00 positive Bewertungen00 negative Bewertungen

6 Ansichten12 Seitendescribed about Retrofit Optimization Framework for Compact Heat Exchangers

Jan 17, 2017

© © All Rights Reserved

PDF, TXT oder online auf Scribd lesen

described about Retrofit Optimization Framework for Compact Heat Exchangers

© All Rights Reserved

Als PDF, TXT **herunterladen** oder online auf Scribd lesen

6 Ansichten

00 positive Bewertungen00 negative Bewertungen

described about Retrofit Optimization Framework for Compact Heat Exchangers

© All Rights Reserved

Als PDF, TXT **herunterladen** oder online auf Scribd lesen

Sie sind auf Seite 1von 12

Publisher: Taylor & Francis

Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,

37-41 Mortimer Street, London W1T 3JH, UK

Publication details, including instructions for authors and subscription information:

http://www.tandfonline.com/loi/uhte20

Exchangers

IGOR BULATOV

Centre for Process Integration, SCEAS University of Manchester Institute of Science and

Technology

Published online: 23 Feb 2007.

To cite this article: IGOR BULATOV (2005) Retrofit Optimization Framework for Compact Heat Exchangers, Heat Transfer

Engineering, 26:5, 4-14, DOI: 10.1080/01457630590927273

To link to this article: http://dx.doi.org/10.1080/01457630590927273

Taylor & Francis makes every effort to ensure the accuracy of all the information (the Content) contained

in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no

representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the

Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and

are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and

should be independently verified with primary sources of information. Taylor and Francis shall not be liable for

any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever

or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of

the Content.

This article may be used for research, teaching, and private study purposes. Any substantial or systematic

reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any

form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://

www.tandfonline.com/page/terms-and-conditions

C Taylor & Francis Inc.

Copyright

ISSN: 0145-7632 print / 1521-0537 online

DOI: 10.1080/01457630590927273

for Compact Heat Exchangers

IGOR BULATOV

Downloaded by [Australian National University] at 17:36 26 January 2015

Centre for Process Integration, SCEAS University of Manchester Institute of Science and Technology, Manchester, UK

A lack of published retrofit methods that account for specifics of plate fin heat exchangers is a serious problem for industrial

projects in which the modernization of such heat exchanger networks is carried out. The author proposes an overall framework

for the retrofit of a plate fin heat exchanger network that has been developed and applied for the debottlenecking of a liquefied

petroleum gas cold box. The framework covers both levels: single exchanger unit and the exchanger network. This retrofit

framework builds on a grassroots design method that has recently been developed and applies a retrofit matrix approach

to obtain volume distribution for new operating conditions. Two sets of heat transfer coefficients, those for existing and

additional volumes, are used. New volume heat transfer conditions are optimized during the procedure; thus, no pre-assumed

coefficients are used, as is the case with previous design and retrofit methods. Pressure drop parameters crucially important

for this sort of equipment are also incorporated into the optimization framework.

INTRODUCTION

this type of exchangers among them. In this paper, the author

addresses one more significant problem: the lack of systematic retrofit methods for heat exchanger networks (HENs) consisting of CHEs that hinders the wider application of plate fin

exchangers.

Industry and academia have gained significant experience in

the design and maintenance of these units. The Energy Efficiency Best Practice program [2], launched by the UK Department, is a very good example of intensive efforts to disseminate

the best practice concerning compact exchangers. Works covering the overall problem as well as various parts of it have been

published by prominent researchers (e.g., Hesselgreaves [3, 4];

Polley et al. [5]). A considerable amount of research on heat

transfer and friction characteristics of various fin geometries

have been published in recent years (e.g., [69]). Very interesting systematic grassroots design procedures were published in

recent years as well [1013]. However, no reliable methods of

retrofit of plate fin heat exchanger networks have been published

so far. The complexity of this problem equals its significance,

as, the majority of projects carried out in modern process industries in industrialized countries deal with retrofit problems. In

this paper, such a retrofit method for plate-fin heat exchanger

networks will be described.

Our retrofit methodology builds on the plate-fin HEN grassroots design approach developed in [12], who adopted parameter correlations from [14], and a retrofit area matrix proposed

in [15] and further updated by [16]. Detailed descriptions of

Compact heat exchangers (CHEs) have been successfully applied by industries for many decades, and recent trends indicate

expansion of the range of their application and blurring sector

divisions. Plate fin heat exchangers proved to be a very efficient

type of compact heat exchangers. They allow the significant

intensification of the heat transfer process due to their construction features, namely, high area density, usage of secondary

surface, counter-current flow arrangement, small temperature

difference, and multistream configurations. Though generally

more expensive in terms of individual manufacturing than conventional shell-and-tube units, compact heat exchangers prove

to be cost effective in terms of weight, space, and their lower

installation costs. Lower utility consumption, less flue gas emission, and reduced compressor power for refrigeration systems

contribute to operating cost savings. Lower quantities of holdup provide improved safety. This said, process industry overall still regards conventional shell-and tube exchangers as the

work horse for heat transfer processes and plate fin exchangers

as suitable for certain applications. There are several obstacles

to the wider spread of plate fin exchangers [1], with fouling

The work has been carried out in the framework of the Royal Society/NATO

Postdoctoral Fellowship Programme. The author would like to thank the Royal

Society/NATO Postdoctoral Fellowship Programme for granting the award.

Address correspondence to Igor Bulatov, Centre for Process Integration,

SCEAS University of Manchester Institute of Science and Technology, P.O.

Box 88, Manchester, M60 1QD, UK. E-mail: igor.bulatov@manchester.ac.uk

I. BULATOV

overview of the two methods will be given here.

PFHE Network Design Method

includes:

1. Temperature optimization in entrance and exit points. Strictly

following an identical entrance-exit temperature arrangement in multistream heat exchangers leads to a considerable amount of unutilized space, extra headers for redistribution of streams, and additional volume and pressure drop.

To overcome this, an extension of streams to other intervals (see Figure 1) has been proposed. However, reduced

unutilized empty space has to be traded off against the reduced temperature driving force (which requires an increase

in heat transfer area). Hence, an optimization procedure is

required to maximize temperature driving forces and

minimize the number of intermediate headers. This will also

minimize the heat transfer area even if the heat transfer conditions in the exchanger to be designed are not yet known.

2. Unit targeting. With entrance and exit point temperatures

for each stream in intervals already determined, the targets

for network volume, frontal area, and pressure drop and the

minimum number of units are calculated.

3. Distribution optimization. This procedure analyzes different stream distributions, because for equal volumes, an exchanger having a smaller number of streams is cheaper than

a unit with more streams. The variables optimized are the

volume of each targeted exchanger and the number of streams

in each exchanger.

4. Fin optimization. Heat transfer conditions for each stream

are revised in an attempt to make lengths of streams in each

exchanger similar.

arrangement that provides the maximum number of heat exchanger units (see Figure 2b). The target area matrix is assumed

to be flexible.

So far, the two networks are hardly compatible, as the existing

area has already been distributed among stream matches and the

target area matrix neglects this fact. An optimization procedure

is applied to maximize compatibility between the two networks

by shifting heat within each enthalpy interval of the spaghetti

arrangement of the target network (see Figure 3). In the resulting

deviation matrix, an excess area emerges when the difference

between the corresponding element of the retrofit and existing

matrix is negative while the additional area required when the

difference between the corresponding element of the retrofit and

existing matrix is positive.

The retrofit area matrix approach was proposed by [15] and

further extended by [16]. The existing area matrix parameters

are considered to be fixed (see Figure 2a). New operating conditions lead to a target network designed using a spaghetti

I. BULATOV

The conventional shell-and tube HEN retrofit has been addressed in a number of works, and quite efficient retrofit procedures exist today. A number of optionsresequencing, repiping, additional area, and stream splittingare applied to overcome the network pinch [17, 18]. However, the specifics of

plate fin heat exchangers do not allow to directly apply those

means to overcome the network pinch. Since plate fin exchangers are brazed or welded units, severe constraints on topology

changes and existing unit design changes can be imposed during

retrofit:

dimensions of the existing unit remain unchanged

stacking pattern of the existing unit remains unchanged

match volume of the existing unit remains constant

no additional intermediate headers can be inserted.

As a result, the conventional means of overcoming the network pinch for this type of exchangers may prove highly constrained or even impossible in some cases. All this makes the

development of a plate fin exchanger retrofit methodology quite

a complex task.

Though the proposed retrofit method builds on the four-step

grassroot design approach discussed earlier, significant differences are required to take into account the existing topology in

the first two steps.

Temperature Difference Maximization

The aim of this stage is to determine temperature difference

profiles in intervals. Thus, partition temperatures in intervals are

regarded as optimization variables.

New operating conditions lead to new temperature intervals

for the retrofitted HEN. But for some cases, new operating conditions will not change temperature intervals, and the temperature

intervals will remain as in the initial HEN (debottlenecking with

a case). This driving force optimization procedure considers it

as a separate branch. Thus, only a new unit with optimal temperature profiles will have to be optimized, which simplifies the

retrofit design task.

The optimal profile search path for new operating conditions

and new temperature intervals is shown in Figure 4. The optimal

profile search path for new operating conditions with same temperature intervals is shown in Figure 5. A target stream structure

that follows vertical heat transfer shows how temperature intervals have changed under the new operating conditions. We can

also see whether the target structure retains the initial temperature intervals or whether they have changed. For a grassroot

design case, the driving force optimization procedure described

in [12] is enough; however, it might be better to retain the existing stream matches. Overall, the final retrofit stream structure

will include two parts: streams in the existing volume and those

in additional volume.

Minimization of unutilized volume may require stream extension to other intervals, which implies a violation of the vertical

heat transfer arrangement. For existing units, a stream extension

may be a solution that leads to full utilization of the existing

volume. For additional units, stream extensions to other intervals will result in a reduction of the number of intermediate

headers and unutilized space, thus reducing the length and the

volume of a new unit (see Figure 1). However, lower driving

forces will result in a volume penalty. The optimization procedure analyzes and avoids temperature cross and transverse heat

conduction by imposing constraints. The procedure also checks

that the temperature driving force should exceed 1 C and that a

reasonable number of intermediate headers are in the heat exchanger. All these constraints can be expressed by Eqs. (16) in

the Appendix.

As [19] shows, the critical ratio of temperature difference,

rcrit , depends on heat conductivity of the fin, k; fin thickness, ;

fin height, ; and heat transfer coefficient, h. This means that the

value of rcrit strongly depends on the exchanger itself. Another

author [12] suggested a range of values assuming standard fin

Figure 4 Retrofit stream structures for different temperature intervals under new operating conditions and temperature intervals.

I. BULATOV

Figure 5 Retrofit stream structures for different temperature intervals under new operating conditions but the same temperature intervals.

transfer coefficient greater than 200 W/m2 K, the value of 0.21

can be assumed as an estimate, and for a gas stream with a heat

transfer coefficient greater than 200 W/m2 K, the value can be

assumed in the ballpark of 0.65. The values can be fine-tuned

later during the design process.

If new temperature intervals appear in the driving force optimization procedure, it means that the existing stream beginning/end needs to be extended or reduced. The range of interval

changes may include:

stream end extension

stream beginning reduction

stream end reduction

operating conditions result in a stream reduced at the beginning

and extended at the end, or vice versa).

The arrangements of temperature interval extensions or

reductions for existing and additional streams can be different. For a set of existing and additional stream structures and

vertical/non-vertical heat transfer conditions, four general combinations can be considered (see Table 1). Hence, the overall

selection of the optimal retrofitted stream structure requires all

four combinations.

As has been pointed out earlier, vertically arranged streams

in plate fin heat exchangers usually require more unutilized

space and extra headers for the redistribution of streams. This

is why only two options are considered here. The first option

is when existing streams remain vertically arranged while ad-

both existing and additional streams are optimized. Thus, in this

procedure, the new volume temperature differences are always

maximized, and the existing volume driving forces can either

follow vertical heat transfer or also be subjected to maximization, as shown in Figure 6.

In practical terms, if the new operating conditions require

an extension of the stream beginning or end, the mass flowrate

should be reduced in the existing stream structure. If, on the

other hand, the stream beginning or end reduction is required,

the mass flowrate should be increased in the existing structure

in compliance with the heat balance. To illustrate this, a cold

stream j beginning extension option is shown in Figure 7. Thus,

in the driving force optimization procedure, we should explicitly

specify or prohibit possible stream splits/mixes in the intermediate intervals of existing and additional structures. As can be

seen from Figure 7, the driving force optimization procedure

needs to choose between the two possible cases.

The objective function of the procedure includes maximization of the temperature driving force profiles in each interval,

minimization of the transverse heat conduction, and a number

of extra intermediate headers (see Figure 8, Appendix Eqs. 7

10).

Existing

Additional

Existing

Additional

Vertical

Non-vertical

Vertical

Non-vertical

Vertical

Non-vertical

Vertical

Non-vertical

Existing

Additional

Existing

Additional

I. BULATOV

In this work, the driving force profile optimization was modelled using General Algebraic Modelling System (GAMS) [20]

and solved using BDMLP, a linear programming and mixed integer programming solver supplied with GAMS, because this is

a typical mixed integer linear programming problem.

First, the approach described in the literature [15, 16] applies constant heat transfer coefficients for both existing and

additional areas. This makes feasible the same basis for area

comparison. Constant conditions may be possible (arguably)

for conventional shell-and-tube exchangers, but this is not the

case with plate fin heat exchangers. The big difference in the

heat transfer coefficient pattern between the plate fin exchanger

Unit, Volume, and Frontal Area Targeting

and a shell-and-tube exchanger is that in a PFHE, heat transfer

This stage determines the minimum number of additional coefficients can vary across different intervals due to changing

units, overall volume, frontal area, and interval pressure drop fin types, frontal area, condensation, and evaporation.

Therefore, the procedure introduces two sets of heat transfer

targets of the network. The insights of the grassroots design

coefficients

that account for the differences between existing and

targeting procedure are used as the basics in this stage of the

new

units.

The

existing volume is calculated via existing heat

retrofit procedure by taking into account the existing structure

transfer

coefficients,

and a set of new coefficients is applied to

and applying the retrofit area matrix methodology. But there are

the

new

volume.

If

new

coefficients were applied to the targeted

two major features that distinguish the current procedure from

area,

they

would

effect

a whole network volume based on new

the retrofit area matrix approach used previously.

coefficients, and no existing matches with different conditions

would be taken into account. Because volume has a different

comparison basis in this case, the optimization of differences

between existing and new volumes will not be correct.

We apply a spaghetti arrangement for each interval and

place an additional match in parallel with the existing match (see

Figure 9) so that the temperature differences of the existing and

new matches will be the same. Theoretically, a feasible, serial

spaghetti arrangement can be implemented only for end intervals

of plate fin HEN because it is impossible to insert additional

matches into the core of the existing unit.

Secondly, the procedure optimizes new volume heat transfer

conditions at the targeting stage, building on the methodology

reviewed earlier in this paper. This is an important difference

between the current procedure and previous methods that used

pre-assumed fin types (hence, pre-assumed heat transfer conditions) at the initial stages of the design.

Figure 8 Entrance and exit points in T-K diagram.

heat transfer engineering

vol. 26 no. 5 2005

I. BULATOV

A brief description of the model is given below, and the equations are presented in the Appendix.

The model keeps energy balances in each interval. The energy balance for hot streams in an interval equals the sum of

heat loads transferred by cold streams of existing and additional

matches (Appendix, Eq. 11) and the energy balance for cold

streams equals the sum of heat loads transferred by hot streams

of existing and additional matches (Appendix, Eq. 12).

Assuming identical fins on hot and cold sides (i.e., applying the identical fin concept developed by [13]), the additional

volume can be explicitly expressed (Appendix, Eq. 13) as a function of two independent parameters: the fin related property, Y,

and combined stacking pattern and process related property, wv

[12]. New heat transfer conditions (rather than conventional heat

transfer coefficients) embedded in the fin properties parameter,

Y, are incorporated within the equation of the additional volume of the match in the interval. This makes redundant any initial specification of fin types and subsequent iteration between

fin selection and network design. Fin types and stacking pattern of an additional unit are unknown prior to design, which

is why both fin related property and stacking pattern/process

related property are subject to optimization during the retrofit

procedure.

As follows from an analysis of the matrix structure, if the additional heat load is negative (Appendix, Eq. 20), no additional

area is required and zero load will be applied, whereas the positive value of the additional heat load indicates additional area

requirements.

As to the existing part of the heat exchanger network, its heat

transfer coefficients, fin efficiency, and geometry of the existing

units are known prior to the retrofit, so these can be incorpo-

Streams

Ts ( C)

TT ( C)

A

B

C

D

E

F

76.09

15.11

5.87

55.11

10.29

13.43

32

33

62.8

40.4

38.2

42.3

33

27

27

14.2

24

42.1

16,398

3,838

2,966

6,862

1,279

1,454

Pall (kPa)

60

100

50

50

30

10

Eq. 21).

The pressure drop is addressed the similar way as the new

volume (Appendix, Eqs. 2228), in that identical fins on both

sides are assumed, and two independent parameters constitute

the equations: fin related properties, which are lumped into Z,

and combined stacking pattern and process related properties,

which are lumped into wPH (wPC ). The number of additional

units (Appendix, Eq. 29) can be determined as the difference

between the total frontal area and existing area less 1, which

means that we assume the maximum possible frontal area for

the new unit to be 1 m2 .

The objective function of the targeting procedure minimizes

the sum of the match additional volumes (Appendix, Eq. 30).

The volume and additional unit targeting stage is a mixed integer non-linear programming problem, with non-linearity coming from Z-Y graph correlation. In this work, the targeting stage

was modelled using GAMS and solved by applying a non-linear

solverin this case, CONOPT2 of ARKI Consulting and Development [22].

The last two stages of the retrofit procedure, aimed at optimization of stream distribution and fin optimization, remain

basically the same as for grassroot design. However, only additional volume should be considered, as the existing units cannot

be modified due to the nature of PFHE.

PFHE RETROFIT CASE STUDY

The proposed retrofit procedure has been applied to a case

study of LPG plant coldbox, which was first introduced in [12]

to demonstrate the PFHE network design method. The original

design and the process data are shown in Figure 10 and Table 2.

Q (kJ/s)

10

I. BULATOV

11

A.B

A.C

A.D

A.E

A.F

B1

B2

B3

B4

B5

B6

B7

B8

B9

B10

B11

0.003

0.001

0.021

0.001

0.003

0.001

0.016

0.001

0.003

0.001

0.081

0.001

0.111

0.001

0.004

0.004

0.003

0.134

0.003

0.003

0.003

0.112

0.003

0.001

0.003

0.010

0.001

0.001

0.003

0.007

0.001

0.001

0.003

0.013

0.003

0.001

0.003

0.012

0.003

0.001

0.003

0.001

0.003

0.001

0.052

B12

B13

B14

B15

0.001

0.001

0.009

0.001

0.013

0.026

0.026

0.026

type selection for the additional unitcan be carried out. The

fins required for the additional unit are shown in Table 5.

The summary data of the performance of the modified network are shown in Table 6. For comparison, two other options,

a conventionally designed additional unit and a new grassroots

design network, are also given. The figures show that for the

conventional design, the additional volume (hence capital costs)

is considerably higher than in the proposed approach, namely

3.16 m3 vs. 0.74 m3 . The reason for this is the vertical heat transfer arrangement of streams leading to a high number of intermediate headers and unutilized volume. If the old network is to be

scrapped and Puas approach to new grassroots design applied

for our debottlenecking case, another viable option (2.75 m3 )

can be suggested, especially when radical plant modernization

is under consideration. But as can be seen from the figures above,

the proposed retrofit approach gives considerable gains in terms

of smaller volume.

The petroleum gaseous feed is cooled down before entering separation stages.

The existing coldbox was designed following the conventional approach, according to which streams are arranged in a

vertical heat transfer manner. Streams D and E exit at the intermediate points of the exchanger, leaving a considerable amount

of unutilized space (see Figure 10). Each unit has the same configuration and same volume of 3.95 m3 .

Our debottlenecking problem was the increase of throughput

by 20% and proportional change in all stream mass flowrates.

Overall there are three options:

1. to add an extra, similar unit to the existing one

2. to add an extra unit designed using the proposed approach

3. to install a newly designed network, having scrapped the

existing one

Economic considerations as well as the overall targets of the

retrofit should dictate the option.

At the first stage, we obtain the optimized driving force profiles (see Table 3). As a result, the general stream structure is

clear from this table.

The maximized temperature difference profile implies minimal volume for the network. At the next stage, the network

volume, frontal area, pressure drops, and number of additional

units are targeted. For this debottlenecking case (see Table 4),

the required targeted additional volume is relatively small, and

only one additional unit is needed. The pressure drops are below

the allowable values due to the Reynolds number limitations for

each stream. Because the units are arranged in parallel and the

additional unit is small, the overall pressure drop for the network

will not change.

According to the results of the unit targeting stage, only one

additional unit is required, so there is no need to carry out

CONCLUSIONS

The aim of this work was to develop a plate fin heat exchanger

network retrofit procedure. Such a procedure has been developed

building on the retrofit area matrix method and recently published grassroot synthesis method. The volume distribution for

the new operating conditions is obtained using the retrofit area

matrix approach, whereas the fin selection and physical insights

are shared with the grassroot design method.

The major difference with the previous retrofit matrix procedures is that it uses two sets of heat transfer coefficients: for

existing and additional volume. Whats more important is that

the new volume heat transfer conditions are optimal ones, obtained during the targeting stage, and not pre-assumed ones, as

it is often the case with previous design and retrofit methods.

A

B

C

D

E

F

B1

B2

B3

B4

B5

B6

B7

B8

B9

B10

B11

B12

B13

B14

B15

SF1

SF2

PF

SF1

PF

SF1

SF2

PF

SF1

PF

SF1

SF2

PF

SF1

SF3

SF1

SF2

PF

SF1

SF1

SF1

SF2

SF1

SF1

SF1

SF1

SF2

SF1

SF1

SF3

SF1

SF1

PF

SF1

SF1

SF1

SF1

SF1

SF1

SF1

SF1

SF1

SF1

SF1

SF2

SF1

SF1

SF1

SF1

SF2

SF1

SF1

SF3

SF1

SF1

LF1

SF1

SF1

SF1

SF1

SF1

SF3

SF3

SF1

LF2

LF1

LF1

SF2

SF1 = SF1/1027.03; SF2 = SF1/925.01; SF3 = 1/813.95; PF = PF12.0T; LF1 = 3/8(a)6.06; LF2 = LF1/26.06.

12

I. BULATOV

Units

(existing + new)

Volume, m2

(existing + new)

Number of headers

(existing + new)

Pressure drops, kPa

A

B

C

D

E

F

Traditional retrofit

approach

New retrofit

approach

Grassroot design

method

4+1

4+1

0+2

15.8 + 3.16

15.8 + 0.74

48 + 12

(6 intermediate)

48 + 12

(4 intermediate)

Fins

48

83

45

45

27

Louvered

0 + 2.75

0 + 18

(6 intermediate)

48

13.45

83

11.4

45

7.5

45

25.1

27

25.6

0.05

2.2

Serrated + Plain Serrated + Plain

+ Louvered

+ Louvered

plate fin heat exchanger design, the procedure meets pressure

drop constraints.

The procedure has been tested with case study data published

in literature. The results of the case study prove its ability to

find better solutions in comparison with new design approaches.

However, more hardware considerations in the retrofit procedure

still need to be included, especially that the greater number of

fin types might make the procedure more comprehensive and

flexible.

NOMENCLATURE

a

A

a

Afr

b

B

b

Bg

c

C

d

D

dh

DTRS

DTS

E

FT

h

H

coefficient in Y -Re relationship, m2

coefficient in fin effectiveness correlation

frontal area, m2

exponent in heat transfer correlation

exponent in Y -Re relationship

exponent in fin effectiveness correlation

ratio of primary heat transfer area between cold and hot

sides

coefficient in friction factor correlation

coefficient in Z -Re relationship, m1

exponent in friction factor correlation

exponent in Z -Re relationship

hydraulic diameter, m

sum of temperature differences between same streams

in MHE, K

sum of temperature differences in MHE, K

enhancement factor, ratio of the total heat transfer area

to plate area

correction factor to the log-mean-temperaturedifference

heat transfer coefficient, W/m2 K

enthalpy change, J

heat transfer engineering

HD

m

P

Q

r

rcrit

Rf

Re

SP

T

T lm

V

w

Y

Y

Ycorr

Z

Z corr

header

mass flowrate, kg/s

pressure drop, Pa

heat transfer rate, W

ratio of temperature difference between hot and cold

streams in MHE

critical r -value where transverse heat conduction will

not occur

fouling resistance, m2 K/W

Reynolds number

stream property (Cp/ Pr2/3 ), kg.m/s3 K

temperature, K

logarithmic temperature difference, K

volume

stacking pattern and process related property in the new

formulation

heat transfer related fin properties, m2

heat transfer related fin properties, m1

correction factor for Y for different basic units

pressure drop related fin properties, m1

correction factor for Z for different basic units

Greek Symbols

V

P

total exchanger volume, m2 /m3

compactness, ratio of total heat transfer area to total

volume of the fin surface, m1

plate spacing, m

fin effectiveness

two-phase correction factor for heat transfer coefficient

two-phase correction factor for pressure drop

viscosity, kg/ms

density, kg/m3

Superscripts

add

ex

init

pos

Retr

Supply

Targ

volume or heat duty of an existing exchanger

relating to initial temperature of the stream

positive heat duty of an exchanger

retrofitted network related

relating to supply temperature of the stream

targeting values from unit targeting

Subscripts

B

i

i1

i2

j

j1

Blok related

hot stream

hot stream 1

hot stream 2

cold stream

cold stream 1

vol. 26 no. 5 2005

I. BULATOV

j2

P

V

cold stream 2

pressure drop relating

volume relating

REFERENCES

[1] Thonon, B., Compact Heat Exchanger: Technology and Application, Proc International Conference on Compact Heat Exchangers

and Enhancement Technology for the Process Industries, pp. 17.

Banff, Canada, 1823 July, 1999.

[2] Energy Efficiency Best Practice Programme, www.energyefficiency.gov.uk

[3] Hesselgreaves, J. E., Compact Heat Exchangers Selection, Design

and Operation, Pergamon, London, 2001.

[4] Hesselgreaves, J. E., An Approach to Fouling Allowances in the

Design of Compact Heat Exchangers, Applied Thermal Engineering, vol. 22, no. 7, pp. 755762, 2002.

[5] Polley, G. T., Wilson, D. I., Yeap, B. L., and Pugh, S. J., Use

of Crude Oil Fouling Threshold Data in Heat Exchanger Design, Applied Thermal Engineering, vol. 22, no. 7, pp. 763776,

2002.

[6] Manglik, R. M., and Bergles, A. E., Heat Transfer and Pressure Drop Correlations for the Rectangular Offset Strip Fin Compact Heat Exchanger, Experimental Thermal and Fluid Science,

vol. 10, pp. 171180, 1995.

[7] Wang, C.-C., Chi, K.-Y., and Chang, Y.-J., An Experimental

Study of Heat Transfer and Friction Characteristics of Typical Louver Fin-and-Tube Heat Exchangers, International Journal of Heat & Mass Transfer, vol. 41, no. 45, pp. 817822,

1998.

[8] Wang, C.-C., Lin, Y.-T., and Lee, C.-J., An Airside Correlation for

Plain Fin-and-Tube Heat Exchangers in Wet Conditions, International Journal of Heat and Mass Transfer, vol. 43, pp. 18691872,

2000.

[9] Sunden, B., and Svantesson, J., Correlations of j- and f-Factors

for Multilouvered Heat Transfer Surfaces, Heat Transfer, I. Chem.

Symp. Series No. 129, vol. 2, pp. 805811, 1992.

[10] Picon-Nunez, M., Polley, G. T., Torres-Reyes, A., and GallegosMunoz, A., Surface Selection and Design of Plate-Fin Heat Exchangers, Applied Thermal Engineering, vol. 19, pp. 917931,

1999.

[11] Picon-Nunez, M., Polley, G. T., and Medina-Flores, M., Thermal

Design of Multi-Stream Heat Exchangers, Applied Thermal Engineering, vol. 22, pp. 16431660, 2002.

[12] Pua, L.-M., Overall Optimisation Framework for Multi-stream

Plate-Fin Heat Exchanger Network Synthesis, Ph.D. Thesis,

UMIST, Manchester, UK, 2001.

[13] Pua, L.-M., and Zhu, (Frank) X. X., Integrate Heat Exchanger

Network and Equipment Design Using Compact Heat Exchangers, Heat Transfer Engineering, vol. 23, no. 6, pp. 1835,

2002.

[14] Picon-Nunez, M., Use of Compact Heat Exchangers in Integrated

Plants, Ph.D. Thesis, UMIST, Manchester, UK, 1995.

[15] Shokoya, C. B., Retrofit of Heat Exchanger Networks for Debottlenecking and Energy Savings, Ph.D. Thesis, UMIST, Manchester,

UK, 1992.

[16] Silva, M. L., and Zemp, R. J., Retrofit of Pressure Drop Constrained Heat Exchanger Networks, Applied Thermal Engineering, vol. 20, pp. 14691480, 2000.

13

of Practical Heat Exchanger Networks, Ph.D. Thesis, UMIST,

Manchester, UK, 1996.

[18] Nie, X. R., and Zhu, X. X., Heat Exchanger Network Considering

Pressure Drop and Heat Transfer Enhancement, AIChE Journal,

vol. 45, no. 6, pp. 12391254, 1999.

[19] Prasad, B. S. V., Fin Efficiency and Mechanisms of Heat Exchange

through Fins in Multi-stream Plate-fin Heat Exchanger: Formulation, International Journal of Heat Mass Transfer, vol. 39, no. 2,

pp. 419428, 1996.

[20] The General Algebraic Modeling System (GAMS), www.gams.

com

[21] Kays, M. W., and London, A. L., Compact Heat Exchangers, 3rd

ed., McGraw-Hill, New York, 1984.

[22] Drud, A. S., CONOPT: A System for Large Scale Nonlinear Optimization, Reference Manual for CONOPT Subroutine Library,

ARKI Consulting and Development A/S, Bagsvaerd, Denmark,

1996.

APPENDIX

Driving Force Optimization Stage Equations

ri,Hj1, j2,k =

Ti,k T j1,k

;

Ti,k T j2,k

(1)

H

ri1,i2,

j,k =

Ti2,k T j,k

;

Ti1,k T j,k

(2)

(3)

C

ri1,i2,

j,k r crit, j

(4)

DTSi, j,k 1 C

(5)

HD B 5

(6)

Maximize

(c1 DTS B c2 DTRS B c3 HD B )

(7)

where

DTS B =

(Ti,k T j,k )

for interval B

(8)

i, j,k

DTRS B =

(Ti1,k Ti2,k )+

i1,i2,k

(T j1,k T j2,k )

j1, j2,k

for interval B

HD B =

(xi + x j )

(9)

for interval B

(10)

i, j

0.5, c3 = 100, are assigned to ensure the priority of the driving

vol. 26 no. 5 2005

14

I. BULATOV

the header parameter, HD.

Existing volume:

Vi,exj,B = {(1/F TL M,i, j,B ) [(1/i,B i,B ) (1/ h i,B + R f i,B )

+ (1/ j,B j,B ) (1/ h j,B + R f j,B )]} Q i,exj,B

(21)

Pressure drop:

Energy balance for hot streams in each interval:

Hi,B =

Q i,exj,B + Q i,add

j,B

(11)

j=1

H j,B =

Q i,exj,B + Q i,add

j,B

(22)

Z = CRe D

(23)

P j = wP j Z j

(24)

where

(12)

i=1

Pi = wPi Z i

Additional volume:

Vi,add

=

j,B

wv,i, j,B

pos

Q i, j,B

Yi,B

C=

(13)

where

c

a a dh2

D = 3d B

Y = ARe

(14)

Re =

4m

(Afr )

(15)

A=

a a E

dh

(16)

B = 1 b + b

(17)

wv iP

2 (1 + Bg)

wP j =

3

m

(27)

i

3d

3

1 (m/)

(28)

2 (1 + Bg)

m j Bg (m/)

i

Bg wV P

j

Unitadd = Afr AExist

1

fr

(29)

obj = min

where

B V

j

(a E) j

1 (m/)

1

j

hR =

Bg

V

(19)

(a E)i

Bg (m/)

i SP R

i

Vi,add

j

(30)

I. Bulatov is a Research Associate at the Centre for Process Integration SCEAS, the University of Manchester, UK. The described research

was done in the framework Royal Society/NATO

Post-Doctoral Fellowship Programme at DPI. He

has an Engineering degree (M.Sc.) and a Ph.D.

in Chemical Engineering. His research interests

include heat recovery process design, heat exchanger network retrofit, and cost optimization

aspects of process design.

The effect of the stacking pattern on heat transfer coefficient, enhanced area, fin effectiveness, and friction factor (for the latter,

see also Eqs. 2226) can be accounted for by applying correction factors, Bg, (a AE) j /(a AE)i , Yi,corr , and Z i,corr , which the

retrofit procedure also incorporates.

Negative additional load:

pos

Q i, j,B = max 0, Q i,add

(20)

j,B

(26)

where a, a , and c are correlation coefficients, and d is a correlation exponent adopted from [14], which in turn was derived

from the original data published in [21].

The stacking pattern and process-related property are described by Eqs. (2728):

wPi =

[14], which in turn were derived from the original data published

in [21].

wv is a stacking pattern and process-related property. The

stacking pattern and process-related property can be found from

Eqs. 1819:

1

1

Q

wv = (1 + Bg)

1+

(18)

FT Tlm

h R iV S Pi Yi,corr

(25)

## Viel mehr als nur Dokumente.

Entdecken, was Scribd alles zu bieten hat, inklusive Bücher und Hörbücher von großen Verlagen.

Jederzeit kündbar.