Simulation Study Of Lube Based Extraction Unit

Paper No. - 098

SIMULATION STUDY FOR ENERGY OPTIMISATION IN SOLVENT EXTRACTION UNIT Ranvir Singh*, Manoj Srivastava, Manoj Kumar, U.C.Agrawal, M.O.Garg & S.M. Nanoti Indian Institute of Petroleum, Dehradun (India)

D.S.Laud HPCL, Mahul Refinery, Mumbai (India)

ABSTRACT Solvent Extraction units for dearomatisation of lube distillates are large energy consuming units in lube refineries. Solvent recovery is a very crucial step in these units which also involves input of large amount of energy. The present paper discusses results of simulation studies carried out for the solvent extraction unit of an existing lube refinery by simulation using available simulation tools (ASPEN PLUS & SPRINT heat exchanger network designs). Results of the case study with two different alternatives have been presented in the paper. In the first of these alternatives addition of a single flash drum has been proposed in extract circuit, which can lead to reduction in hot utility target by as much as 35.2%. In the other case, pseudo component approach has been used to predict the physical properties (specific gravity, mean average boiling point, pseudo critical temperature & pressure, Watson characterization factor, heat capacity etc.) of different hydrocarbon mixtures viz feed, raffinate and extract. The 500N distillate stock, which is complex mixture of hydrocarbons coming from vacuum unit, has been considered as feedstock in the study. Physical properties obtained through simulation have been compared with the actual property values obtained in test run under controlled process conditions such as throughput, pressure drop, temperature and other parameters. Good agreement has been found between the simulation results and the test run results. It has been established that the existing unit can be operated with as much as 20% higher throughput without any hardware changes or provision of any new hot or cold utilities. Key words: solvent extraction, optimization, simulation, heat integration, flash drum. *To whom correspondence may be addressed: Ranvir Singh Address & Tel. No: Scientist, lube & Wax Area, IIP, Dehradun Ph.No-0135-2660113-115 (extn.: 272) Email: [email protected]

1.0

INTRODUCTION The primary objective of solvent extraction of lube distillates is to remove the aromatics. This removal of undesirable aromatic part leads to tremendous improvement in product quality in terms of VI as well as oxidative & thermal properties of finished lube oil base stock (LOBS). This paper deals with simulation and energy conservation on solvent extraction unit (SEU) of one of the existing refinery involved in production of lube oil base stock (LOBS). Lube oil is a complex mixture of different constituents where prediction of average molecular weight/molecule size distribution and also it behavior is quite difficult due to presence of some non-hydrocarbon materials therefore, pseudo component approach which obviates need for such prediction, has been used to simulate the existing SEU. The SEU currently consumes a total of nearly 13.5 MW energy including steam and fuel oil. Bulk of this energy goes to vaporize extract solution to recover solvent from extract solution, having around 85% solvent. To reduce this energy consumption and parameter optimization of SEU, a broad based investigation has been carried out with the objective of feasibility of further improving unit performance are establishing throughput maximization in treater tower. Energy optimization using pinch analysis has also been attempted and reported elsewhere. Findings of subsequent work on simulation studies aimed at utilization of energy saving established feasibility of increasing throughput by as much as 20 %. Thorough analysis of plant test run data & existing duties of the furnace as well as heat exchangers in raffinate & extract circuits through their design sheet has been carried out to extract out relevant data which formed the basis of study. Selection of suitable thermodynamic model which gives good interaction between lube & NMP solvent equilibrium data is a key step as such data provides important input for simulation by matching of data for actual feedstock and model hydrocarbons representing this feedstock. ASPEN PLUS version 13.1 & SPRINT Heat exchanger Network-2003 version have been used as simulation tool to simulate existing flow sheet.

PINCH ANALYSIS OF EXTRACT CIRCUIT 2.1

Case – I : Base Case The unit process flow diagram has been subdivided for convenience into two different sections (1) Raffinate Section (2) Extract section. This was due to the fact that most of solvent goes with extract stream and therefore large amount of energy recovery is expected to be possible through improved extract section design. Simplified flow diagram has been represented in Fig2. Stream data obtained through plant data sheet and analytical data generated in-house has been simulated with the simulation tool SPRINT. Composite curve & Grand composite curve based on these data has been prepared, Information obtained through CC & GCC has been presented in Table 1. Change in minimum approach temperature leads to change in hot & cold utility duty. TABLE- 1 PROGRAMME TABLE FOR EXTRACT CIRCUIT BASE CASE ΔTmin Minimum Hot Utility Minimum Cold Utility Pinch Temperature Cold Pinch Temperature Hot Pinch Temperature

10 9571.6 10411.6 250.4 245.4 255.4

[˚C] [kW] [kW] [˚C] [˚C] [˚C]

After Steam balance & cross pinch study for the base case in Extract circuit, base heat exchanger Network diagram has been established as shown in fig 4. The key findings from the Extract circuit data extraction and pinch analysis are as follows: 1. The composite curve and grand composite curve (GCC) for this extract circuit have been plotted from above stream data and exhibited in figure 3 and 4 respectively. These plots show scope for heat recovery from hot stream to cold stream respectively. 2. The hot utility target is 9571.6 kW against the existing heat duty of 10850kW. 3. The cold utilities target is 10411.6 kW against the existing cold duty of 11690kW. 4. Difference of hot utility target and actual operation is 1278.4 kW which can be utilized in the process. TABLE-2 CROSS PINCH REPORT FOR EXTRACT CIRCUIT BASE CASE Δ Tmin Minimum Hot Utility Minimum Cold Utility Total Hot Utility Total Cold Utility Total Cross Pinch Heat Transfer

10 9571.6 10411.6 10850.0 11690.0 1278.40

[˚C] [kW] [kW] [kW] [kW] [kW]

RAF

EXT

FIG 2 :INDEPENDENT PROCESS FLOW DIAGRAM WITH ADDITIONAL FLASH DRUM

FIGURE 4: GRAND COMPOSITE CURVE FOR EXTRACT CIRCUIT BASE 400

FIGURE 3 : COMPOSITE CURVES FOR EXTRACT CIRCUIT BASE 350

350

TEMPERATUTRE

300

TEMPERATU

2.0

250 200 150 100

0

0

5000

10000

15000

20000

ENTHALPY kW

25000

250 200 150 100

HOT COMPOSITE COLD COMPOSITE

50

300

50 30000

0

0

2000

4000

6000

800

ENTHALPY kW

10000

12000

187.79 297.44

22,23ext

1

E 4215 N:1

N:28 90.62

188.37

25,27ext

2

28-31Solv

3

33-2 Solv

4

E4212 N:2

N:16 255.58

224.20

214.34

155.65

149.42

260.40

E4205

E-4204 N:3

N:12

E4214 N:18

E-4203 N:20

Reflux N:14

N:24

94.01 149.00

E4202 N:4

N:22 224.55

325.83

182.25 85.00

19ext phs

F4201

E-4204

E-4203

N:0 A:287.9514

N:15 N:0 A:534.0814

N:13

N:27

*Q:1759.0

5 N:5 126.59 85.00

*Q:4044.0

18ext phs

6

Reflux N:25 185.60

bfw

N:6

N:0 A:52.26148 *Q:265.0

165.60

7

E4214 N:21 186.10

steam

N:7

N:0 A:107.5373 *Q:1046.0

185.68 185.60

E4205 N:19 N:0 A:495.265 *Q:6388.0

8

E 4215 71.00

rcw

N:29 N:0 A:270.6385 *Q:1209.0

N:8

54.40

9

E4212 N:17

N:9 N:0

49.00

saltwater

A:153.2925 *Q:910.0

10

E4202 N:23 380.00 400.00

furnace main

35.00

11

N:10

N:0 A:66.27748 *Q:2137.0

F4201 N:11

N:26 N:0 A:511.1005 *Q:10850.0

FIGURE 5: HEAT EXCHANGER NETWORK FOR EXTRACT CIRCUIT BASE CASE

3.0

PROCESS MODIFICATION 3.1

Case –II : Pinch analysis of extract circuit with additional flash stage Further examining Grand composite curve of Extract circuit reveals that large enthalpy below the pinch is unutilized which is mainly the source stream. To efficiently utilize the above heat, introduction of an additional flash drum for partially vaporizing the run down bottom has been envisaged as process modification, as shown schematically in Fig 5. Stream no. 4 comes out from treater tower T4201 bottom. This stream is preheated up to 221 °C at gauge pressure of 2.5 bars by flash overhead vapors. The condition for flashing has been achieved through simulation of data in ASPEN and to get the temperature & pressure data at different condition of this modification. At the condition selected, 21% of the rundown bottoms of T4201 vaporize under adiabatic conditions maintained in the flash drum. Bottom of the flash drum is pumped and preheated through exchanger. After preheating, stream no.4 is designated as stream 19 and proceeds towards furnace F4201 and is redesignated as stream no.20 on exit. Part of stream 4 is routed to T4202 (Extract Recovery Tower) top as a reflux at 125°C which is receiving heat from heated Extract solution ex E4204. This is basically a mixer system where two streams are mixing, resulting in a new stream no.18 at 125°C. For pinch analysis, this mixing system is considered as an invisible heat exchanger. On examining the composite curve and flow sheet data, it is observed that hot stream coming out from T4202 & T4203 top (stream no 28) at temperature 260°C looses heat in Exchangers E 4204,E4205, E 4214 & E 4203, and finally comes out as stream no. 31 at temperature 150°C. This heat loss in exchanger E4204 & E4203 is utilized to preheat Extract solution stream no. 4; the resulting heated streams are stream no. 18 & 19. For this case with addition of flash drum, the entire stream data has been revised. For hot utility in Extract circuit, furnace F4201, heat supply temperature is taken as 400 °C and target temperature is taken as 380°C. The other hot utility considered is steam available at 250 °C with target temperature of 230 °C. Similarly for refinery cooling water, a cold utility, supply temperature is taken as 40 °C with target temperature of 75 °C. Stream data of has been analyzed for pinch analysis, taking ΔΤmin to be 10 °C and the results obtained are reported in. The Network diagram from the above stream data has been established and exhibited in figure 8. Following are the key findings from the Extract circuit data extraction and pinch analysis : 1. 2. 3. 4. 5. 6.

Pinch temperature is reduced to 211˚C in the modified from 250 °C in base case Pinch occurs at a temperature of 2160C and 2060C for hot & cold streams respectively. The composite curve and grand composite curve (GCC) for this modified circuit have been plotted from above stream data and exhibited in figure 7 and 8 respectively. The hot utility target is reduced to 7030 kW against the existing heat duty of 8480kW. The reduced cold utilities target is 6577 kW against the existing cold duty of 8042kW. Difference of hot utility target and actual operation is 1460 kW which can be utilized in the process.

7.

The addition of Single flash drum leads to reduction in hot utility target to 7030 KW against the existing base case heat duty of 10850 which shows around 35.2% reduction in heat duty.

Additional flash drum leads to Reduction in total hot utility. This modification thus s to reduction in duty of furnace F4201 by 2820kW, which is about 35.2% lower than current operation. Table – 3 Problem Table ΔTmin Minimum Hot Utility Minimum Cold Utility =

10 8480.39 8042.02

[C] [kW] [kW]

Figure – 7 : Grand Composite Curve

Figure – 6 : Composite Curve 197.62

95.00

221.00

103-104 I f

1

2 N:1

Rcw N:26

N:24

229.76

116.68

95.03

260.00

128-108 II f

2

1 N:2

3 N:18

Rcw N:10

230.54

N:22

188.10

297.00

126-129 r/d

3

129-130 r/d

4

5 N:3

Rcw N:12

N:16

99.84 188.00

rcw N:4

N:20 221.00

210.00

214.43

164.63 85.00

ext sol-100

1 N:19 N:0 A:74.76106 *Q:2500.0

HP

N:27 N:0 A:312.2093 *Q:3600.0

furnace

N:14 75.00

230.00 250.00

9

N:6

duty

N:7

cu steam

6

N:13 N:0 A:25.23295 *Q:650.0

380.00

7

N:5

221.00

5

duty

N:15

400.00 furnace

N:11 N:0 A:77.56682 *Q:3600.0

242.62

327.00

109-86 B24 furnace

5

3

2

steam

N:29

HP

N:0 A:59.24859 *Q:5330.0

49.78

45.87

43.66 40.00

Rcw

rcw

N:25 N:0 A:51.83297 *Q:4742.0

N:21 N:0 A:8.204612 *Q:735.0

Rcw N:17 N:0 A:2.53468 *Q:415.0

8

Rcw N:23 N:0 A:10.83367 *Q:689.0

N:8

steam

N:9

N:28 N:0 A:62.49553 *Q:1690.0

Figure 8 : Heat Exchanger Network Process Modification

3.2

CASE-III base case Simulation In the extraction step, the selectivity and capacity of the solvent is a strong function of hydrocarbon type composition of the feedstock e.g. saturates aromatics, naphthenes and carbon number. The most important input to the Aspen for correct predictions is the Binary Interaction parameters. For LLE prediction, NRTL/UNIFAC/UNIFAC-LL model were chosen and the existing plant data was matched with the inbuilt Aspen parameters; the discrepancies were subsequently corrected by choosing the model which predicted data closer to the actual plant data. ASTM D1160 data, Specific gravity, Mean average boiling point are the parameters used as pseudo component model parameters.

4.0

Conclusions Simulation study has been carried out with base case study of existing plant data & with close matching with simulation data. • • • • •

The addition of Single flash drum leads to reduction in hot utility target up to 7030 KW with reference to the existing base case heat duty of 10850 which shows around 35.2% reduction in existing heat duty in extract circuit. Experimental data indicates production of LOBS of desired purity feasible with increase in throughput of up to 20 % without change in existing system & without installation of any new exchanger or any other unit ,However this require use of all the hot & cold utilities effectively Simulation carried out for test run data indicates a good match The results have been compared in terms of key parameters e.g. purity and recovery of products, heat exchanger & furnace duties of extract recovery unit & raffinate recovery units. The results further indicate that LOBS of desired purity can be produced for 500N feedstock.

5.0

References 1. B. Linnhoff et al., User Guide on Process Integration for the Efficient Use of Energy, IChemE, Rugby, UK, 1982 (last edition 1994). 2. Gorazd Soboˇcan, Peter Glavic, A simple method for systematic synthesis of thermally integrated distillation sequences, Chemical Engineering Journal 89 (2002) 155–172 3. Anna Fritzson, Thore Berntsson, Efficient energy use in a slaughter and meat processing plant-opportunities for process integration, Journal of Food Engineering 76(2006) 594–604. 4. Majda Krajnc , Anita Kovac-Kralj, Peter Glavic, Heat integration in a specialty product process, Applied Thermal Engineering 26 (2006) 881–891 5. Kejin Huanga, Koichi Iwakabe, Masaru Nakaiwaa, Atsushi Tsutsumi, Towards further internal heat integration in design of reactive distillation columns—part I: The design principle, Chemical Engineering Science 60 (2005) 4901 – 4914 6. Douglas J.M.,Conceptual, Design in process engineering, McGraw Hill publication (2rd edition) 7. Gorsek A, P.Glavic, M.Bogataj, Design of the optimal total site heat recovery system using SSSP approach Chemical Engineering & processing 45(2006),372-382.