Você está na página 1de 6

Pergamon

Computers them. Engng Vol. 19, Suppl., pp. S137S142.1995 Cotwrinht 8 1995 Elsevier Science Ltd Printed in &eat Britain. All rights reserved 009%1354(95)00165-4 0098-1354/95 $9.50 t 0.00

A NEW

DESIGN

METHOD FOR SEGMENTALLY HEAT EXCHANGERS and S. ZAGERMANNt

BAFFLED

M. REPPICH

* F. Ambs GmbH & Co. KG, Postfach 1560, D-79305 Emmendingen. t Rohrleitungs- und Anlagenbau GmbH, Leipziger StraSe 2, D-04439 Engelsdorf.

ABSTRACT The optimization of shell-and-tube heat exchangers from an economic viewpoint gains in importance because of the current and future share of the market of this heat exchanger type in various industries. The developed computer-based design concept assuming from the calculation of optimum tubeside and shellside pressure drops allows to determine the optimum dimensions of segmentally baffled shell-and-tube heat exchangers. As the objective function that is minimized are chosen the total annual costs of the heating or cooling system consisting of the heat exchanger and pumps or compressors. The discussed design method can be used for single-phase liquid or gaseous stream applications. KEYWORDS Shell-and-Tube Heat Exchanger, Optimization, Heat Transfer, Pressure Drop.

INTRODUCTION The market of heat exchangers is increasing continuously because of the necessity of energy saving, their multiple using and the current overcoming of the economic crisis in the chemical industry in Europe. The shell-and-tube heat exchanger is the most commonly used basic heat exchanger type in process industries. In the past few years there were realized several developments to improve the thermal effectiveness of shell-and-tube heat exchangers. New types of tubeside turbulence promoters (e. g. tube inserts, corrugated tubes) and tube supports (e. g. helical baffles) have been successfully introduced. At present several commercial programs are available to design and rate shell-and-tube heat exchangers, however, this tools do not offer any optimization strategies from economic viewpoint. In the field of heat exchanger design methods several works dealed among others with the determination of optimum allowable tubeside and shellside pressure drops from economic viewpoint that are in design algorithms usually treated as constraints. The optimum heat exchanger will be the one which makes full use of both optimum pressure drops. On the basis of simplified expressions describing the interaction between heat transfer and pressure drop are calculated tubeside and shellside pressure drops so that the total annual costs of a heating or cooling system consisting of capital, operating and maintenance costs are minimized. The pressure drop equations published in a recent paper (Reppich, 1994) are now modified to allow the suggestion of geometric dimensions of such heat exchanger that has under given economic and process conditions minimum total annual costs.
s137

S138

European Symposium on Computer Aided Process Engineering-S

PRESSURE DROPS Reppich (1994) published the following simplified equations to determine the tubeside and shellside pressure drop of segmentally baffled shell-and-tube heat exchangers. The tubeside pressure drop is given by dp, = a 1 2 (c, ,,,,,) (2) ; $+W~
.

(1)

The shellside pressure drop can be expressed as

In deriving the relationship for shellside pressure drop by eq. (2) it is assumed that bundle crossflow and baffle window flow aress resp. velocities are equal to achieve the best possible use of available pressure drop in the promotion of heat transfer.

Figure 1. Schematic flow distribution for baffled shellside flow. Based on Tinker (1958), Palen and Taborek (1969). Unfortunately, eq. (2) does not cover the influence of leakage and bypass streams shown in Fig. 1 on shellside heat transfer and pressure drop. In recommended design methods there are introduced correction coefficients td cover the effects of bypass and leakage streams (see VDI-U%rmeatZas, 1994). Considering these correction coefficients eq. (2) was modified as follows

where zl, z2, q and z2 are complex terms corresponding to the coefficients in eq. (2). The correction coefficients (pl and cp2are depending on geometric parameters of apparatus and they are determined according to the calculation method described in VDI-W&meatlas (1994). The pressure drops calculated using eqs. (1) and (3) corresponds very well to the results obtained with mentioned method. The eqs. (1) and (3) are suitable for optimum design of heat exchangers since the only unknown variables they contain are the tubeside and shellside heat transfer coefficients a1 and a2_ The correction coefficients Q~ and cp2are determined iteratively during optimization because of their dependence on the heat exchanger geometry.

European Symposiumon ComputerAided Process Engineering-5

s139

OVERALL

OPTIMIZATION

OF A HEAT

EXCHANGER

SYSTEM

Using of economic criteria allows the optimum design of heat exchanger system consisting of heat exchanger, pumps and/or compressors. As the objective function which is to be minimized is chosen the total annual cost equation of heat exchanger system. The total annual costs include capital, operating and maintenance costs for the heat exchanger and pumps or compressors resp. (Schnell, 1991). Capital costs of any process equipment can be estimated with reasonable accuracy using non-linear correlations based on installed costs containing characteristic equipment parameters. According to the type of process equipment the maintenance costs are usually assessed as a percentage of capital costs. Operating costs can be predicted depending on the power consumption for overcoming the pressure drops. Capital costs on the one hand arise nonrecuringly as purchase costs, operating and maintenance costs on the other hand arise during the overall operating period of equipment. Often are emphasizing the capital costs. however, the operating costs are undervalued or absolutely neglected. Therefore, the total costs of any process equipment over the total equipment life should be considered +s a whole.
OPTIMUM DESIGN OF SHBLL-AND-TUBB HBAT BXCBANGBRS
Process

File Name:

A-S*

Shellside Tubeside Cooling Water Process Water 10.00 90.00 Inlet IV1 60.00 70.00 Outlet 20.00 50.00 Fluid Quantity [kg/s1 1.80 (10--4m'.K/W] 1.80 Fouling Resistance Average Fluid Properties: 994.120 971.650 Density [kg/m-31 0.72325 0.36255 Kinematic Viscosity [10*-6-ma/sl 4178.24 4196.36 Spec. Heat Capacity IJ/(K-kg1 0.6218 0.6665 Therm. Conductivity IW/(m.f)l Design & Construction Data: Orientation (O-Eor./l-Ver.) 0 No. of Shells Parallel/Series 1 / 1 n m Head Type (L, M or S) 2 Number of Tubeside Passes 6000 MaximumTube Length [mm1 2.000 Tube Thickness Dutside Tube Diameter tnml 25.000 Tube Pattern o 30 32.000 Tube Pitch 1"41] tm1 No. of Sealing Strips (Pairs) 0 Therm. Conductivity [W/(m.K)l 48.00 (O-by Progr./I-0pposite2;;de&-Satne Side) 0 Shellside Nozzle Orient. 250 w In/Out Nozzle Sizes: Tubekide 150 w 150 rml Shellside In/Out Requirements: Fluid Name Temperature

Impingement Protection Tubeside/Shellside (O-No/l-Yes) Q/Q Design Pressure Shell tMPag1 ;;yyl 1.60 External 0.00 Young's Modulus Max. Allowable Stress tMPa1 IMPal o.io Poissonls Ratio Tolerance [WI 1.00 Circularity Corrosion Allowance tw1 [*I Joint Bfficiency 0.85 Safety Factor Capital Costs: Heat Exchanger K$:~:xs::: IMI 1' tg;
a+b -Aa

202854 0.30 1.50 1.50 0.~10 0.680 0.680

a+b-(bp-VI* a+b-(Clp-V)n ziyen;

33T20.0 2397.0 2397.0 Rate of Interest [*I 8.00 8.00 8.00

b 1713.60 7.09 7.09

Maintenance Bz:$n-

Heat

costs t,1
0.05 0.05 0.05 70 70

Bacchanger Pun&Compressor 1 Puuq/Caapressor 2

15.0 15.0 15.0

Operating Costs: [DM/kWhl 0.150 Power Costs Cool./Iieat. Fluid Costs [DM/m^31 0.000 Cool./Heat. Fluid Quant.[m^3/hl 0.000 Plant Availability 0.90

Plant Attairnnent[h&r] 8000 Full Load 100 EalfLoad [Z 0 Quarter Load 0 [%I

Figure

2. Input

data required for the heat exchanger optimization program.

The total a.n.nualcosts of heat exchangers are a function of bothside pressure drops, heat transfer area and overall heat transfer coefficient (which is a function of both heat

S140

European Symposium

on Computer Aided Process Engineering-5

transfer coefficients).

Using the basic heat exchanger &=k.A.AT,

design equation (4)

as a funcand eqs. (1) and (3) we will get the objective function in the final formulation tion of two independent variables - the tubeside and shellside heat transfer coefficient:

The non-linear objective function (5) and its first and second partial derivatives are continuous - for exact formulation of the total annual cost equation see Reppich (1993). According to the necessary conditions for an extremum existence of the objective function

the optimum heat transfer coefficients can be obtained by solution of an equation system The optimum tubeside and shellside pressure drops as well using numerical methods. as the corresponding flow velocities are calculated on the basis of process data, fluid properties, design parameters and economic data. Basic input data needed for the optimization program are shown in Fig. 2. Required geometric informations of the heat exchanger to be specified by the user are: heat exchanger type, maximum allowable tube length, tube diameter, tube wall thickness, tube pitch, tube layout, number of tube passes, number of sealing strips, nozzle diameters (see Fig. 2). Some of these parameters By using the Newton-Raphson method are constraints for the optimization problem. for solving the equation system (6) and by analysing the second order derivatives of the objective function (5) the momentary optimum heat transfer coefficients are obtained. From eqs. (1) and (3) then can be calculated the corresponding momentary optimum pressure drops. On the basis of optimum pressure drops are finally determined such gee metric dimensions of the heat exchanger that entail momentary optimum flow velocities. These six optimized design parameters of apparatus are: number of tubes, tube length, shell diameter, number of baffles, baffle cut, baffle spacing.
SOLUTION OF RQUATION SYSTRM AND RRROR ANALYSIS: 7.218B+0003 fl= 8.883B-0012 x[2]= .3.429B+0003 f2= -2.736B-0010

x[11-

ml I 7218.06 W/(ma-lC)
Y' x

OPTIMUM

PROCBSS DATA:

Nul= 227.4

= 3420.19

W/(ma-K)

Nu2= 137.9
.I

1.23 m/s 6.75 ( 8.59) kPa

Rel= Prl= 70908 2.22 No. of Shells: 1


A: Q:
Par./

19.83 0.68 m/s (20.521 kPa


1 Ser.

Re2= Pr29 23391 4.03


horizontal

Type: Fixed tubesheet

Orientation:

mat Transfer Surface per Shell Heat Exchanged

95.98 m*

Corrected Mean Temperature Difference at: Overall Heat hanefer Coefficient k:


OPTIMUM

4196.36 kW 39.19 R 1115.74 W/(m-K)

GEOMRTRIC DETAILS: 557.7 mm No. of Tubeside Passes: 2 Outer Tube Diameter: 571.6 ua~ Inside Shell Diameter: 5050 lmn No. of Tubes: 242 (25.0 x 2.0 / A 30) Tube Length: Cut: 27% Spacing c/c: 228.0 mp No. of Baffles: 20 Inlet: 308.7 mn
COST ANALYSIS: Capital Costs: Operating Costs: Total Costs:

20053.91 M/year 3257.06 DM/year 23311.77 DUlyear Figure 3. Report

HRATRX~RRACCORDINGTO No. of Tubes: 250 Outer Tube Diameter: Inside Shell Diamater:

CODRS 575.5 rmn 508.0 mm

of optimization

results.

The whole optimization problem - consisting of the above mentioned solution of equation system (6), the analysis of second order derivatives and the determination of geometric

European Symposium on Computer Aided Process Engineering-5


dimensions

s141

that results from calculated pressure drops - is solved as an iterative procedure that is continued until the absolute values of variable changes are less than a specified tolerance:

1,; - a;-lI&,

j=

1,2

where i is the loop number. In each loop are taken into consideration the defined constraints (maximum allowable tube length, maximum shell diameter provided by standards etc.) and, if necessary, the number of shells in parallel or series is changed. In Fig. 3 is summarized a brief result report of the problem defined by the input data from Fig. 2. Detailed results including all informations about the optimized heat exchanger are saved in a text file. Table 1 contains the results of a case study analysing the influence of changed operating and capital costs on the total costs and the dimensions of heat exchanger. In all cases is transferred the same heat rate between two liquid streams s = 4196.3 kW. The input and output temperatures, flow rates, fouling resistances and physical properties of the two streams are shown in Fig. 2. It is assumed always the same effective mean temperature difference AT,,, = 39.19 K for two tube passes. The basic case 1 in Tab. 1 contains the results of problem defined in Fig. 2 with the given capital and energy costs. In second case are assumed as opposed to case 1 energy costs with 0.25 DM/h. The optimized heat exchanger has because of higher energy costs lower pressure drops in comparison with first case (energy costs in case 1 are 0.15 D&f/h). The overall heat transfer coefficient in second case is because of lower flow velocities lower too by which the heat transfer area increases about 5 percent.
CASE 1 tubeside Heat transfer coefficient ]w/(mK)l Pressure drop 7218 shellside 3428 CASE 2 tubeside 6297 shellside 3136 CASE 3 tubeside 8770 shellside 3957

[ kPa]

8.6 1.23 1116

20.5 0.68

6.0 1.03 1055

13.1 0.58

14.7 1.56 1207

37.3 0.86

Flow velocity [m/s] Overall heat transfer coefficient ]W/(m2K)l Number of tubes Shell diameter [mm] Tube length (mm] Number of baffles
II

242 572 5050 20


I

287 618 4500 16


I

190 511 5945 28 27

Baffle cut

[ %]
II

27

28

] Heat transfer area I m2 I Capital costs [ DM/yr] Operating costs [ DM/yr] Total costs [ DM/yr]

95.9
20054 3258 23312

101.5
20600 3619 24219

88.7
37532 5667 43199

Table 1. Results

of optimization.

In the third case are assumed the same energy costs as in first case, however, doubled capital costs, e. g. by using of more expensive construction materials. The optimized heat exchanger has in comparison with the basic case 1 a reduced heat transfer area about approximately 8 percent. A reduced heat transfer area is only possible due to a higher overall heat transfer coefficient. Therefore, the flow velocities and the heat transfer coefficients increase. The higher pressure drops are caused by higher flow velocities. Table 1 contains for each case the following optimized design details: tube length, tube number, shell diameter, number of baffles and baffle cut.

S142

European Symposium on Computer Aided Process Engineering-S

CONCLUSIONS
On the basis of optimum pressure drops and resulting flow velocities the main dimensions of shell-and-tube heat exchangers. The derived sign method allows to show the interaction between process data, and geometric parameters of heat exchangers. With the new design to find the optimum technical solution from an economic viewpoint transfer applications. REFERENCES Palen, J.W. and J. Taborek (1969). Solution of shellside flow pressure drop and heat transfer by stream analysis method. Chem. Eng. Prog. Symp. Ser., Vol. 65, No. 92. Reppich, M. (1993). Optimale Auslegung von Rohrbiindelwkmeiibertragern segmenten. PhD Thesis, Technical University of Brno. Reppich, M. (1994). Optimale Auslegung von Rohrbiindelwarmeiibertragern. meaustauscher (H. Schnell and B. Thier, Eds.), 2nd Ed., Vulkan-Verlag, mit Umlenk. In: Essen. WC+there are optimized computer-based deeconomic conditions concept it is possible for single-phase heat

Schnell, H. (1991). Technisch-wirtschaftliche Optimierung von Wkmeaustauschern. In: Wirmeoustauscher (H. Schnell and B. Thier, Eds.), 1st Ed., Vulkan-Verlag, Essen. Tinker, T. (1958). J. Heat Transfer, Vol. 80, pp. 36-52. Dusseldorf.

VDI- Wtirmeatlas (1994). NOMENCLATURE A a, 4 cl


CT d i

7th Ed., VDI-Verlag,

b-l21

k kL/,kgf,k~,k~,k1,kz,ks ml, m2, w22, nl


PT

P/(m2K)1

AP ci AT,,, i,
2J1,22r =1,-Q a & x v

Pal WI

WI
bVs1

W(m2K)1
W(m VI

P
'pl,PZ

b2/sl k/m31

heat transfer area coefficients total annual costs tube diameter iteration number overall heat transfer coefficient coefficients coefficients Prandtl number pressure drop total heat exchanged effective mean temperature difference volumetric flow rate coefficients heat transfer coefficient stop criterion thermal conductivity kinematic viscosity density correction coefficients

Você também pode gostar