Sensitivity analysis and optimization of a toluene recovery system

F.V.O. Nr:

28:s

Cl Technische Universiteit Delft

Vakgroep Chemische Procestechnologie

Verslag behorende bij het fabrieksvoorontwerp

van

J.J. Feddes P ,J", Ferwerda

onderwerp:

SENSITIVITY ANALYSIS AND OPTIMIZATION OF

A TOLUENE

RECOVERY SYSTEW

Gasthuissteeg 15 2611 RH Delft

Jacoba van Beierenlaan 53 2613 JA Delft

opdrachtdatum: November '89 verslagdatum: April '90

'....)

TECHNICAL UNIVERSITY of DELFT

DEPARTMENT of CHEMICAL TECHNOLOGY

SENSITIVITY ANALYSIS AND OPTIMIZATION

OF A TOLUENE RECOVERY SYSTEM

J.J. FEDDES

P.J. FERWERDA

April, 1990

. ./ -' TABLE OF CONTENTS PREFACE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . lil SUMMARY . . . .. . . .. .. . . .. .. . ... ... . IV 1 INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 1 2 PRINCIPLES IN FLOWSHEETING . . . .. . . .... .. . 2 2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 2

2.2 The sequential method . . . . . . . . . . . . . . . . . . . . . . . . . .. 2

2.3 The simultaneous modular method . . . .. .. . . .. 3

2.4 The equation-oriented method .. . .. .. .... . ... .... . . . .. .. . . 3

3 THE TOLUENE RECOVERY SYSTEM . . . . . . . . . . . . . . . . .. 5

4 SIMULATION CALCULATIONS WITH CHEMCAD II . . . . . . . . . . . . . . .. 7

4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

4.2 Calculation of the feed composition . . . .. ... . . . 7

4.3 Result of the flash calculation . . . . . . . . . . . . . . . . . . . . . . . .. 11

4.4 Results of the distillation calculations . . . .. .... . 12

5 SENSITIVITY ANALYSIS OF THE TOLUENE RECOVERY SYSTEM ... 14

5.1 Distillation tower TW201 . . .. . ... . . .. ... .. . ... . . . .. 14

5.1.1 The effect of the feedstage position . . . 14

5.1.2 The effect of the pressure . . . .. ... . .... ... .. 16

5.1.3 The effect of the number of stages ... . . . .. . .. . . . . . 18

5.2 Distillation tower TW202 .... .. . . .. ... . .. . . .. .... . .. . 21

5.2.1 The effect of the reflux ratio .... . .. . . .. . . . .. ... . 21

5.2.2 The effect of the bottom flow . . . . . . . . . . . . . . .. 24

5.2.3 The effect of the pressure ... . . .. .. . . 25

5.2.4 The effect of the feedstage position . ... .. . . . ... . . . 27

5.3 Distillation tower TW203 . .. . . .... . .. . .. ... ... . . .. . 28

5.3.1 The effect of the reflux ratio ... . ... . . .... . . .. ... 29

5.3.2 The effect of the bottom flow. . . . . . . . . . . . . . . . .. 31

5.3.3

The effect of the number of stages and the feedstage position . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

5.4 Conclusions from the sensitivity analyses . . . . . . . . . . . . . . . . . . . .. 33

6 OPTIMIZATION OF THE TOLUENE RECOVERY SYSTEM . . . .. .. 35

6.1 Calculation of the flash feed composition with ASPEN + . . . 35

6.2 Optimization of distillation tower TW201 . . . 38

6.3 Optimization of the recovery system . . . 39

7 COLUMN DESIGN . . . 41

8 COSTS CALCULA TION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 42

9 LITERATURE ... . . .... . . 46

10 SYMBOLS . . . 47

11 TABLE INDEX. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 49

12 FIGURE INDEX. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 51

APPENDIX A: FLOWSHEET AND GIVEN PROCESS DATA. . . . 52

APPENDIX B: RESULTS OF THE SIMULA TION CALCULA TION WITH ASPEN+ . . . 55

APPENDIX C: COMPOSITION OF THE STREAMS IN THE FLASHSECTION. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 57

APPENDIX D: COMPOSITION OF THE OPTIMISED PRODUCTSTREAMS OF COLUMN TW201. . . . 59

APPENDIX E: COMPOSITION OF THE OPTIMISED STREAMS OF THE RECOVERY SECTION . . . 60

APPENDIX F: MASSA EN WARMTEBALANSEN. . . . 63

APPENDIX G: SPECIFICATIE WARMTEWISSELAARS H 1 EN H2 . . . , 66

APPENDIX H: OVERZICHT WARMTEWISSELAARS EN SMOORKLEPPEN. . . . . . . . . . . . . . . . . . . . . . .. 68

PREFACE

This report contains the work done in accordance with the preliminary plant design for students of the Department of Chemical Technology of the Technical University of Delft.

The subject was accomplished in corporation with Pyrotec B.V. at Zoetermeer and concerns the analysis of a toluene recovery system. The work was done over the period december 1989 - april 1990.

We'd like to thank J.H. Bredée for his assistance while working with the ASPEN + flowsheeting package.

SUMMARY

The sensitivity analysis and optimization calculations of the toluene recovery system were carried out with ASPEN

+

(release 8.2). It is possible 10 simulate the process and to meet the given specifications, except for the C₈-specification.

The toluene production is 433.2 [kmoljhr] (3.5 105 [tonjyear)) with a purity of 99.09 [mole%]. The total toluene recovery is 86.44 [mole%].

lINTRODUCTION

The purpose of this preliminary plant design is to analyze and to optimise a toluene recovery system. The toluene recovery system is defined by Pyrotec B.V. at Zoetermeer. Pyrotec B.V. is developing a TISFLO distillation tower subroutine, suggested by prof. Pierrucci (Milan, Italy). TISFLO is an equation-oriented method flowsheet package. For this purpose first the toluene recovery system had to be analyzed and optimised extensively. The results than can be used as reference while testing the TISFLO subroutine mentioned above.

The toluene recovery system consists of one flash vessel and three distillation towers straight forward. The purpose of the recovery system is 10 obtain high purity toluene. The product stream obtained from a catalytic reforming section contains hydrogen, i-alkanes, n-alkanes, naphthenes and aromatics. This gas/liquid mixture is separated in a flash vessel into a hydrogen rich vapor-fraction and a C₄ + liquid-fraction. The hydrogen fraction is recycled to the reforming section. Part of this fraction is purged to prevent build-up of inerts. The C₄

+

-fraction is separated by distillation in a LPG-fraction and a Cs

+

-fraction. The Cs

+

-fraction is than separated in a next column into a toluene poor fraction and heavier hydrocarbons. The toluene is isolated by distillation in a third column from the heavier hydrocarbons.

Specifications are given for the purity of the toluene (> 99.0 [mole%]) and a minimum overall loss of toluene is desired. Specifications are also given for the purity of the top and bottom product of the first distillation tower. The 10luene recovery system is described more in detail in chapter 3. Alternatives for the recovery system are described in "The extractive and azeotropic distillation technique in the separation of aromatics" of A. Harms.

The simulation calculations have been carried out with CHEMCAD II (version 2.3) and ASPEN + (release 8.2). The sensitivity analyses and optimization calculations bave been carried out with ASPEN +. Sensitivity analyses and optimization calculations are not as easy to carry out with CHEMCAD 11 as with ASPEN +, sa the system is

transferred from CHEMCAD II to ASPEN +. CHEMCAD II and AS PEN + are

simultaneous modular method flowsheet packages.

2 PRINCIPLES IN FLO\VSHEETING

2.1 Introduction

Flowsheeting is defined as performing calculations based upon flowsheet topology. This may range from simulation to optimization of plant operation, piping design and process unit sizing, costing, etc ..

A flowsheeting problem, by definition, involves the solution of a large set of equations related to the flowsheet topology and the unit opera ti on mode Is therein.

The possibilities of a program package for flowsheeting depend on the method of flowsheet ealculation. Three different methods are used for su eh ealculations:

- the sequential method;

- the simultaneous modular method; - the equation-oriented method.

2.2 The sequential method

The sequential method may be described as an automated manual eomputation method. The calculation starts at a known point in the flowsheet. The calculation order usually follows the logical flow direction. The mode Is used in th is method have the following form:

where,

o

=

f(l,f)

o

is the output flow data;

I

is the input flow data;

f

are unit parameters.

For the unknown recycle flows values are estimated. This procedure yields calculated values for those flows. On the basis of the difference between the calculated and estimated values, new estimates are made and the calculation is then repeated. The calculation is considered to be complete when for each estimate the difference between the calculated and estimated values is sufficiently smal!. An example of a sequential method package is PROCESS.

~.

0_'

2.3 The simultaneous modul ar method

In simultaneous modular method the flowsheet is seen as a set of equations, compri-sing:

- balance equations; - unit-operation equations.

The balance equations are linear, while the unit equations are usually non-linear. By substituting the non-linear equations with linear ones, a set of linear equations is obtained. This set of equations can be quite large, ranging from hundreds to thousands of equations.

As in the sequential method, a 'trial and error' procedure is used: non-linear equations are replaced by linear equations which are then solved. This is giving what is known as the 'linear balance'. On the basis of this linear balance the unit mode Is describing the various unit operations are calculated. Based on the results of the calculation of the units, the set of linear equations is updated. This procedure is repeated until the linear system no longer needs to be revised. The models are mostly identical to those of the sequential method. An example of a simultaneous modular method package is ASPEN + .

2.4 The equation-oriented method

In the equation-oriented method, too, non-linear equations are linearized. Here the linearization is performed according to Newton's method. The equations describing the units are in the form of functions which in the solution have a value of zero. Instead of linear coefficients, as in the simultaneous modular method, we have to deal with function values and partial derivatives. The mode Is are of the following form:

where,

f(X,.e)

=

Q

X are process variables; .e are unit parameters.

Af ter solving the set of equations, the models are onee again ea\culated (with the linear balanee as a basis) and all funetions va lues and derivatives are redetermined. Convergence is said 10 be reached when all functions approximate the zero value within an acceptable range.

These models are fundamentally different from the previous two types. They are not used for solving sets of equations. They are only used for deriving funetion values and partial derivatives. Solving the whole set of equations is left to the system. There are two methods for determining partial derivatives: numerical by the system, or analytical (or semi analytical) by the model. Examples of a equation-oriented method package are SPEED UP and TISFLO.

3 THE TOLUENE RECOVERY SYSTEM

The toluene recovery system consists of one flash vessel and th ree distillation towers straight forward. A flowsheet of the system and the given process data are shown in appendix A. Figure 3.1 shows a simplified flowsheet with the unit and stream names of the recovery system.

RECYCLE PURGE V-TW201 V-TW201 V-TWZ03 V-VE201 F-VE201 F-TW201 F-TW202 F-TW203 ---.t V E 2 0 1 TW201 TW202 TW203 VI V2 V3 L-VE201 L-TW201 L-TW202 L-TW203

Figure 3.1: A simplified flowsheet of the toluene recovery system.

The feed of the recovery system (F -VE201) is obtained from the catalytic reforming section. Af ter cooling by heat exchange, the gas/liquid mixture is separated in flash vessel VE20l. Table 3.1 shows the components of the feed and their component ID. The flasher operates at a pressure and temperature of 14.79 [bar] and 38.33 [0C] respectively. At these conditions almost all the hydrogen, methane, ethane and propane are separated from the heavier hydrocarbons. The vapor product (V-VE201) is recycled to the reforming section after recompression. Some of the gas is vented to prevent build-up of inerts in the reforming section. The hydrogen fraction is 85.30 [mole%] in the recycled and purged stream.

TABLE 3.1: The components of the feed of the toluene recovery system and their component ID. Name Comp 10 Name Comp 10 Name Comp 10 Hydrogen H2 i-Pentane ICS n-Heptane NCl Methane Cl n-Pentane NCS i-Octane ICB Ethane CZ i-Hexane IC6 Toluene TOL

Propane C3 n-Hexane NC6 n-Octane NCB

i -Butane IC4 Benzene BZ p-Xylene PX

n-Butane NC4 i-Heptane ICl Mesitylene MES

The liquid product (L-VE201) is separated in distillation tower TW201 into a LPG-fraction (V-TW201) and a Cs+-LPG-fraction (L-TW201). This column operates at lower pressure. The pressure of the liquid product of the flash vessel is reduced by expansion through a throttle valve. The i-pentane in the LPG-fraction is Ie ss than 1 [mole%] and the i-butane in the

Cs

+ -fraction is less than 2 [mole%].

Af

ter reduction of the pressure, also by expansion through a throttle valve, the C

_s

+ -fraction enters the next column (TW202). The gas/liquid feed is distilled into a toluene poor fraction (V-TW202) and heavier hydrocarbons (L-TW202). Toluene is the component with the lowest boiling point of the bottom product. The bottom product enters distillation tower TW203 af ter reduction of the pressure. High purity toluene (> 99.0 [mole%]) is isolated in th is column from the heavier hydrocarbons (p-xylene and mesitylene). The toluene contains a low percentage

«

0.4 [mole%]) of Cs-components. Table 3.2 shows some specifications and operating conditions of the recovery system.

TABLE 3.2: Some specifications and operating conditions of the recovery system. Stream Component Specification Unit Operating conditIons

V-VE201 Hydrogen 85.3 [mo le%] VE201 _PVE201 ; 14.79 [bar]

V-TW201 i-Pentane < I [mo le%] _{T VE201} ; 38.33 [·C]

L-TW201 i-Butane < _{2 [mole%] TW201 14.79} ~ P_TW201 ~ P_TW202 [bar]

V-TW203 Toluene > 99.0 [mo le%] TW202 P_TW201 ~ PTW202 ~ 1.0 [ba r]

Ca-components < 0.4 [mo le%] TW203 P

TW203 ; 1 0 [bar]

4 SIMULATION CALCULATIONS WITH CHEMCAD 11 4.1 Introduction

The flash vessel, the throttle valve and the distillation tower are simulated with the FLASH, V AL V and TOWR model respectively. The behaviour of the gas and liquid phases is described with the Soave-Redlich-Kwong (SRK) equation of state. The enthalpy is calculated also with the SRK equation of state. During all the simulation calculations no pressuredrop is assumed in the columns.

The flowrate and the composition of the feed of the flash vessel are not known.

It is not possible to calculate the feed composition with the given process data. The netto feed composition of the flash vessel (NF-VE201) is measured and shown in table 4.1. The netto feed stream is equal to the purge stream (PURGE) and the liquid product stream (L- VE201) together.

TABlE 4.1: The netto feed composition of the flash vessel VE201.

Comp [wt%] [mo le%] Comp [wt%] [mo le%] Comp [wt%] [mo le%]

IQ 10 10

H2 2.10 4.69E+l IC5 2.26 1.41 NC7 1. 87 8.47E-l

Cl 7.38E-l 2.07 NC5 1. 42 8.88E-l IC8 2.15 8.47E-l

C2 1. 78 2.66 IC6 6.94 3.63 TOL 2.44E+l 1.19E+l

C3 3.50 3.57 NC6 3.70 1.94 NC8 5.12E-l 2.03E-l

IC4 1. 50 1.16 BZ 1. 07E+l 6.15 PX 2.13E+l 9.01

NC4 1.72 1.33 IC7 5.50 2.47 MES 7.97 2.94

4.2 Calculation of the feed composition

The feed contains hydrogen from two different sourees, hydrogen produced in the reforming section (H2-PROD) and hydrogen from the recycle gas. The produced hydrogen is equal to the hydrogen amount in the netto feed. With the given it is possible to ca1culate the flowrate of the netto feed. <PF-REFORM is the flowrate of the feed of the catalytic reforming section. Table 4.2 shows the given flowrates of the recovery system and the H2-PROD/<PF_REFORM ratio.

TABlE 4.2: The given flowrates of the given recovery system.

Stream F lowrate [MSCFO] H2-PROO [MSCFO]

= 1.052

F-REFORM 37145 _<PF-REFORM [BBL/DJ

RECYCLE 162322

BBl/O = Barrels per Oay. MSCFO = Mille Standard Cubic Feet per Oay. The standard state of a gas is 60 ['FJ and 1 [atm].

To solve the overall mass balance and component mass balances the following assumptions are made:

1. The density of a gas can be calculated with the ideal gaslaw;

2. The vapor product of the flash vessel contains all the hydrogen, methane, ethane and propane;

3.

The liquid product of the flash vessel contains all the heavier hydrocarbons.

The hydrogen produced in the reforming section is ca1culated with equation [4.1] and

[4.2]. The flowrates of the netto feed and the products of the flash vessel are ca1culated with the equations [4.3], [4.4] and [4.5].

where, RhoGAS

_,

i H2-PROD ~NF-VE201 Rho_GASi MwGAs; H2-PROO p. J PF-REFORM Si P R*(273.15 + T) 1.052*~REFORM*CF1*RhoGAS _, H2 H2-PROD/MwGAS H2

,

=

is the gasdensity of component i [kg/m3]; is the molweight of component i [kg/kmol); is the hydrogen production [kg/hr];

is the flowrate of stream j [kmol/hr];

[ 4. 1]

[ 4.2 ]

[ 4.3 ]

[ 4.4 ]

[ 4.5 ]

is the flowrate of feed of the reforming section [BBL/D]; is the netto feed composition of the flash vessel [mole%]; is the pressure [bar];

T

R

CF1

is the temperature [0C].

is the gas constant

=

8.314 103 [J jK.kmol]

is conversion factor 1

=

1.1311. From [MSCFD] 10 [m3jhr] at 0 [0C] and 1 [bar].

With the second and third assumption it is possible to calculate the composition of the gas and liquid streams. The results are shown in table 4.3.

where, y. l x· _l Yi x· ₁ s· _l S. l [ 4.6] [ 4. 7 ]

is the composition of the streams V-VE201, RECYCLE and PURGE [mole%];

is the composition of the stream L-VE201 [mole%].

TABLE 4.3: The fractian af the campanents in the gas and liquid streams. Stream V-VE201, RECYCLE & PURGE

Camp 10 [ma le%] Camp 10 [ma le%] Camp 10 [ma le%]

H2 8.50E+l C2 4.82

Cl 3 75 C3 6.47

Stream L-VE201 & F-TW201

Camp 10 [ma le%] Camp 10 [ma le%] Camp 10 [ma le%]

IC4 2.60 NC6 4.33 TOL 2.65E+l

NC4 3.07 SZ 1.37E+l NC8 4.52E-l

lC5 3.15 lC7 5.52 PX 1.97E+l

NCS 1. 98 NCl 1. 88 MES 8.33

lC6 8.09 lC8 1.89

The overall molebalance is solved with the equations [4.8], [4.9] and [4.10] and the results are presented in table 4.4.

where, ~RECYCLE ~V-VE201 ~F-VE201 162322*CF1*---R*(273.15 + T) ~PURGE + ~RECYCLE

=

~L-VE201 + ~V-VE201

~j is the flowrate of stream i [kmoljhr].

TABLE 4.4: Calculated flowrates [kmol/hr]. Stream Flowrate H2-PROO 1941.9 NF-VE201 4140.6 PURGE 2285.6 RECYCLE 8084.8 V-VE201 10370.4 L-VE201 1855.0 F-VE201 12225.4 [ 4.8 J [4.9 J [4.10 ]

It is now possible to calculate the feed composition of the flash vessel with equation [4.11]. The results are shown in table 4.5.

where, z· _l

~F-VE201

Zj is the feed composition of the flash vessel [mole%

J.

Simulation calculations with CHEMCAD II

[4.11J

TASLE 4.5: The Assumed Feed composition of the flash vessel VE201. The total flowrate 1S

12225.4 [kmol/hr].

Comp [kmo l/hr] [mo le%] Camp [kmol/hr] [mo le%] Comp [kmol/hr] [mo le%]

10 JO JO

H2 8806.6 7.20E+l JC5 58.6 4.79E-l NC7 35.0 2.86E-l Cl 389.0 3.18 NC5 36.8 3.01E-l IC8 35.2 2.88E-l C2 499.6 4.09 IC6 150.5 l. 23 TOL 495.1 4.05 C3 670.6 5.49 NC6 80.5 6.58E-l NC8 8.4 6.90E-2 IC4 48.3 3.95E-l SZ 255.2 2.09 PX 374.0 3.06 NC4 56.9 4.65E-l IC7 102.6 8.39E-l MES 122.5 l. 00

4.3 Result of the flash calculation

The composition of the assumed feed is used

as

the

feed

composition

of

the flash

vessel.

The results of the P,T-flash calculation are shown

in

table 4.6.

TABLE 4.6: Results of the flash calculation at 14.79 [bar] and 38.33 [OC].

Enthalpy in [MJ/hr].

Stream F-VE201 V-VE201 L-VE201

Comp JO [kmol/hr] [mo le%] [kmo l/hr] [mo le%] [kmol/hr] [mo le%] H2 8806.6 7.20E+l 8795.4 8.43E+l 11 1 6.20E-l

Cl 389.0 3.18 385.4 3.69 3.6 2 03E-l

C2 499.6 4.09 474.3 4.55 25.3 1.41 C3 670.6 5.49 568.7 5.45 101.9 5.69 JC4 48.3 3.95E-l 34.9 3.30E-l 13.9 7.76E-l NC4 55.9 4.55E-l 35.5 3.41E-l 2l.4 1 19 JC5 58.5 4.79E-l 24.2 2.32E-l 34.4 ) 92 NCS 35.8 3.01E-) 12.8 1. 23E-) 24.0 1. 34 IC6 150.5 1.23 29.8 2.86E-l 120.7 6 73 NC6 80.5 6.58E-l 12.4 1.19E-l 68.1 3.80 SZ 255.2 2.09 25.8 2.47E-l 229.4 1.28E+l

IC7 102.6 8.39E-l 8.2 7.83E-2 94.4 5.27

NC7 35.0 2.86E-l 2.0 l.91E-2 33.0 1.84

JC8 35.2 2.88E-l 2.3 2.23E-2 32.9 1. 83

TOL 495.1 4.05 16.7 1.50E-l 478.4 2.57E+l NC8 8.4 6.87E-2 0.2 1. 67E-3 8.2 4.59E-l

PX 374.0 3.06 4.1 3.97E-2 359.9 2.05E+l

MES 122.5 1.00 0.4 3.90E-3 122. I 6.81

Total 12225.4 10432.7 1792.6

Enthalpy 212368.0 213864.0 -1495.5

4.4 Results of the distillation calculations

The configuration, operating conditions and product compositions followed from the

simulation calculations are shown in table 4.7, 4.8 and 4.9.

TABLE 4.7: The standard configuration and operating conditions of the toluene

recovery section. Unit VE201 TW201 TW202 TW203 Temperature [. C~ 38.33 Pressure [bar 14.79 8.0 1.0 1.0 Number of Stages 30 30 30 Feedstage Position 15 15 20

Condenser Type Partial Partial Partial

Temp."",. [OCl 27.3 80.4 110.4

Duty [MJ/hr -5042.8 -57268.6 -45660.5

Reflux Ratio [L/D 1. 45 2.5 3.5

Top Flow [kmol/hr 178.25 702.80 385.69

Reboi ler Temperature [,C] 185.7 124.9 140.3

Duty [MJ/hr] 56547.8 49439.4 59515.2

Boilup Ratio [V/B~ 1.17 1.5 3.0

Bottom Flow [kmol/hr 1614.37 911.56 525.88

Thermo-model SRK SRK SRK SRK

SRK: Soave-Redlich-Kwong.

TABLE 4.8: The composition of the flows [kmol/hr] in the distillatlon section. Pressure in [bar]. Temperature in ['C] and Enthalpy in [MJ/hr].

Unit/Str TW201 TW202 TW203 Comp 10 F V L F V L F V L H2 11. 1 11. 1 0.0 Cl 3.6 3.6 0.0 C2 25.3 25.3 0.0 C3 101.9 101. 9 0.0 IC4 13.9 13.9 0.0 NC4 21.4 21.4 0.0 IC5 34.4 1.0 33.4 33.4 33.4 0 0 NC5 24.0 0.1 23.9 23.9 23.9 0 0 IC6 120.7 0.0 120.7 120.7 120.7 0 0 NC6 68.1 0.0 68.1 68. I 68.1 0.0 BZ 229.4 0.0 229.4 229.4 229.4 0.0 IC7 94.4 0.0 94.4 94.4 94.4 0.0 NC7 33.0 0.0 33.0 33.0 32.6 0.4 0.4 0.4 0.0 IC8 32.9 0.0 32.9 32.9 32.7 0.2 0.2 0.2 00 TOL 478.4 0.0 478.4 478.4 67.5 410.8 410.8 380.4 30.4 NC8 8.2 0.0 8.2 8.2 0.0 8.2 8.2 4.7 3.5 PX 369.9 0.0 369.9 369.9 0.0 369.9 369.9 0.0 369.9 MES 122.1 0.0 122.1 122.1 0.0 122.1 122.1 0.0 122.1 Total 1792.6 178.2 1614.4 1614.4 702.8 911.6 911.6 385.7 525.9 Pressure 8.0 8.0 8.0 1.0 1.0 1.0 1.0 1.0 1.0 Temp 38.7 27.3 185.7 107.7 80 4 124.9 124.9 110.4 140.3 Enthalpy -1495.4 4485.0 45524.5 45523.6 28038.6 9657.1 9657.1 15542.7 7970 6

TABLE 4.9: The composition of the flows [mole%] in the distillation section. Pressure in

[bar]. Temperature in ["C]. Enthalpy in [MJ/hr] and Total in [kmol/hr].

Unit/Str TW201 TW202 TW203 Comp 10 F V L F V L F V L H2 0.62 6.24 0.0 Cl 0.20 2.05 0.0 C2 1. 41 14.2 0.0 C3 5.69 57.2 0.0 lC4 0.78 7.80 0.0 NC4 1.19 12.0 0.0 lC5 1.92 0.57 2.07 2.07 4.75 0.0 NC5 1.34 0.01 l. 49 l. 49 3.41 0.0 lC6 6.73 0.0 7.47 7.47 17.2 0.0 NC6 3.80 0.0 4.22 4.22 9.69 0.0 BZ 12.8 0.0 14.2 14.2 32.6 0.0 lC7 5.27 0.0 5.85 5.85 13.4 0.0 NC7 1.84 0.0 2.04 2.04 4.65 0.04 0.04 0.09 0.0 lC8 1. 83 0.0 2.04 2.04 4.65 0.02 0.02 0.04 0.0 TOL 26.7 0.0 29.6 29.6 9.61 45.1 45.1 98.6 5.79 NC8 0.46 0.0 0.51 0.51 0.0 0.90 0.90 1.23 0.66 PX 20.6 0.0 22.9 22.9 0.0 40.6 40.6 0.0 70.3 MES 6.81 0.0 7.56 7.56 0.0 13.4 13.4 0.0 23.2 Total 1792.6 178.2 1614.4 1614.4 702.8 911.6 911 .6 385 7 525.9 Pressure 8.0 8.0 8.0 1.0 1.0 1.0 1 .0 1 .0 1.0 Temp 38.7 27.3 185.7 107.7 80.4 124.9 124.9 110.4 140.3 Enthalpy -1495.4 4485.0 45524.5 45523.6 28038.6 9657.1 9657.1 15542.7 7970.6

5

SENSITIVITY ANALYSIS OF THE TOLUENE RECOVERY SYSTEM

The simulation calculations were done first with CHEMCAD Il. CHEMCAD II does not have the possibility to carry out sensitivity analyses or optimization calculations in a proper way. For this reason AS PEN

+

is used instead of CHEMCAD

IJ.

The simulation results given by CHEMCAD II we re first checked with those given by AS PEN +. The results of the ASPEN + simulation calculations are presented in appendix B. The conclusion is that all the results are within a percent equal. The mode Is used in ASPEN

+

for the flash vessel and the distillation tower are respectively FLASH2 and RADFRAC. The behaviour of the throttle valve is described with an adiabatic flash

vessel and simulated also with the FLASH2-model. Ouring the calculations no pressuredrop is assumed in the columns and the Soave-Redlich-Kwong equation of state is used to describe the behaviour of the vapor and the liquid phases. The standard configuration and operating conditions presented in appendix B (tabie B2) are used as base case during the sensitivity analysis.

5.1 Distillation tower 1W201

The liquid product of the flash vessel is separated in this column into a LPG-fraction (V-TW201) and a Cs+-fraction (L-TW201). The molefraction of i-pentane in the LPG-fraction and i-butane in the bottom product are specified. The i-pentane and i-butane fraction must be less than 1 and 2 [mole%] respectively. Also a condenser-temperature of about 30 [OC] is desired, because watercooling than becomes possible. A configuration has 10 be found for which the separation is as good as possible. The effects of the following variables are investigated:

Columnpressure; Number of Stages; Feedstage position.

5.1.1 The effect of the feedstage position

The effect of the feedstage position on the moleflow and molefraction of the pentanes in the top product is shown in table 5.1 and figure 5.1. The effect on the moleflow and molefraction of the butanes in the bottom product is shown in table 5.2 and figure 5.2. The stages are numbered from the top to the bottom of the 1Ower. The condenser is stage 1 and the reboiler is the last stage.

TABLE 5.1: i-Pentane and n-pentane in the distillate of column TW201.

Feed- ICS NCS ICS+NCS ICS NCS

stage [kmo l/hr] [kmo l/hr] [kmo l/hr] [ma le%] [ma le%]

3 6.38E-1 1.41E-l 7.79E-l 3.56E-1 7.86E-2

5 8.34E-1 6.76E-2 9.02E-l 4.65E-1 3.77E-2

10 8.87E-l 2.43E-2 9.11E-l 4.95E-1 1. 35E-2 15 9.11E-l 1.19E-2 9.23E-1 5.08E-1 6.62E-3

20 1. 05 7.67E-3 1.06 5.87E-1 4.27E-3

25 2.95 2.03E-2 2.97 1. 64 1.13E-2 28 1.17E+1 1. 97 1.37E+1 6.53 1. 10 14.13 13.9 12.9 ~ _11.0 c ,4

.,

18.8 c OIO û. 9.13 J c 8.8 w ( _i.ê ,4

"

C _6.8 011 û. J 5.13 " " 4.13 J I I _3.13 I 2.0 1.9 ---

---

0- . --

-

0

--

_ . - -

-. __ . - t r. -. -_.-.. - -'--0-- - --+- ----... - ---13.13 9 29 ---> Feed Stage [CS+NCS [ma le%] 434E-l 5.03E-l 5.08E-l 5.14E-1 5.91E-1 1. 65 7.63 o ;' / ,.{ , .. 0, I ; t , /

Figure 5.1: i-Pentane + n-pentane in distillate versus feedstage position of column Tw201. [.:kmol/hr, +:mole%l

TABLE 5.2: i-Butane and n-butane in the bottom product of column TW201.

Feed- IC4 NC4 IC4+NC4 [C4 NC4 [C4+NC4

stage [kmo l/hr] [kmo l/hr] [kmo l/hr] [mo le%] [mo le%] [mo le%]

3 0.00 8.98E-4 8 98E-4 0.00 5.56E-5 5.56E-5

5 0.00 4.20E-4 4.20E-4 0.00 2.60E-5 2.60E-5

10 1. 27E-5 1.21E-3 1.22E-3 0.00 7.50E-5 7.50E-5

15 3.52E-4 1.20E-2 1. 24E-2 2.18E-5 7.42E-4 7.64E-4

20 1.08E-2 1. 39E-1 1. 50E-1 6.67E-4 8.61E-3 9.27E-3

25 3.40E-1 1. 58 1. 92 2.10E-2 9.79E-2 1.19E-1

28 2.33 5.57 7.90 1. 44E-1 3.45E-l 4.90E-l

38

~ c

'"

.,

J al I c ~ c

'"

.,

:l al I

..

1\ I I 8.13 7.0 6.0 5.9 4.8 3.0 ;' i / 2.8 I I 1.0 0.8

/~

8 18 28 ---) Feed Stage

Figure 5.2: i-Butane + n-butane in bottom product versus feedstage of column T~201.

[.:kmol/hr, +:mole%l

30

Table 5.1 and figure 5.1 show that the best results were obtained when the feed enters the column on stage 3. Stage 3 is the second stage of the column and to get more flexibility an optimum feedstage position of 5 is chosen. This gives no problem for the specifications, because the molefraction of i-pentane and i-butane in the top and bottom product respectively are also at this feedstage position smaller than the specified values. From th is moment the feed of the first column enters the column on stage 5 instead of stage 15.

5.1.2 The effect of the pressure

The effect of the columnpressure on the moleflow and molefraction of i-pentane in the top product is shown in table 5.3 and figure 5.3. It can be concluded from this table and figure that the moleflow and molefraction of i-pentane in the distillate are independent of the columnpressure.

~

TABLE 5.3: i-Pentane in distillate of column T\.I201.

P les les P les les

[bar] [kmol/hr] [mo le%] [bar] [kmol/hr] [mo le%]

5.0 1. Ol 5.64E-1 8.0 1. Ol 5.66E-1 6.0 1. Ol 5.64E-1 9.0 1. Ol 5.67E-1 7.0 1. 01 5.65E-1 10.0 1. 02 5.70E-1 1.20 1.18 1.00 fl fl fl EI fl 0.98 Ib O.Be c fd

.,

_0.78 c 41 a. I _0.60 A _0.50 I I I 0.40 I 0.30 0.20 0.10 0.00 4.8 6.0 8.8 ---> Pressure (bar)

Figure 5.3: i-Pentane in distillate versus pressure [bar) in column T\.I201.

(.:kmol/hr, +:mole%)

I;l

----+--10.0

The effect on the condenser and reboiler temperature is shown in table 5.4 and figure 5.4. The results show the higher the pressure the higher the condenser and reboiler temperature. Watercooling becomes possible when the column operates at a pressure

beyond 9.0 [bar].

TABLE 5.4: Condenser and reboiler temperature of column TW201.

P Condenser Rebo i ler P Condenser Rebo i ler

[bar] temp [·e] temp [·e] [bar] temp [·e] temp [·e]

8.0 26.90 185.58 10.0 34.80 199.20

8.5 29.01 189.20 10.5 36.58 202.28

9.0 31.02 192.67 11.0 38.29 205.27

9.5 32.95 195.99

---' 21B.B , - - - -- - - -- - - -- -- -- - -- -- ---, 2BB.0 19B.B 18B.B 6B.B ~ ~ 50.0

"

~ ..., ~ " 4B.B ~ Cl. E ~ >-3B.B A I I I I _20.0 10.B O.B 7.0 9.0 11.0 ---) Pressure [b.rl

Figure 5.4: Reboi ler [_) and condenser [+) temperature lOC) versus pressure [bar) in column Tw201.

5.1.3 The effect of the number of stages

The effect of the number of stages on the molefraction of the pentanes in the top product is shown in table 5.5 and figure 5.5. The effect on the moleflow of the pentanes is shown in table 5.6 and figure 5.6. The effect on the butanes in the bottom product is shown in table 5.7 and 5.8.

TABLE 5.5: i-Pentane and n-pentane [mole%] in the distillate of column TI/20l.

P Number of Stages = 20 Number of Stages = 25 Number of Stages = 30 [bar] IC5 NC5 IC5+NC5 IC5 NC5 IC5+NC5 IC5 NC5 IC5+NC5

8.0 4.84E-l 4.73E-2 5.31E-l 4.65E-l 4.03E-2 5.05E-l 4 65E-l 3.77E-2 503E-l

8.5 4.87E-l 4.94E-2 5.36E-l 4.64E-l 4.16E-2 5.05E-l 4 63E-l 3.88E-2 5.02E-[

9.0 4.91E-l 5.15E-2 5.42E-l 4.63E-l 4.29E-2 5.05E-l 4 62E-l 3.99E-2 5.02E-l

9.5 4.95E-l 5.38E-2 5.49E-l 4.62E-l 4.43E-2 5.06E-l 4 60E-l 4.11E-2 5.01E-[

10.0 5.01E-l 5.61E-2 5.57E-l 4.61E-l 4.57E-2 5.06E-l 4. 59E-l 4.23E-2 5.00E-l

10.5 5.07E-l 5.87E-2 5.65E-[ 4.60E-[ 4.72E-2 5 07E-l 4 57E-[ 4.3 5E - 2 5 ODE -1

11.0 5.14E-l 6.13E-2 5.75E-l 4.59E-l 4.87E-2 5.08E-l 4 55E -1 4.47E-2 5 OOE-l

0.70 0.60 ~ c ~ B B U - cr-u 0 -e

"

c 0.50

*

*

*

~ a. I c + 0.40 ~ c ~ " c _0.30 ~ Q. I .~ 1\ _0.20 I I I I 0.10 8.88 7.8

',.8

11.8 ---) Pressure [blr]

Figure 5.5: i-Pentane + n-pentane [mole%] in doistillate versus pressure [bar] of column T\.I201. Number of stages is .:20, +:25, *:30.

TABLE 5.6: i-Pentane and n-pentane [kmol/hr] in the disti llate of column T\.I201.

P Number of Stages = 20 Number of Stages = 25 Nurnber of Stages = 30

[bar] IC5 NC5 IC5+NC5 IC5 NC5 IC5+NC5 IC5 NC5 I C5+NC5

8.0 8.68E-l 8.49E-2 9.53E-l 8.33E-l 7 .22E-2 905E-l 8 34E-l 6.76E-2 902E-l

8.5 8.74E-l 8.85E-2 9.63E-l 8.31E-l 7 .46E-2 9.06E-l 8 31E-l 6.96E-2 9. 01E-l 9.0 8.80E-l 9.23E-2 9.72E-l 8.30E-l 7 .70E-2 9.07E-l 8. 28E-l 7.16E-2 9 OOE-l 9.5 8.88E-l 9.64E-2 9.84E-l 8.28E-l 7.95E-2 9.08E-l 8 25E-l 7.37E-2 8.99E-l 10.0 8.98E-1 1.01E-1 9.99E-1 8.26E-1 8.20E-2 9.08E-l 8.22E-1 7.58E-2 8.98E-l 10.5 9.09E-l 1.05E-1 1. 01 8.25E-1 847E-2 9.10E-1 8. 19E-1 7.80E-2 8.97E-l 11. 0 9.21E-1 1. lOE-1 1. 03 8.24E-1 8.74E-2 9. 11E-1 8. 16E-1 8.02E-2 8.96E-1

---

'

'-' 1.18,---, 1.00 011 c: rj

.,

C 111 0.99 't==

*

~ -

_*

I I a. )( )( I C 011 c _0.80 ol

.,

C 011 Cl. I ... 0.78 1\ I I I I 0.60 O.~+_---_r---~---,_---_,---~ 7.0 9.0 11.0 ---) Pressure [bar]

Figure 5.6: i-Pentane + n-pentane [kmol/hr) in distillate versus pressure [bar) in column T\oI201.

Number of stages is .:20, +:25, *:30.

TABLE 5.7: i-Butane and n-butane [mole%] in the bottom product of column TW201.

P Number of Stages = 20 Number of Stages = 25 Number of Stages = 30

[bar] IC4 NC4 IC4+NC4 IC4 NC4 IC4+NC4 IC4 NC4 IC4+NC4

8.0 6.29E-5 3.18E-3 3.25E-3 0.00 2.94E-4 2.94E-4 0.00 2.60E-5 2.60E-5

8.5 8.08E-5 3.81E-3 3.89E-3 0.00 3.70E-4 3.70E-4 0.00 3.44E-5 3 44E-5

9.0 1.03E-4 4.52E-3 4.62E-3 0.00 4.62E-4 4.62E-4 000 4.49E-5 4.49E-5

9.5 1.29E-4 5.32E-3 5.45E-3 0.00 5.74E-4 5.74E-4 000 5.81E-5 5.81E-5

10.0 1.61E-4 6.22E-3 6.38E-3 0.00 7.07E-4 7.07E-4 0.00 7.46E-5 7.46E-5

10.5 1. 98E-4 7.23E-3 7.43E-3 0.00 8.66E-4 8.66E-4 0.00 9.53E-5 9.53E-5

11.0 2.43E-4 8.34E-3 8.59E-3 1.29E-5 1. 05E-3 1.07E-3 0.00 1.21E-4 1.21 E-4

TABLE 5.8: i-Butane and n-butane [kmol/hr] in the bottom product of column TW201.

P Number of Stages = 20 Number of Stages = 25 Number of Stages = 30

[bar] IC4 NC4 IC4+NC4 IC4 NC4 IC4+NC4 IC4 NC4 IC4+NC4

8.0 1.02E-3 5.14E-2 5.24E-2 3 .55E-5 4.74E-3 4.78E-3 0.00 4 20E-4 4 .20E-4

8.5 1.30E-3 6.15E-2 6.28E-2 4.90E-5 5.97E-3 6.02E-3 o 00 5 55E-4 5. 55E-4

9.0 1.66E-3 7.30E-2 7.47E-2 6.68E-5 7.47E-3 7.54E-3 0.00 7 .24E-4 7 .24E-4

9.5 2.08E-3 8.59E-2 8.80E-2 9.00E-5 9.26E-3 9.35E-3 0.00 9 .38E-4 9 38E-4

10.0 2.59E-3 1.00E-l 1. 03E-l 1.20E-4 1.14E-2 1.15E-2 0.00 1 .20E-3 1.20E-3

10.5 3.20E-3 1.17E-l 1. 20E-l 1. 59E-4 1.40E-2 1.42E-2 0.00 1.54E-3 1 54E-3

11.0 3.92E-3 1.35E-l 1. 39E-l 2.08E-4 1.70E-2 1.72E-2 1.02E-5 1.95E-3 1.96E-3

-'

From the results it can be concluded th at decreasing the number of stages results in higher molefraction and moleflow of both pentanes and butanes. Nevertheless the specifications are still met.

Table

5.9

shows the effect on the reboiler and condenser temperature. lt can be concluded from th is table that the reboiler and condenser temperature are independent of the number of stages.

TABLE 5.9: Condenser and reboiler temperature of column TW201.

Number of Stages = 20 Number of Stages = 25 Number of Stages = 30 P

[bar] Condenser Rebo i ler Condens er Rebo i ler Condenser Reboi ler temp ['C] temp ['C] temp ['C] temp ['C] temp ['C] temp ['C] 8.0 26.96 185.58 26.91 185.58 26.90 185.58 8.5 29.07 189.19 29.02 189.20 29.01 189.20 9.0 31.09 192.66 31.03 192.67 31. 02 192.67 9.5 33.03 195.99 32.96 195.99 32.95 195.99 10.0 34.90 199.19 34.81 199.20 34.80 199.20 10.5 36.69 202.28 36.59 202.28 36.58 202.28 11.0 38.42 205.26 38.31 205.27 38.29 205.27 5.2 Distillation tower 1W202

The second distillation tower is used to separate the bottom product of column TW201 into a toluene poor fraction and heavier hydrocarbons. The toluene poor fraction contains all the components having lower boiling points as compared 10 1Oluene. A configuration has to be found for which this separation is as good as possible, because in the third column toluene has to be isolated as a 99.0 [mole%] pure product at the top.

5.2.1 The effect of the reflux ratio

The effect of the reflux ratio on the toluene loss in the columns TW202 and TW203 and on the toluene purity in the top product of column TW203, is analyzed. The results are shown in table

5.10

and the figures

5.7

and

5

.8

.

A

I I I I

TABLE 5.10: Toluene in the bottom product of column TW202 and the distillate of the

columns TW202 and TW203.

Reflux Toluene [kmo l/hr] Ta luene [ma le%]

Ratio

Top TW202 Bot TW203 Top TW203 Top TW202 Bot TW203 Top TW203

2.0 71.58 30.14 376.03 10.18 5.73 97.49 2.1 69.77 30.17 377 . 82 9.93 5.74 97.95 2.2 68.62 30.19 378.94 9.76 5.74 ···98.24 2.3 67.91 30.20 379.64 9.66 5.74 98.42 2.4 67.46 30.21 380.09 9.60 5.74 98.54 2.5 67.16 30.22 380.38 9.56 5.75 98.62 2.6 66.96 30.22 380.57 9.53 5.75 98.67 2.7 66.82 30.22 380.71 9.51 5.75 98.70 2.8 66.73 30.22 380.81 9.49 5.75 98.73 2.9 66.65 30.22 380.88 9.48 5.75 98.75 3.0 66.60 30.23 380.93 9.48 5.75 98.76 ::::: r_--<88---<>"" _ --t:".---"....__- __eB- ---O- - - · 0 -- ---. . --.0 - -. 378.8r 80.0 70.0 60.0 58.8 48.8

30.a+---~-- -- ..y.- -_.- -.---<)-- ---~--~-_.--- _ ... ----.,e. --~~:r----_.:r_- -20.0 10.0

I

o.o+----.---.---,---.---,---.---,---.---,---~ 2 2.2 2.4 2.'; 2.8 ---) R~f'lu)( R .. tlO

Figure 5.7: Toluene [kmol/hrl versus Reflux Ratio of column TWZOZ.

+:in distillate of TWZOZ, ~:in bottom product of TWZ03, .:in distillate of TWZ03.

',,--J 1 0 0 . 0 . , . . . - - - , 98.0 _ _ -'0- - '0- - ' - !l [] 96.0 14.9 ~ ~ 12.0 J 10.0 - -- +--.-. '.', ... _ ... _{+-- --+--- -- - - - + -- . ,+" ' - - - ' _ . --t._.} À : 8.9 I I + ._. 6.8 +----*----"----4---~~--0____---_--.-9 __ --~' ._ .... _-. 9 4.0 2.0 O.O~--.----,--~----._--._--,_--_r--~---._-~ 2 2.2 2.4 2.6 2.8 ---> Refl ux Rat i 0

Figure 5.8: Toluene [mole%] versus Reflux Ratio of column T~202.

,': in disti llate of T~202, +: in bottom product of T~203 .: in disti llate of T~203.

As can be seen, the results for higher reflux ratias are better, although the improvements are only up to a few percents. Increasing the reflux ratio from 2.0 to 2.4 gives the largest irnprovement.

5.2.2 The effect of the bottom flow

The effect of the bottom

flow

was

analyzed by varying

this flow between

800

and

1000

[kmoljhr). The results

are

shown in table

5.11 and

figures 5.9

and 5.10.

,

,

,

,

TABLE 5.11: Toluene ln the bottom product of column TW202 and the distillate of the columns TW202 and TW203.

Bottom Toluene [kmo l/hr] Toluene [mole%]

Flow

[kmo l/hr] Top TW202 Bot TW203 Top TW203 Top TW202 Bot TW203 Top TW203

800.0 177.79 30.02 269.94 21.83 5.71 98.47 820.0 157.87 30.04 289.84 19.87 5.71 98.54 840.0 137.96 30.07 309.73 17.81 5.72 98.60 B60.0 118.08 30.10 329.58 15.65 5.72 98.64 880.0 98.23 30.15 349.38 13.37 5.73 98.66 900.0 78.50 30.19 369.07 10.98 5.74 98.65 920.0 59.00 30.24 388.52 8.50 5.75 98.57 940.0 40.24 30.27 407.24 5.97 5.76 98.33 960.0 24.34 30.27 423.15 3.72 5.76 97.47 980.0 14.95 30.17 432.64 2.36 5 74 95.26 1000.0 9.75 30.02 437.99 1. 59 5 71 92.37 500.0 450.0 400.0 o 350.0 3G0.a 250.0 200.0 158.8 100.6 56.6 t -- --<>-- - - . > - - - -<>-.--~ ... O.O+----r---,---,----r---,---,----.---,---,---~ 0.8 0.84 O.BB 0.92 ( Thous~nds) ---) Bottom FJo~ (kM01/hrJ 0.96

Figure 5.9: Toluene [kmol/hrl versus Bottom Flow [kmol/hrl of column TW202 .

• :in distillate of TW202, +:in bottom product of TW203, ~:in distillate of TW203.

108.0 .---~ A _, i 99.0 ---·~u~--~e--~g~--~g---~~--~---~_ %.8 94.8 92.8 20.0 18.8 16.9 14.9 12.9 18.0 8.9 6.9 4.8 2.0 8.0 9.8 0.B4 O.BB 9.92 (Thousa,nds)

---) Bottom flow [kmol/hr)

0.96

Figure 5.10: Toluene [mole%l versus Bottom Flow [kmol/hrl of column TW202. +:in distillate of TW202, o:in bottom product of TW203, .:in distillate Tw203.

It follows from these results th at with increasing bottom flow the toluene loss decreases enormously, but so does the toluene purity in the top of column TW203 beyond a bottom flow of 940 [kmol/hr].

5.2.3 The effect of the pressure

The effect of the pressure was analyzed by varying it between l.0 and 3.0 [bar]. This was done for three different reflux ratios. The results are shown in table 5.12 and figures

5.11 and 5.12.

TABLE 5.12: Toluene in the distillate versus the pressure in column TW202. P Toluene [kmo l/hr] in top TW202 Toluene [ma le%] in top TW202 [bar] RR = 2.0 RR = 2 5 RR = 3.0 RR = 2.0 RR = 2.5 RR = 3.0 1.0 71.58 67.16 66.60 10.18 9.56 9.48 1.5 73.53 67.72 66.79 10.46 9.64 9.50 2.0 75.49 68.42 67.06 10.74 9.74 9.54 2.5 77 .42 69.29 67.41 11. Ol 9.86 9.59 3.0 79.34 70.29 67.84 11.29 10.00 9.65

A I I I I ~~---~

-

-78 76 74 ~ _______ 72~ 78

----

_...

--

--64 62 w+---r---.---.---r---.---.---r---.---~----~ 1.13 1.4 1.8 2.2 ---) Pressure [bar]

Figure 5.11: Toluene [kmol/hr] in distillate versus pressure [bar] in column T\J2ü2.

Reflux Ratio of tower T\J2ü2 is .:2.13, +:2.5, 0:3.13.

3.0 15.---, 14 13 12 ~ ~ ₁₁ ~ ___ - - ----0 ----' . J ._ . . - -. . - ; : r - -~ 0

.1---

- - ---

..

-

0 r- ₁₈ ._. ---+--- ---..--/\ o I 9 I I I 8 7 6 5 1.8 1.4 1.8 2.2 2.6 ---) Pressure (bar]

Figure 5.12: Toluene (mole%) in disti llate versus pressure (bar) in column T\J202.

Reflux Ratio of tower T\J202 is .:2.0, +:2.5, ó:3.0.

--~

3.0

The higher the pressure, the higher the toluene loss in column TW202, at constant

reflux ratio. To obtain the same separation results one needs higher reflux ralios 10 compensate for higher pressures.

5.2.4 The effect of the feedstage position

Finally the effect of the feedstage position was analyzed. The results are shown in table

5.13 and figure 5.13 and 5.14.

... ' i (Jl ~ ...J "IJ c , ~ ~ c w J 0 f-1\ I I I I

TABLE 5.13: Toluene in the distillate and the bottom product of column TW202.

Feed- Toluene in top TW202 Light in bottom TW202 stage

[kmol/hrJ [mo le%} [kmol/hrJ [mo le%} 3 93.46 13.29 44.00 4.83 5 74.28 10.56 10.00 1.10 10 66.71 9.49 0.00 0.00 15 67 16 9.56 0.00 0.00 20 69.05 9.83 2.00 0.22 25 77.49 11 . 02 11.00 1. 21 28 96.37 13.71 29.00 3.18 188.8 98.8 80.0 70.0 u 68.8 50.0 40.0 \ 30.0

\

20.0 _'.. 18.8 0.0 0 18 ---) Feed Stage 20

Figure 5.13: Toluene [kmol/hrl in disti Ilate and "1 ight" components [kmol/hrl in bottam product versus feedstage position of column TW202.

30

,-,' 28.8 18.8 c 0 .., 16.0 cl M ~ 14.0 u. .., r. IJl _12.8 .~ ....J '0 10.8 c M lil c _8.8 lil J 0 6.8 f-1\ I _4.0 I I I 2.0 8.8 ----~---~-- ---8 18 28 ---) Feed Stage ~ ----~ o o

Figure 5.14: Toluene [mole%l in disti llate lil ight" components [mole%l in bottom product versus

feedstage of column T~202.

38

Between feedstage positions 5 and 15 a rather shallow minimum is obtained for both the toluene in the top product and the "light" components (C₆-C_{8 )} in the bottom product of column TW202.

5.3 Distillation tower 1W203

The bottom product of column TW202 is fed to distillation tower TW203. From this final tower a top product containing 99.0 [mole%] toluene has 10 be obtained and the Cs-content has 10 be less than 0.4 [mole%]. The bottom product of column TW203 can be mixed up with the top product of column TW202 to yield a gasoline fraction. To come to an optimal configuration, the effects of the reflux ratio, the bottomflow, the number of stages and the feedstage position on the toluene loss and purity is analyzed.

5.3.1 The effect of the reflux ratio

The effects of the re flux ratio are shown in table 5.14 and the figures 5.15 and 5.16.

41 c: ~ J .... 0 f-A I I I I

TABLE 5.14: Toluene in the distillate and the bottom product

of column TW203.

Reflux Toluene in top TW203 Toluene in bot TW203

Ratio

[kmo l/hr] [ma le%] [kmo l/hr] [ma le%]

3.0 379.97 98.51 30.62 5.82 3.1 380.06 98.53 30.54 5.81 3.2 380.14 98.55 30.45 5.79 3.3 380.22 98.57 30.37 5.78 3.4 380.30 98.60 30.29 5.76 3.5 380.38 98.62 30.22 5.75 3.6 380.45 98.63 30.14 5.73 3.7 380.53 98.65 30.06 5.71 3.8 380.60 98.67 29.99 5.70 3.9 380.67 98.67 29.92 5.68 4.0 380.74 98.71 29.85 5.68 400.9.---, ib---o---n-u---!l---t:r---o---ttO---10~---_tct_-·-·---·c--- -300.8 280.8 100.0 + - - -... - ----10-- - -+--- - - ----+-- ----1---- +---+---1...-- - ----+--.- -O.O~---r----~---r_----~Ir---~I---r---r----~---.---~ 3.8 3.2 3.4 3.6 3.8 4.0 ---) Ref1ux Ratio

Figure 5.15: Toluene [kmol/hrl versus Reflux Ratio of column TW203 .

• :in distillate of TW203, +:in bottom product of TW203.

.... _ .. 1 8 8 . 0 r - - - , 98.8 96.8 14.0 '11 c 12.9 Go ::J ~ 0 >- _18.8 A I 8.9 I I I 6.8 t - - -- 01-- ----+---..j..---- -_ - - ... ---_---+---~---.---,,_ --.- _ 4.0 2.0 0,0 3.0 3.2 3.4 3.6 ---) Reflux R4tio

Figure 5.16: Toluene [mole%l versus Reflux Ratio of column T~203 . • :in distillate of T~203. +:in bottom product of T~203.

3.8 4.0

From the reflux ratio analysis follows, as could be expected, that the higher the reflux ratio, the higher the toluene purity in the top product and the lower the toluene loss. However, the effect within the range analyzed is smal!.

5.3.2 The effect of the bottom flow

The effect of the bottom flow on the toluene purity and loss are shown In table 5.15

and figure 5.17 and 5.18.

A ,

I

I

I

TABLE 5.15: Toluene in the distillate and the bottom product of column TW203.

Bottom Toluene in top TW203 Toluene in bot TW203

Flow

[kmo l/hr] [kmo l/hr] [ma le%] [kmo l/hr] [ma le%]

450.0 410.22 88.87 0.37 8.32[-2 460.0 410.15 90.82 0.44 9.60[-2 470.0 410.07 92.86 0.53 1.12[-1 480.0 409.96 94.99 0.63 1.31E-1 490.0 409.82 97.21 0.77 1.57E-1 500.0 404.07 98.17 6.52 1. 30 510.0 395.24 98.42 15.36 3.01 520.0 385.95 98.56 24.64 4.74 530.0 376.44 98.65 34.16 6.44 540.0 366.79 98.71 43.80 8.11 550.0 357.07 98.75 53.52 9.73 428.8 418.13 488.0 0 390.0 380.8 378.13 360.0 350.0 60.() 50.9 40.0 30.a 28.() 18.() 0 [] [] o.o+---~----~---~----_r~--_,---,_----_,----_,r_----,_----~ 459.0 470.8 490.8 510.0

- - - j BottoM flow [kmol/hrl

Figure 5.17: Toluene [kmol/hrl versus Bottom Flow of column TW203 .

• :in distillate of TW203, +:in bottom product of TW203.

Sensitivity analysis of the toluene recovery system

530.0 550.0

109.9~---, .. 0- -

_--

-D--B B B ---0--- - - --_ .. 95.9 99.9 85.9 20.0 15.0 10.0 5.0 O.O~--~--~--~--~~=--.---.---~---.---r--~ 450.0 470.0 490.8 510.0

---} Bottom Flow [kmol/hrJ

Figure 5.18: Toluene (mole%l versus Bottom Flow of column T~203 .

• :in distillate of T~203, +:in bottom product of T~203.

538.0 550.0

From these results it can be concluded that increasing the bottom tlow from 450 to 550 [kmoljhr] sure increases the purity of toluene (from 88.87 to 98.75 [mole%]), but on the other hand also increases the toluene 10ss dramatically.

5.3.3 The effect of the number of stages and the feedstage position

The effects of the number of stages and the feedstage position on the toluene purity and loss are shown in table 5.16. The effects on the n-octane content in the top product are shown in table 5.17.

TABLE 5.16: Toluene in the distillate of column TW203.

Feed Number of Stages = 20 Number of Stages = 30 Number of Stages = 40 Number of Stages = 50 Stage

[kmo l/hr] [ma 1e%] [kmo l/hr] [ma 1e%] [kmo l/hr] [ma 1e%] [kmo1/hr] [mo 1e%]

5 367.77 95.35 368.16 95.45 - - - -10 379.04 98.27 380.18 98.56 380.37 98.61 380.39 98.62 15 378.80 98.21 380.90 98.75 381.38 98.87 381.45 98.89 20 - - 380.38 98.62 381 .44 98 89 381.62 98.94 25 - - 379.18 98.30 381.20 98.83 381.66 98.95 30 - - - _{- 380.55 98.66 381.58 98.93} 35 - - - _{- 381.30 98.85} 40 - - - _{380.64 98.68}

~)

TABLE 5.17: n-Octane in the distillate of column TW203.

Feed Number of Stages = 20 Number of Stages = 30 Number of Stages = 40 Number of Stages = 50

Stage

[kmo l/hr] [ma le%] [kmo l/hr] [ma le%] [kmo l/hr] [ma le%] [kmo l/hr] [ma le%]

5 4.27 1.11 3.93 1. 02 - - -

-10 4.88 1. 27 3.83 9.92E-1 3.64 9.44E-l 3.61 9.37E-l 15 5.92 l. 53 3.94 1.02 3.47 8.99E-l 3.40 8.81E-l 20 - - 4.52 l. 17 3.46 8.97E-l 3.28 8.50E-l 25 - - 5.72 l. 48 3.70 9.59E-l 3.24 8.41E-l 30 - - - - 4.35 1. 13 3.32 8.60E-l 35 - - - - _{- - 3.60 9.33E-1} 40 - - - 4.26 1. 10

From these results it can be concluded that increasing the number of stages results in a higher purity of 1Oluene; nevertheless, the economics of increasing the number of stages overshadows the small positive effect on the purity when doing so.

From the results of varying the feedstage position it can be seen th at there IS

a shallow optimum when the feed enters the column just in the middle.

5.4 Conclusions from the sensitivity analyses

The improved and operating conditions followed from the sensitivity analyses are shown in table 5.18. The simulation results with respect 10 the specifications are shown in table 5.19.

TABLE 5.18: The improved configuration and operating conditions of the toluene recovery section.

Unit VE201 TW201 TW202 TW203

Temperature [. Cj 38.33

Pressure [bar 14.79 9.0 1.0 1.0

Number of Stages 20 30 30

Feedstage Position 5 10 20

Condenser Type Partial Total Total

Temperature [·C] 31 . 1 69.9 109.7 Duty [MJ/hr] -1 .380E+6 -2 630E+7 -1 .543E+7 Reflux Ratio _[L/Dj 1. 45 3.0 4.0 Top Flow [kmo l/hr 179.34 734.37 330.0 Reboiler Temperature [. C] 192.7 125.2 138.5 Duty [MJ/hr] 1 .654E+7 1 .705E+7 1 .563E+7 Boilup Ratio _{[V /B~} 1. 26 1. 92 2.71 Bottom Flow [kmol/hr 1614.37 880.0 550.0 Thermo-model SRK SRK SRK SRK

SRK: Soave-Redlich-Kwong.

TABLE 5.19: Specifications calculated with the improved configuration.

Component Stream Specification Calculated [ma le%} [ma le%] [kmo l/hr]

Hydrogen Recycle 85.3 84.31 6940.63

i-Pentane Distillate of T\-I201 1.0 4.91E-l 1.05 n-Pentane Distillate of T\-I201 -- - - - - _{5.15E-2 1.34E-l}

i-Butane Bottom Product of T\-I201 2.0 1. 03E-4 6.82E-3

n-Butane Bottom Product of T\-I201 - - - - _{4.52E-3 2.57E-l}

Toluene Distillate of T\-I203 99.0 99.06 326.92 C8-comp Distillate of T\-I203 0.4 9.14E-l 3.02

Toluene Total 10ss minimum 154.74

Purge 3.90

Distillate of T\-I201 4.70E-5

Distillate of T\-I202 97.84

Bottom Product of T\-I203 53.00

A final optimization of the system is still necessary, because the toluene toss 1S unacceptable and the hydrogen specification is still not met yet.

6 OPTIMIZATION OF THE TOLUENE RECOVERY SYSTEM

Optimization of a system is defined as the calculation of the steady-state mass and heat balance with the addition of an objective function and imposed constraints. The user must define an objective function to maximizejminimize and the variables that can be varied within a user defined range. The objective function may be economie, e.g.

minimum toluene 10ss or purely technical, e.g. minimum temperature in a condenser. The user can also define constraints of one or more variables, e.g. toluene purity of 99.0 [mole%] in the distillate. The variables that are defined as constraints are called constraint variables.

6.1 Calculation of the flash feed composition with ASPEN

+

It can be concluded from the results of the flash calculation (see table 4.6 and B2) that the last two assumptions made in chapter 4 were not correct. The optimization option in ASPEN + makes it possible to calculate the feed composition in such a way that the measured and calculated netto feed composition are almost the same. The simplified

flowsheet shown in figure

6.1

is used to ca1culate the feed composition.

H2 C1 C2 NCB PX MES , , , , I

..

RECYCLE PURGE Y-YE201 ' -F-YE201 NF-YE201 YE201 - r -L-YE201

Figure 6.1: A simplified flowsheet, used to calculate the feed composition.

With the given process data and the assumption that all mole-gasdensities are equal it is possible to solve the overall hydrogen molebalance. This makes it also possible to calculate the moleflow for each component in the netto feed. The results are shown in table

6.1

and

6.2

respectively.

TABLE 6.1: The total and hydrogen moleflow [kmol/hr].

Stream Total Hydrogen

RECYCLE 8084.8 6896.2 NF-YE201 4147.1 1946.3

F-YE201 12231. 9 8842.5

TABLE 6.2: The measured component mo1ef1ow [kmo1/hr] of the netto feed.

CaMP ID Moleflow CaMP 10 Moleflow CaMP ID Moleflow

H2 1946.3 IC5 58.7 NC7 35.1 Cl 85.9 NC5 36.8 IC8 35.1 C2 110.4 IC6 150.4 TOL 494.9 C3 148.2 NC6 80.5 NC8 8.4 IC4 48.2 BZ 255.0 PX 373.9 NC4 55.0 IC7 102.6 MES 121.8

The molefraction of hydrogen in the vapor product of the flash vessel is defined as a constraint variabie with a value of 0.853. The moleflow of the recycle stream and the hydrogen in the feed are fixed values with a value of 8084.8 and 8842.5 [kmoljhr] respectively. The variables that can be varied to calculate the optimum are shown in tab Ie 6.3. The objective function chosen to minimize is:

MES

Objective Function _{L: (Prneasured i Pcalculated,i)} 2

where,

P measured,i

q, calculated,i

i=Cl '

is the measured flowrate of component i in stream NF- VE201 [kmoljhr];

is the calculated flowrate of component i in stream NF-VE201 [kmoljhr];

TABLE 6.3: The varlables that can be varied to ca1cu1ate the opt1mum.

COMP 10 Starting 1i it COMP 10 Starting 11 i t

va1ue lower upper va1ue lower upper

NC6 92.7 10.0 200.0 Cl 387.9 200.0 550.0 BZ 281.5 150.0 400.0 C2 432.4 200.0 550.0 IC7 111. 3 50.0 200.0 C3 438.2 200.0 550.0 NC7 37.4 10.0 100.0 IC4 107.9 10.0 200.0 IC8 37.7 10.0 100.0 NC4 107.2 10.0 200.0 TOL 519.2 300.0 650.0 IC5 86.8 10.0 200.0 NC8 8.7 1.0 50.0 NC5 50.7 10.0 200.0 PX 385.1 200.0 500.0 IC6 179.9 50.0 350.0 MES 124.8 50.0 200.0

The results of the optimization calculation are shown in table 6.4. The moleflow and molefraction of the components in the different streams are shown in appendix C.

TABLE 6.4: The Assumed Feed. Calculated Feed. Measured Netto Feed and Calculated Netto Feed of the toluene recovery system.

Feed of the flash vessel VE201 Netto Feed of the flash vessel VE201 COMP

ID Assumed Calculated Difference Measured Ca lcu la ted Difference

[kmo l/hr] [kmo l/hr] [kmol/hr] [ma le%] [kmo l/hr] [kmo l/hr] [kmol/hr] [ma le%]

H2 8806.6 8842.50 - _35.90 - _{0.41 1946.3 1900.99 45.31 2.38} Cl 389.0 395.68 - 6.68 - 1. 69 85.9 87.61 - _{1.71 - 1.95} C2 499.6 433.15 66.45 15.34 110.4 111.43 - _{1.03 - 0.92} C3 670.6 432.96 237.64 54.89 148.2 148.57 - _0.37 - _0.25 lC4 48.3 105.08 - 56.78 - 54.04 48.2 47.90 0.30 0.63 NC4 56.9 106.99 - 50.09 - 46.82 55.0 56.04 - 1. 04 - _{1. 86} IC5 58.6 86.73 - 28.13 - 32.43 58.7 60.00 - _{1. 30 - 2.17} NC5 36.8 49.04 - 12.24 - 24.96 36.8 36.42 0.38 1.04 lC6 150.5 176.29 - 25.79 - 14.63 150.4 150.79 - _0.39 - _0.26 NC6 80.5 90.61 - _{10.11 - 11.16 80.5 80.52 - 0.02 - 0.02} BZ 255.2 275.17 - 19.97 - _{7.26 255.0 254.96 0.04 0.02} IC7 102.6 107.60 - 5.00 - 4.65 102.6 101.38 1. 22 1.20 NC7 35.0 35.95 - 0.95 - 2.64 35.1 34.48 0.62 1. 80 IC8 35.2 36.22 - _{1. 02} - 2.82 35.1 34.49 0.61 1.77 TOL 495.1 507.61 - 12.51 - 2.46 494.9 494.64 0.26 0.05 NC8 8.4 8.25 0.15 1.82 8.4 8.13 0.27 3.32 PX 374.0 375.91 - 1. 91 - 0.51 373.9 372.91 0.99 0.27 MES 122.5 121 .04 1.46 1. 21 121. 8 120.75 1. 05 0.87 TOT 12225.4 12186.78 4147.2 4101.97

It can be seen from these results that the measured and calculated netto feed are equal within a few percent. The Soave-Redlich-Kwong equation of state describes the behaviour of the vapor and liquid phases with a two percent error. This explains the difference between the measured netto feed and the calculated netto feed with ASPEN+.

Table 6.4 shows also that the assumption: the vapor product of the flash vessel contains all the hydrogen, methane, ethane and propane is not correct. This assumption is only right for hydrogen and methane. Propane and the C_4, Cs and C₆ hydrocarbons are leaving the flash vessel with both product streams. The difference between the assumed and calculated moleflow of hydrogen, ethane and benzene till mesitylene in the feed of the flash vessel is acceptable. The assumption th at propane leaves the flash

vessel as a vapor and the butanes as a liquid causes the greatest errors.

The difference between the assumed and calculated feed has a neglectible effect

on the sensitivity analyses which were carried out. This because of the main differences

appear in the amount of light components, which wiU be easily removed in the first distillation tower and sa the impact on the se para ti on behaviour in the next columns is neglectable. From this point the optimization calculations were carried out with the calculated feed.

6.2 Optimization of distillation tower 1W201

The composition of propane, butane, pentane and hexane in the feed of th is column is changed. The new feed composition has a positive effect on the condenser-temperature. The feed contains now less propane and more heavier alkanes. Because of this the boiling point of the top product becomes higher and watercooling is still possible. The reflux ratio and the bottom flow must be optimised to meet the given specifications for this column. Two optimization calculations with the improved configuration are carried out.

During the first optimization the molefraction of the pentanes in the top product

and the butanes in the bottom product are defined as constraint variables. The

molefractions must be less than 0.5 and 0.05 [mole%] respectively. The reflux ratio and the bottom flow are defined as variables that can be varied. The objective function chosen to maxi mi ze is the moleflow of the pentanes in the bottom product. The system

converged to an unacceptable solution. In this solution only the top product contains

pentanes and even part of the benzene and toluene are leaving the column with the top

product. This solution is not acceptable because of the great loss of toluene.

The second optimization calculation is the same as the first, except the bottom flow is nowa fixed variabie. The bottom flow [kmoljhr] is calculated with equation [6.1].

IC4

Bottom flow _{~F-TW201 - 1.05*~(~F-TW201 i)}

i=H2 I

[6.1]

The re su lts of the optimization calculation are shown in table 6.5. The composition of the streams are shown in appendix D.

TABLE 6.5: The optimised configuration

and operating conditions of distillation

tower TW201.

Unit TW201

Pressure [bar] 9.0

Number of Stages 20

Feed Stage 5

Condenser Type Partial

Temperature [·Cj 42.33 Duty [Watt -1.956E+6

Reflux Ratio [L/D] 1.95

Top Flow [kmol/hr 186.90

Reboiler Temperature [·C 188.94

Duty [Watt 1.772E+7

Boilup Ratio [V/B 1.27

Bottom Flow [kmol/hr 1715.60

Thermo-model SRK

SRK: Soave-Redlich-Kwong.

6.3 Optimization of the recovery system

Af ter the optimization of the first distillation tower, the only thing left is the optimization of the recovery system, distillation towers TW202 and TW203. It follows from the calculations already done, that the specification with respect to the C

_s

-components in the top of the column can't be met. This because of "azeotropic like" behaviour of n-octane and toluene. Therefore this specification has been dropped.

The system is regarded as a toluene recovery system and so the goal of the optimization will be to re ach the toluene purity specification and at the same time to keep the toluene loss as low as possible.

During the optimization the reflux ratios and the bottom flows of the columns TW202 and TW203 are varied while having the toluene purity constrained. The objective function to minimize is the toluene loss in bath of the columns. Because of the positive effect of higher reflux ratias on the degree of separation and sa on the toluene purity, the reflux ratias will probably reach their upper limits. The optimization settings are shown in table 6.6. The re su lts of the optimization calculation are shown in table 6.7.

TABLE 6.6: Optimizatlon settings.

Configuration

Feed

Object function

Constraint 1

Varied variable Range Varied variable 2 Range Varied variable 3 Range Varied variable 4 Range

Improved + Optimised column TW201

Calculated Feed Minimize Toluene loss

Fraction Toluene = 0.991 + 0.001 (top of tower TW201) -Reflux Ratio TW202 Lower = 2.5 Upper = 3.5 Bottom Flow TW202 Lower = 850.0 Upper = 950.0 Reflux Ratio TW203 Lower = 3.5 Upper = 4.5 Bottom Flow TW203 Lower = 500.0 Upper = 600.0

TABLE 6.7: The optimised configuration and operating conditions of the toluene recovery section.

Unit VE201 TW201 TW202 TW203

Temperature [ 'C] 38.33

Pressure [bar] 14.79 9.0 1.0 1.0

Number of Stages 20 30 30

Feedstage Position 5 10 15

Condenser Type Partial Total Total

Temperature ['C 42.3 65.7 109.7

Outy [MJ/hr -I .956E+6 -3.039E+7 -2 228E+7

Reflux Ratio [LID I. 95 3.5 4.5

Top Flow [kmol/hr 186.90 766.6 433. 19

Reboiler Temperature ['C 188.9 123.8 141.3

Duty [MJ/hr~ 1 .772E+7 2 075E+7 2 254E+7

Boilup Ratio [V/B 1. 27 2.19 4.16 Bottom Flow [kmol/hr] 1715.60 949.0 515.81

Thermo-model SRK SRK SRK

SRK: Soave-Redlich-Kwong.