtha and the operating conditions used, the ethylene yield varies anywhere from 26 to 36 wt% of the feed. To design the cracking coil, a good model for naphtha pyrolysis is essential. A detailed model can be used to predict the effect of operating variables on the product yields, to predict the concentration of various products along the reactor length, and to optimize the plant. In order to model the reactor, it is necessary to know the reactions taking place in the system. Naphtha contains a large number of hydrocarbons, and owing to the complexity of the feed and the radical nature of the reactions, hundreds of reactions can occur among the varous freeradical species. The preexponential factors and activation energies for most of these free-radical reactions are not precisely known. Furthermore, difficulties are enco-untered with the integration of stiff differential equations associated with radical reactions. Taking all these factors into account, a mechanistic model is not practical, if possible at all. Although many investigators (e.g., Ennis et al., 1975; Chambers and Potter, 1974) have measured the product yield in naphtha pyrolysis, very few detailed kinetic models have been published. Modeling of naphtha pyrolysis has been attempted at various levels of sophistication. The simplest models correlate the product yields with some parameter such as the kinetic severity function used by Zdonik et al. (1967) or the cracking severity index developed by Shu and Ross (1982). On the other extreme are the simulation programs based on fundamental free-radical reaction kinetics. Recently, Goosens et al. (1978) have reported the existence of a simulation model, which is able to predict the coking rates and product yields during pyrolysis for any hydrocarbon feed ranging from light hydrocarbons to gas oils. However, due to proprietory reasons no details have been disclosed. To reduce the number of reactions, some investigators have preferred to replace the radical reactions by molecular reactions. Hirato and Yoshioka (1973) have discussed such a model for naphtha, gas oil, and kerosene pyrolysis. They represented the primary and secondary reactions by gross molecular models, with the initial selectivities determined experimentally. The rate constants were adjusted so that the predicted product distribution matched the experimental data.
Department of Chemical Engineering, HBTI, Kanpur, India.
0196-4305/85/1124-0774$01.50/0

A more detailed model, based on individual molecular reactions, has been outlined by Van Damme et al. (1981). Naphtha pyrolysis was simulated by writing the continuity equations for 23 pure hydrocarbons present in naphtha and a number of pseudocomponents. The initial selectivities for normal and isoparaffins were obtained by using the Rice and Kossiakoff theory (1943) as modified by Murata and Saito (1974). The initial selectivities for naphthenes were estimated from the limited literature sources and their own experimental data. The secondary reactions were an extended version of the scheme published by Sundaram and Froment (1977). In all, the model contained 49 continuity equations. However, neither the details of the reaction scheme, nor the numerical values of the various parameters such as the preexponential factors, activation energies, and initial selectivities, were disclosed. Related to this study are the models developed for the pyrolysis of nornial and isoparaffins and their mixtures (Murata and Saito,1974,1975; Murata et al. 1974; Arai et al., 1977; Tanaka, 1975; Tanaka et al., 1976a). The aim of the present study was to investigate the effect of temperature, naphtha partial pressure, and space time on the product yields in naphtha pyrolysia and to develop a pyrolysis model which could be used to predict the major product yields including the aromatics.

Experimental Section The pyrolysis experiments were conducted in a continuous annular tubular reactor. A schematic diagram is shown in Figure 1. The annular reactor was constructed by inserting a 7 mm 0.d. stainless steel tube inside a 15 mm i.d. stainless steel tube. The length of the reactor was 60 cm. The inside tube served as a thermowell, in which a chromel-alumel thermocouple was moved to measure the axial profile. The reactor was placed inside a 4-kW furnace, which had provision for independently varying the heat input to the top and bottom sections; the temperature in each section was regulated with proportional controllers. Steam, which was uged as an inert, was generated in a vaporizer and mixed with the naphtha before the preheater. To avoid cracking in the preheat section the temperature of the preheated mixture was kept below 550 "C, and this mixture was then fed to the reactor. The effluent from the reactor was quenched in two water-cooled condensers placed in series. The noncondensables were passed through a sampling valve, a wet testmeter, and then vented. The gas and liquid samples were further analyzed. To measure the amount of coke deposited in edch run, after the pyrolysis run was completed, the reactor was flushed with nitrogen and then the coke was burnt with
0 1985 American Chemical Society

Ind. Eng. Chem. Process Des. Dev., Vol. 24, No. 3, 1985 775

mocouplo wells

Table I. Characteristics of the Naphtha Feed (i) density (15 "C) 0.701 e/mL .. (ii) ASTM histillition IBP 47 O C 5% 59 O C 10% 63 "C 50 % 87 o c 90% 119 o c 95% 129 O C FBP 156 "C (iii) PONA normal paraffins 33.7 wt % 73.6 wt % total paraffins naphthenes 17.4w t % aromatics 9.0wt % (iv) total sulfur 0.05 w t % 91 average molecular weight
I ,

Figure 1. Schematic of the pyrolysis equipment.

dried and preheated air. The flue gases, consisting of CO, C02,H20, and N2,were passed through a catalytic packed bed reactor, which oxidized the carbon monoxide to carbon dioxide. For this purpose, a 15 mm i.d. stainless steel tube of 300 mm length was filled with cupric oxide. The temperature of the converter was maintained at 250-300 "C. The effluent gases from the catalytic converter were then passed through a magnesium perchlorate column to absorb the water vapor and then through a 40% NaOH solution to absorb the carbon dioxide. The amount of water absorbed was determined gravimetrically and the carbon dioxide by double titration. A similar method for measuring the deposited coke has been used by Newsome and Leftin (1979). Gas a n d Liquid Analysis. The gaseous products formed in the pyrolysis were analyzed on a gas chromatograph using a thermal conductivity detector. A 3 mm i.d., 3 m long Duropak column operated at 45 "C was used to analyze the C3and C4 gases. Hydrogen, methane, ethane, and ethylene could not be completely separated on this column and were analyzed on a 3 mm i.d., 2 m long Poropak T column also operated a t 45 "C. Due to the complexity of the liquid producta and the nonavailability of a suitable capillary column, detailed analysis of the liquid fraction was not possible. The aromatic content of the liquid was determined on a 3 mm i.d., 3 m long column packed with 25% diethylene glycol succinate on Chromosorb P. In addition, the liquid products collected in each run were separated into three fractions having the boiling range of IBP -156, 156-200, and >200 "C, respectively. The density of each fraction was also determined. Results a n d Discussion In this study, the effect of space time, temperature, and inlet steam to naphtha ratio on the conversion and product yields in naphtha pyrolysis at atmospheric pressure was investigated. The experiments covered the following range of variables: temperature, 993 to 1073 K; naphtha flow rate, 0.5 to 30 cm3/min, and thus a space time of 0.024 to 0.76 s; steam to naphtha ratio, 1.43 to 0.71 (kg/kg). The naphtha was obtained from Bharat Refineries, Bombay, which possesses Arabian crude, and the feed characteristics are given in Table I. The analysis of the experimental data in any pyrolysis study is rendered difficult by the inevitable temperature profile which exists in the reactor. The concept of equivalent reactor volume as introduced by Hougen and Watson (1947) and applied by Van Damme et al. (1975, 1981) was used to analyze the data by a pseudoisothermal approach. Another problem in the kinetic analysis is the definition of conversion. Some investigators have used

methane yields as a measure of conversion; others have defined the conversion on the basis of various severity functions (e.g., Zdonik et al., 1970; Illes et al. 1976; Shu and Ross, 1982) whereas Van Damme et al. (1981) based the naphtha conversion on the disappearance of 15 main components present in the feed. Since, in this study, due to experimental limitations, the detailed analysis of the product liquid and the feed could not be done, an intermediate approach was adopted. The naphtha conversion based on the total feed was defined as follows wt % conversion = w t % gaseous products + (wt % of total aromatics in liquid effluent wt % aromatics in feed) + (wt % of liquid products boiling above 156 "C w t % of aromatics boiling above 156 "C) (1) where 156 "C was the FBP of the feed. This definition of conversion does not account for the decomposition products, other than the aromatics, which have the same boiling range as that of the feed. The error in neglecting these products is not expected to be significant because at low conversions (<30%), conversion is approximately the same as the gaseous yields; i.e., the liquid effluent is essentially unconverted feed. This has also been reported by Hirato and Yoshioka (1973) and can also be concluded from the data published by Van Damme et al. (1981). On the other hand, at high conversions the amount of aromatic-free liquid effluent boiling in the same range as the feed is not appreciable. The variation of naphtha conversion with VE/Fo a for steam/naphtha ratio of 1.43 kg/kg, in the temperature range 993-1073 K, is shown in Figure 2, where VE is the equivalent reactor volume. At a constant temperature, the conversion depends on both VE/Fo and the dilution ratio (kg of steam/kg of naphtha) as is also shown in Figure 2. At a fixed temperature and VE/Fo,increasing the dilution ratio reduces the conversion. Determination of t h e Overall Rate Constant. The overall kinetics for naphtha pyrolysis was determined by calculating the equivalent reactor volume by a pseudoisothermal approach. The equivalent reactor volume, VE, is defined as the volume, which at a reference temperature, TR, would give the same conversion as the actual reactor with its temperature profile (Hougen and Watson, 1947). Thus, we have

For each run, the actual temperature profile was measured at 3-cm intervals and a typical temperature profile is shown in Figure 3. Due to heat losses, the temperature

776

Ind. Eng. Chem.

Process Des. Dev., Vol. 24, No. 3, 1985

65

J

Temp H StearrJmphtt 993 143 (kgh' 0 1013 1 43 h 1033 'A3
T

o
0

1053 1073 1073 1073

143 1 L3 095 071

a io73 v 1073
'0
LO

035 071

80

120

160

200

2bO

280

320

3MI

V/F,(

mss)/kg mole

Figure 2. Variation of naphtha conversion with V,/F,.
Equivalent space tlme s

Figure 4. Determination of the overall first-order rate constant a t
different temperatures.
Y increased from 3.35 at approximately 25% conversion to 3.95 at approximately 90% conversion. The aromatics presents in naphtha are very refractory and in general do not undergo significant pyrolysis until 1073 K. In this analysis, the aromatics present in the feed have been assumed to act as inerts and the overall rate constants have been calculated on an aromatic-free basis. The conversions used in eq 3 were based on aromatic-freenaphtha and were calculated from eq 1by dividing the total conversion by 0.91 (9% aromatics in feed). e on an aromatic-free basis was determined experimentally for all the runs. Figure 4 shows the plot of the right-hand side of eq 3 as a function of space time. The plots for various temperatures were straight lines for the whole range of conversion, indicating that the rate constant is independent of conversion. Van Damme et al. (1981) also reported the first-order rate constant for naphtha pyrolysis to be independent of conversion. Varying the initial partial pressure of naphtha does not significantly change the slope of the line. The rate constants were independent of partial pressures at other temperatures also, thus confirming the assumption that naphtha pyrolysis can be represented by first-order kinetics. The activation energy and frequency factors were obtained from an Arrhenius plot as shown in Figure 5. The overall decomposition of aromatic free naphtha can be represented with an activation energy of 220.80 kJ/gmol (52 600 kcal/g-mol) and a preexponential factor of 6.6 X 10" s-l. The activation energy and preexponential factor are in the same range as reported by Van Damme et al. (1981) and Hirato and Yoshioka (1973) for naphtha pyrolysis. The validity of the equivalent reactor volume analysis was further checked by using these values of the preexponential factor and the activation energy, together with the measured axial temperature profile, to calculate the conversion in the nonisothermal reactor by numerically integrating the reactant continuity equation. In all cases, the calculated conversions were within *3% of the conversions determined experimentally. Product Distribution. The effect of temperature, partial prwure, and space time on the product yields and composition was investigated. As expected, the major gaseous products were methane, ethylene, propylene, 1,3-

600 7001

LOO0

500L
12
'26

36

L8

60

Diotcnce t r o m tho top o t tho rocctor.cm

Figure 3. Typical axial temperature profile in the tubular reactor.

gradients a t the inlet and exit were of the order of 25 to 35 K/cm. To utilize the equivalent reactor volume concept, a good estimate of the activation energy is required. The activation energy was assumed to be 218.4 kJ/g-mol (52000 cal/g-mol), which was very close to the value subsequently obtained from an Arrhenius plot so that a second iteration was not required. It is well-known that the pyrolysis of pure hydrocarbons follows a first-order rate law and naphtha pyrolysis was also assumed to be first order. This assumption was later checked. For a first-order, irreversible reaction A v products taking place in a plug flow reactor, Levenspiel (1974) has derived the following equation

-

Jr = z

(1

+ e) In -- e 1-x x

(3)

where r is the space time and was calculated using the equivalent reactor volume; e is the expansion factor and represents the relative change in the volume of the reaction mixture from zero to 100% conversion and includes the effect of the diluents. e can be related to v and 6 as follows
e=-

v-1 l + b

(4)

Ind. Eng. Chem. Process Des. Dev., Vol. 24, No. 3, 1985

777

02-

'
0

1073 1073

095 071

O;
692 093 094 095 0 9 6 097 (1iT)x IO3(]i K ) 098 099 100 101

,b

:o

io

Lo

o :

$0

0 :

8 1 0

s

Convorsion ,%

Figure 7. Variation of product selectivities with conversion at different temperatures and inlet steam to naphtha ratio.

Figure 5. Arrhenius plot for the pyrolysis of naphtha.

b

o
A
0

Conversion ~ o t o pyrolysis gas i Ethylene Propylene Methone

0

0

005

01

015

020

025

030

035

OLO

Space t i m e , s

Figure 6. Variation of conversion and product yields with space time at 1073 K. Inlet steam to naphtha ratio = 1.43 kg/kg.

butadiene, 1-butene, and hydrogen. The aromatic yield increased with conversion and (including the aromatics present in the feed), constituted 9 to 20% of the total feed. For a given conversion, the total pyrolysis gas yield did not significantly depend on temperature. The gas yields varied from 20% at 20% conversion to approximately 73% at 90% conversion; i.e., at low conversions the liquid product was mainly unconverted feed. Figure 6 show the variation of the conversion and the yields of the major producta with space time at 1073 K. The yields of thermally stable methane, hydrogen, and aromatics increase monotonically with conversion, whereas propylene, l,&butadiene, and 1-butene show maxima with increasing space time. The concentration of olefinic gases decreases due to the secondary pyrolysis and polymerization reactions. It is ex-

pected that at higher conversion, ethylene yield will also pass through a maximum; however, due to experimental limitations this maximum could not be observed in this study. The effect of inlet naphtha partial pressure on the product yield was not appreciable. Ethylene yield increased, whereas methylene yield decreased as the steam dilution was increased. For example, at 1073 K and 85% conversion the ethylene yield decreased from 25.5 to 23.5 w t % as the steam/naphtha ratio was changed from 1.43 to 0.71 kg/kg. There was no effect of inlet naphtha partial pressure on propylene yield until the maxima; beyond that point increasing the steam dilution decreased the propylene yield. The selectivity of a product is defined as the number of moles of component formed per mole of raw material pyrolyzed. The effects of conversion, temperature, and partial pressure on selectivities of various products were experimentally determined and it was found that for all components the selectivities converged to a constant value at low conversions. There was no significant effect of partial pressure and temperature on initial selectivities. Typical results for ethylene and propylene are shown in Figure 7. For all the runs, the initial selectivities of ethylene and propylene can be represented as 0.88 f 0.02 and 0.60 f 0.01, respectively. Similar trends have been observed for pure hydrocarbons by Tanaka et al. (1976b). The initial selectivities of all the products are shown in Table 11. An attempt was made to predict the initial selectivities using R-K (Rice-Kossiakoff) theory (1943) as modified by Murata and Saito (1974). An average carbon number of 6.5 was calculated for the feed. The initial selectivitiesfor 6 and 7 carbon n-paraffms and isoparaffins were then calculated. For naphthene pyrolysis, since no rules for predicting the initial selectivities are available, the initial selectivities were estimated from published literature values. The overall selectivities for naphtha were estimated by weighting the selectivities for the individual hydrocarbons by their respective mole fractions. Values thus obtained are also reported in Table 1 . In some cases, 1 the differences between the experimental and calculated selectivities are appreciable. However, it is possible that

778

Ind. Eng. Chem. Process Des. Dev., Vol. 24, No. 3, 1985

Table 11. Initial Selectivities of the Primary Products Obtained in Naphtha Pyrolysis
component hydrogen methane ethane ethylene propane propylene n-butane l-butene 1,3-butadiene other Cis isobutane l-pentene 2-pentene 3-methyl- 1-butene l-hexene 2-hexene 4-methyl-l-pentene selectivity exptl calcd 0.58 0.44 0.68 0.45 0.10 0.011 0.88 0.73 0.02 0.026 0.60 0.49 0.035 0.2 0.14 0.07 0.026 0.09 0.13 0.069 0.036 0.042 0.027 0.019 0.014

with a detailed analysis of the feed a better estimate of the initial selectivities can be obtained without resorting to experiments. Model Development. Due to the incomplete knowledge of free-radical reactions, a mechanistic model for naphtha pyrolysis was not attempted. For developing a model for naphtha pyrolysis, analogy was taken from modeling of pure hydrocarbon pyrolysis using molecular reaction schemes. In such models, the pyrolysis is represented by a primary step followed by a set of molecular reactions between the primary products. Such an approach has been successfully applied in modeling the pyrolysis of ethane, propane, butane, and their mixtures (Sundaram and Froment, 1977) and extended up to C8 normal and branched paraffins (Murata and Saito, 1975; Tanaka et al., 1976a,b). For pure components, the selectivities in the primary decomposition step are usually calculated by using R-K theory or its extension (Rice and Kossaikoff, 1943; Murata et al., 1973; Murata and Saito, 1974). For this study, it was assumed that naphtha could be represented as a pseudo-pure compound and the primary decomposition of naphtha was represented by a single reaction with the initial selectivities determined experimentally. Based on the experimental results, the primary reaction was represented by an overall firsborder reaction for the whole range of conversion and the initial selectivities were assumed to be constant. The secondary reactions of the primary products were also represented by molecular reactions. It has been further assumed that the aromatics in the feed do not react. The first-order rate constant for the primary reaction was calculated using a pseudoisothermal approach as discussed earlier. The primary reaction can thus be represented as naphtha 0.58H2 + 0.68CH4 + O.88CzH4 + O.lCzH6 + 0.6C3H6 + 0.02C3H8 + 0.035C4Hlo + 0.2C4Hs 0.07C4H6 -k 0.09C4S

ki

In the pyrolysis of hydrocarbons, an innumerable number of secondary reactions can take place between the various primary products. An attempt has been made to include the important secondary reactions that wl account for the il major products obtained in naphtha pyrolysis. Sundaram and Froment (1977) have published a set of secondary reactions which accurately predict the product yields in the pyrolysis of mixtures of ethane, propane, and n-butane. For our purpose, this set was modified by deleting the reactions of i-C4H8,2-C4H8,and C3H4,because these com-

ponents were negligible and could not be identified properly. Additional reactions were included for aromatic formation and the depletion of l-butene and l,&butadiene. The probable reactions of l-butene were based on the reaction scheme proposed by Murata and Saito (1975) and Tanaka et al. (197613). Aromatic formation reactions were estimated based on the reaction scheme proposed for hexane pyrolysis (Isbern et al., 1981) and for the high conversion pyrolysis of n-paraffins (Murata and Saito, 1975). The final set of secondary reactions account for ethylene formation from the dehydrogenation of ethane and n-butane, the scission of propane and n-butane, and from propylene. Propylene can be formed from the reversible dehydrogenation of propane, the reaction of ethylene with ethane or propane, and from the scission of n-butane to propylene and methane. The decrease in propylene selectivity is accounted for by its reaction with ethane to form l-butene and methane, the formation of aromatics, c,+, and methane or ethylene from proylene. Butadiene formation is represented by the reaction of acetylene with ethylene and the dehydrogenation of l-butene, whereas l-butene formation is accounted mainly from the dehydrogenation of n-butane and the reaction of propylene and ethane to form l-butene and methane. The monotonic increase in methane yields is represented by several reactions such as the scission of propylene to acetylene and methane, the disproportionation of ethane, scission of propane and n-butane, and the reaction of ethane with ethylene or propylene. In this scheme, the c6+ fraction is formed either from propylene or l-butene. Aromatics are also formed in these two reactions. In addition, aromatics can also be formed by the reaction of butadiene with ethylene, propylene, l-butene, or from two moles of butadiene. The final set of reactions, together with the preexponential factor and activation energy for each reaction, is shown in Table 111. In reaction numbers 11 and 17 of 6 Table 111, c + denotes the aromatics-free product boiling above the end point of the feed. The equilibrium constants for the reversible reactions were calculated by standard procedures and checked with the published values. The reactions of other Cq)s were not included because these were not properly identified and their yields were also low. With this set of reactions, the continuity equations for the 14 components were written and solved using a sixth-order Runge-Kutta-Verner method. The experimental and predicted product yields were compared and the rate constants adjusted, by trial-and-error, to minimize the deviation between the experimental and predicted yields. In order to match the experimental c +yields, for 6 reaction numbers 1 and 17, the ratio of the stoichiometric 1 coefficients of c,+ to aromatics was adjusted to 1:2.14. The actual stoichiometric coefficientsof C,+ and aromatics in these two reactions was determined from a carbon balance. The average carbon number for the aromatic fraction and c,+ fraction was assumed to be 6.5 and 7.0, respectively. Moreover, the average molecular weights of c,+ and aromatics were taken to be 91 and 86, respectively. The best results for all the experimental conditions investigated were obtained using the reaction scheme shown in Table 1 1 Compared to the reaction scheme of Sun1. daram and Froment (19771, the activation energies and preexponential factors of reaction numbers 3, 10, 11, and 17 had to be modified to satisfactorily simulate the variation of ethylene, propylene, and l-butene yields with temperature. In addition, the activation energies and

Ind. Eng. Chem. Process Des. Dev., Vol. 24, No. 3, 1985 779

Slaomlnophtho :143(kq/kq)

4
993
0

Steam / naphtha

=I3 ( k g / kg) L

A
V

n
I
I I I

0

1013 1033 1053 1073 Simulated
I I I

0
1 I
I I

01

02

03

1 01,

I

I

05

06

I 07

I

08

I 3

Space t i m e .

s

0

01

02

03

04 05 Spoce lima.

06

07

08

09

s

Figure 10. Predicted and experimental propylene yields vs. space time at different temperatures.
Steam/nophtha=l L 3 ( k g / k g )
L

Figure 8. Predicted and experimental ethylene yields w. space time

a t different temperatures.
5-

, Ter$3K
0

1013

1

A
V

13 03
1053 1073
Simulated

Steom/nophtho:l

L3(kg/kq)

/

-

0

Temp K

, /

1 ;:
h

1033
1053 1073 Simulated
I

9
Spoce t i m e , s

0
0

0
I I

Figure 11, Predicted and experimental l-butene yields vs. space time at different temperatures.
I

1

0

01

02

03

04 05 06 Spoce tima.s

07

08

I

Figure 9. Predicted and experimental methane yields vs. space time at different temperatures.

preexponential factors for aromatic formation reactions were also changed. The initial wsumed values and the final adjusted values of the kinetic parameters that were modified are given in Table IV. For the other reactions, the initially assumed values were the same as the final values given in Table 1 1 Some reactions, viz. 1. C4H6 UC6+ + b aromatics l-C4H8 2C2H4 l-C4H8 H2 0.73(CH4 + CSHJ oe27(C2H4+ C2H6) were deleted, because even with various combination of rate constants, the product distribution was adversely affectkd. Model Prediction. The simulation model developed was used to compare the experimental and predicted yields

+

- -

+

at various temperatures and partial pressures of naphtha. Figures 8 to 15 show the comparison between the experimental and calculated yields of the major products at different temperatures for a steam/naphtha ratio of 1.43 (kg/kg): As can be seen from these figures, there is very good agreement between the experimental and predicted yields and, for most of the components, the predictions were within f5% of the experimental yields. The product yields were predicted with a similar accuracy at steam to naphtha ratio of 0.95 and 0.71 (kg/kg) also. For these comparisons, the pseudoisothermal approach was used which is strictly valid for a simple reaction. To check the validity of the equivalent reactor volume concept, the product yields using the actual temperature profiles were also calculated and compared with the isothermal product yields. Comparisons for methane, ethylene, and propylene are shown in Figure 16. For other components, the difference was even smaller. The sensitivity of the model for the rates of secondary reactions was also checked. It was observed that the model is not sensitive to the rates of secondary reactions. No

780

I d . Eng. Chem. Process Des. Dev., Vol. 24, No. 3, 1985

Table 111. Reaction Scheme for Naphtha Pyrolysis
no. 1. equation" naphtha 0.58Hz + 0.68CH4 + 0.88CzH4+ 0.1CzH6+ 0.6C3H6+ 0.02C3H~ 0.035C4H10 + 0.2C4H~+ 0.07C4H6 + 0.09CiS CzH6 + C2H4 + Hz C3H6 CzH2 + CH4 C2Hz + CzH4 C4H6 2CzH6 C3H8+ CHI CzH4 CzH6 C3H6 CH4 C3HB + C3H6 + Hz C3H8 CzH4 + CH4 C3H8 + CzH4 CzH6 + C3H6 2C3H6 3CzH4 2C3& 0.3CnHzn4+ 0.14c6' + 3CH4 C3H6 + C2H6 1-CdHs + CH4 n-C4Hio CsH6 + CH, n-C4H1o 2C2H4 + Hz n-C4Hio -*CzH4 + CzH6 n-C4Hlo+ l-C4Hs + H2 1-C4Ha O.41C"Hz, + 0.19c6+ l-C4HB H2 + C4H6 * C2H4 + C4H6 B -I2Hz C4H6 + C3H6 T + 2H2 C4H6 + l-C4H8--C EB + 2Hz C4H6 + C4H6 ST + 2Hz

-

E, kcal/g-mol 52.58
65.21 65.33 41.26 65.25 60.43 51.29 50.6 59.06 64.17 56.90 60.01 59.64 70.68 61.31 62.36 50.73 50.00 34.56 35.64 57.97 29.76

ko, 8-l 6.565 X 10" 4.652 X 1013 7.284 X 10l2 (1.026 X 1015)b 3.75 x 10'2 (7.083 X 1016)b 5.888 X 1O'O 4.692 X 1O'O (2.536 X 1016)b 7.386 X lor2 2.424 X 10" (1.0x 1017~ 7.0 X 10l2 7.0 x 1014 4.099 X 10l2 1.637 X 10l2 2.075 X 10" 1.0 x 10'0 (8.385 X (9.74 x 10")b (6.4 x 1 0 9 5 (1.51 X

2. 3.
4.

4

+

-

+

+

+

"B:benzene; T: toluene; EB: ethylbenzene; ST: styrene. *Units: cm3/(mol 8).

I

Steam1 naphtha

i

1 L3 (kgl k g )

I

I

o

01

02

03

I I 05 c6 Space time s

I 06

Figure 12. Predicted and experimental 1,3-butadiene yields vs.
space time a t different temperatures.

01 0

/

I

I

01

02

03

1 OL

Space l i m e s

Figure 13. Predicted and experimental hydrogen yields VB. space time at different temperatures.

+ -

+ -

+ -

5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22.

- -

+

-+

-

I

Table IV. Initial Estimates and Final Adjusted Values of the Kinetic Parameters
initial estimate

E,
reaction no. 3 10
11

final value

17 19 20 21 22

kcal/gmol 59.39 55.80 45.50 29.84 28.80 29.70 52.70 24.80

ko, 8-l 3.794 x 10"
1.514 X 10" 1.423 X lo8 7.685 X lo6 (9.58 X 10'O)" (1.02 X (9.36 X ,OU)" (2.57 X 1O'O)"

kcal/mol 65.33 64.17 56.90 50.73 34.56 35.64 57.97 29.76

E,

ko, s-' 7.284 X 1OI2 7.386 X 10l2 2.424 X 10" 2.075 X 10" (8.385 X (9.74 x 10")" (6.4 x 10q0 (1.51 X 10")"

"Units: cm3/(mol 8).

I
07

I

1

oa

03

I

05

I 06

I

07

08

09

Figure 14. Predicted and experimental aromatic yields vs. space time a t different temperatures.

significant change was observed by changing the rates of each of the secondary reactions by *lo%. However, the

0

TOIOI

O

/

gas

80-

o
A

Total oletins C2HL

Temp I 1073 K Steam I nophtha = 0 6 ( k g I kg )

l

50

i

30

10
I

3

'1
0
Space time.

&,
002 004

o-(benzene.toluene) proedcec
I

s

0

008

016 032 0 6 L Residence time.s

128

256

512

Figure 15. Predicted and experimental C + yields vs. space time , a t different temperatures.

Figure 17. Predicted and published experimental product yield vs.
residence time for naphtha pyrolysis.

I-I
2L

o Total gos

Temp ~ 1 0 7 3 K Steam1 kerosene=Ol (kgIkg)

TRmp = t 0 1 3 K , S t e o m I n o p h t h a : l W ( k g I k g ) protila

-___ temperature Actual
- Equivalant

Totol olefins 0 C2HL
0

Isothermal tamparalure

, 18 t
c
1

3

p >

15

12

9
6
l-C4Hg

3
l 82 I

A 1,3-~41ig
1
0 - (benzene +taiuene)

Total oromaticsoromotics in teed

0 0

produced

10

20

30

LO

50

60

70

80

I
L -

Convarslon 7.

Figure 16. Predicted methane, ethylene, and propylene yields vs.
conversion using the pseudoisothermal approach and actual temperature profile.

I

I

I

Residence time, s
~

model is quite sensitive to initial selectivities of the components. Therefore, any error in calculating the initial selectivities may lead to poor estimates of the product yields. To further check the generality of the model, the experimental results of Hirato and Yoshioka (1973) for naphtha, kerosene, and gas oil pyrolysis were also simulated. The initial selectivities and rate constants reported by these authors for each feed were used for the primary reaction step, whereas the set of secondary reactions was the same as given in Table 1 1 1. Figures 17-19 show the comparison between their experimental yields and predicted yields by our model. As can be seen from these figures, the predicted yields for most of the major components are in very good agreement with the experimental values. The experimental yields for the total aroamtics were not available and hence could not

Figure 18. Predicted and published experimental product yield vs.
residence time for kerosene pyrolysis.

be compared. In the original paper, the yields were reported as a function of the residence time calculated at the outlet conditions and therefore the simulated results have also been calculated on the basis of this definition. Compared to the model proposed by Hirato and Yoshioka (1973),our model predictions are as good asor even better than their predictions. This indicates that the secondary reactions are more or less independent of the nature of the feedstock and may be utilized for other hydrocarbon pyrolysis also. This virtual independence of secondary reactions was also observed by Hirato and Yoshioka (1973). However, secondary reactions considered by Hirato and Yoshioka (1973) were gross molecular reactions, which are

782

I d . Eng. Chem. Process Des. Dev., Vol. 24, No. 3, 1985
Temp =1073K S t e a m l g o s i C 7(kglkgl
0 1 1

Such a model together with an expression for the coking kinetics (Kumar and Kunzru, 1985) can also be used for modeling the coke profiles in naphtha crackers. Nomenclature E = activation energy, kJ/kmol or kcal/g-mol Fo = molar feed rate of naphtha, kmol/s k = rate constant, l / s R = gas constant T = process gas temperature, K TR = reference temperature, K V = reactor volume, m3 VE = equivalent reactor volume, m3 X = fractional conversion based on aromatic-free naphtha
Greek Letters
6 = molar dilution ratio, mol of inert/mol of reactant

’*t ’
0

e
Total a r o m a t i c s a r o m a t i c s i n teed (benzene toluene), produced
T
0

v = stoichiometric coefficient; mol of product/mol of naphtha

= expansion factor = equivalent space time, s

cracked

Literature Cited
4 -

0 001
1

002

004

008

016 032 064 Residonce t i m e . 5

128

256

5

Figure 19. Predicted and published experimental product yields w. residence time for gas oil pyrolysis.

limited to a narrow range of operating conditions. Other dekiiled models for naphtha pyrolysis do not disclose the reaction scheme because of proprietory reasons. The advantage of this model is that the yields of the major products of naphtha pyrolysis can be predicted by a relatively small number of reactions. This reduces the computational effort considerably. Compared to gross molecular schemes, the individual molecular reactions reflect the true nature of the reactions better and increase the range of applicability of the model. A limitation of this model is that the initial selectivitieshave to be determined experimentally. Conclusions From the above discussion, it can be concluded that the pyrolysis of naphtha can be satisfactorily modeled by an overall first-ordei primary step and a set of secondary reactions, which are virtually independent of the feed stock. Such a model has wide applicability in design of cracking coils. I t can also be used to predict the product distribution, provided the initial selectivities, rate constant, and preexponential factor for the primary step are known.

Aral, Y.; Murata, M.; Tanake, S.; Salto, S . J . Chem. Eng. Jpn. 1977, 10, 303. Chambers, L. E.; Potter, W. S. tfyhcarbon Ibocess. 1974, 53. 121. Ennls, 8. P.; Byod, H. 6.; Oriss, R. Chem. Techno/. 1975, 5, 693. 57, 227. Gwsens, A. G.; Dente, M.; Ranzi, E. &&ocarbon process. 1@78, Hlrato, M.; Yoshloka, S. Int. Chem. Eng. 1973, 13, 347. Hougen, 0. A.; Watson, K. M. “Chsmlcal Rbcess Prlndples”; WHey: New York, Part 111, 1947; p 884. Illes, V.; Salal, 0.; Csermely, Z.ACS Symp. Ser. 1978, 32, 423. Isbern, G.; Ederer, H. J.; E M , K. H. I n “Modeling of Chemical Reactlon Systems”, Ebert, K. H.; Deuflhard, P.; Jager, W., Ed.; Springer-Verlag: Heidelberg, 1981; p 235. Kumar, P.; Kunzru. D. Can. J . Chem. Eng.,in press. Levenspiel, 0. “Chemical Reactlon Englneering”; Wley: New Delhi. 1974. Amano, A.; Maeda, S. J . Chem. Eng. Jpn. 1973, 6 , Murata, M.; Salto, S.; 252. Myata. M.; Takeda. N.; %no, S. J . & e m . EM. Jpn. 1974, 7 , 286. Mbrata, M.; Saito, S. J . Chem. Eng. Jpn. 1874, 7 , 389. Murata, M.; Salto, S. J . Chem. Eng. Jpn. 1875, 8 , 39. Newsome, D. S.; Leftln, H. P. “Cdthg Rates In a Laboratoty Pyrolysis Furnace”; paper presented at 72nd Annual AIChE Meeting, San Franclsco, Nov 1979. Rice, F. 0.; Kosslakoff, A. J . Am. Chem. Soc, 1943, 65, 590. Sundaram, K. M.; Fromnt, 0. F. Chem. En#. Scl. 1977, 32, 609. Shu, R. W.; ROSS, L. L. Ind. Eng. Chem. Rocess D e s . Dev. 1982.21, 371. Tanaka, S. J . Chem. Eng. Jpn. 1975, 8 , 3059. Tanaka, S.;Aral, Y.; Salto, S.J. Chem. Eng. Jpn. 1976a, 9 . 161. Tanakq, S.; Arai, Y.; Saito. S. J . Chem. Eng. Jpn. 197Sb, 9 , 504. . Froment, G. F. AIChEJ. 1975, 6 , 1065. Van Damme, P. S : Narayanan. S.; Froment, G. F.; Balthasar, W. E. I n d . Eng. Chem. ProVan Damme, P. S.; cess Des. D e v . 1981, 20, 366. Zdonik, S.E.; Green, E. J.; Halle, L. P. 011 Cies J . Jucn 26, 1967, 96. Zdonik, S. 6.; Green, E. J.; Hallee, L. P. “Manufacturing Ethylene”; The Petroleum Publishing Co.;Tulsa, OK, 1970.

Received for review August 8, 1983 Revised manuscript received August 13, 1984 Accepted August 29,1984


相关文章:
英文期刊缩写与全称对照(J)
Pyrolysis Journal of Analytical and Applied Pyrolysis J. Anal. At. Spectrom...Journal of Molecular Modeling J. Mol. Spectrosc. Journal of Molecular ...
专题论文
Reduction of Format ion during Naphtha Pyrolysis Using Triethyl Phosphit e. Ind. Eng.Chem.Res., 1988, 4(27):559-565. [10]Lyle F. Albright , ...
主要石油系统中英对照表
pyrolysis aviation gasoline residual oil motor ...naphtha platforming reactor B-K dewaxing cracked ...modeling SBR ethanol 乙烷 乙苯,苯乙烷 乙烯 环氧...
Synthesis-of-Au-ZnO-and-Pt-ZnO-nanocomposites-by-one-step-flame-spray-pyrolysis-and-it
Zinc naphthanate (Aldrich, <50% in mineral spir-its) and gold (III) ...The particle morphology was observed using JEOL Model JEM-2010 transmission ...
更多相关标签:
journal of pyrolysis | naphtha | naphtha price | naphthas | naphthate 酸 | pyrolysis | spray pyrolysis | biomass pyrolysis |