Efficient Retrofitting Approach for Improving Heat Recovery in Heat

Subscriber access provided by Old Dominion University Libraries & VIVA (Virtual Library of Virginia)

Article

An efficient retrofitting approach for improving heat recovery in heat exchanger networks with heat transfer intensification Ming Pan, Igor Bulatov, and Robin Smith Ind. Eng. Chem. Res., Just Accepted Manuscript • DOI: 10.1021/ie501202w • Publication Date (Web): 05 Jun 2014 Downloaded from http://pubs.acs.org on June 24, 2014

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Industrial & Engineering Chemistry Research is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Industrial & Engineering Chemistry Research

An Efficient Retrofitting Approach for Improving Heat Recovery in Heat Exchanger Networks with Heat Transfer Intensification Ming Pan*, Igor Bulatov, Robin Smith Centre for Process Integration, School of Chemical Engineering and Analytical Science The University of Manchester, Manchester, M13 9PL, United Kingdom ABSTRACT In this paper, an efficient design method is presented to implement heat transfer intensification technologies for increasing energy saving in industrial scale heat exchanger networks (HENs). First of all, the detailed models of shell-and-tube heat exchangers are utilized to estimate exchanger performances under normal and intensified conditions, and an optimization method based on simulated annealing is proposed to find appropriate retrofitting options in the existing HENs. Then, a novel retrofit approach based on the exchanger models and optimization method is introduced to retrofit HENs with several possible retrofit strategies, including exchanger relocations, heat transfer intensification, additional area, and stream repiping. Moreover, a new efficient software package (int-HEAT®) is developed for the addressed retrofit problems. The interface of int-HEAT® is designed to focus on a clear presentation of HEN retrofit procedure, which allows users to easily control all the retrofit stages. An industrial problem of preheat train for crude oil distillation is solved with the use of the new developed software through all the main steps of the proposed retrofit procedure, namely reducing energy consumption, identifying retrofitted exchangers, selecting retrofit strategies, and implementing heat transfer intensification techniques. The case study demonstrates the validity and efficiency of the developed software tool and the proposed approach.

Keywords: Heat exchanger networks (HENs); Retrofit; Heat transfer intensification; Heat recovery; Software tool

1 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 43

1. INTRODUCTION Retrofitting heat exchanger networks (HENs) has been widely studied in industrial and academic research. Improving energy saving is crucial to the modern process industries, not only due to the increasing cost of energy consumption but also because of the growing social concern about the sustainability of current industrial development. Commonly, the conventional techniques for HENs retrofit include implementing heat transfer intensification, adding exchanger area, installing new exchangers, and reconfiguring heat recovery structure (e.g. repiping). The existing methods for retrofitting HENs can be divided into three categories, thermodynamic analysis, mathematical programming, and the combination of both. Regarding the thermodynamic analysis of process, Tjoe and Linnhoff 1 proposed two steps for HEN retrofit based on Pinch Analysis, where retrofit targeting and modifications were determined. Wang et al. 2 proposed five heuristic rules to identify suitable heat exchangers for intensification and considered detailed exchanger performances during the proposed retrofit procedure. Rather than giving a qualitative indication of how much heat would possibly be recovered, the researchers working on mathematical programming methods tried to give a quantitative answer to retrofitting costs. Ciric and Floudas 3 proposed a two-stage optimization approach for HEN retrofit, where a mixed-integer linear programming (MILP) model was used to determine a superstructure in the first stage, while a nonlinear programming (NLP) problem was solved for finding a detailed solution in the second stage. Sorsak and Kravanja

4

proposed an MINLP optimization model for HEN retrofit, which

considered the selection of different exchanger types simultaneously, such as, double pipe exchangers (DP), shell and tube exchangers (ST), and plate and frame exchangers (PF). Rezaei and Shafiei 5 used genetic algorithms (GA) coupled with NLP and integer linear programming (ILP) to address network structure modifications, maximum energy recovery, 2 ACS Paragon Plus Environment

Page 3 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


and minimum investment cost, respectively. To combine mathematical programming methods with the thermodynamic analysis, Kovac Kralj and Glavibvc 6 used Pinch Analysis to identify many alternative retrofit designs, then optimized material and energy flow rates sequentially to obtain a superstructure by using an ASPEN PLUS simulator based on the given energy and material bounds, and finally obtained the optimal solution, including flashes and compressors with mixed-integer nonlinear programming (MINLP) or NLP algorithms. Most early works considered network topology modifications without heat transfer intensification. However, using intensification techniques has practical advantages in HEN retrofit, as intensification can avoid physical modification of the exchanger itself. The implementation of intensification techniques is a relative simple task that can be easily achieved within a normal maintenance period when production losses can be kept in a minimum level and the relevant civil works can be also reduced. Polley et al.

7

applied

intensification techniques into the retrofit of heat exchanger networks, and analyzed the aspects of fouling and pressure drop, but only the potential for using intensification techniques in the heat recovery systems was discussed without providing systematic procedures for implementing intensification in the retrofitting of heat recovery network. Zhu et al.

8

presented a methodology for applying intensification in HEN retrofit based on

network pinch approach. The proposed procedure aimed to find out which exchanger should be intensified and the augmentation level of intensification. However, the intensified devices were used only for additional area identified from the network pinch method, without using full benefits of intensified heat transfer. Recently, some new approaches of implementing intensification technologies have been reported to provide realistic and practical solutions for debottlenecking of HENs. This leads to significant energy saving without many structural modifications of heat recovery system configuration. Pan et al. proposed a novel MILP-based method to solve small scale HEN 3 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

retrofit problems

9, 10

and the large scale problems

11, 12

Page 4 of 43

with intensified heat transfer

techniques. The combination of different types of intensification techniques in suitable exchangers has been also addressed to increase network energy recovery with very low retrofit cost.13 To take the advantages of intensification and conventional retrofit methods, Pan et al. successfully developed an optimization framework to facilitate the automated design of HEN retrofit with rigorous consideration of the conventional topology modification strategies

14-16

, and validated potential structures for retrofitting to achieve

significant energy saving without the expensive cost from too many modifications.

17

Moreover, intensified heat transfer techniques are also considered to reduce fouling effect in enhanced exchangers.18, 19 In those work, the new retrofit approach can provide realistic and practical solutions for industrial HEN retrofit problems as detailed performances of tube inserts (heat transfer enhancement and fouling mitigation) were systematically considered. This leaded to substantial capital saving not only due to the significant energy reduction under low retrofit costs, but also due to the less production loss related to the increase in exchanger operating time. As discussed above, most of the reported methods are still in the research stage and far from being used in practical application, as they assume constant heat transfer coefficients and neglect fluid pressure drops, which is not suitable for practical application and retrofit, because heat transfer coefficients and fluid pressure drops vary with design. Even a few have been programmed in some commercial software, none of them cover comprehensive HEN retrofit problems involving heat transfer intensification, exchanger reconfiguration, new exchanger installation, and network topology modification, specially for industrial application. In this paper, a new software package (int-HEAT®) and retrofit method is proposed to improve heat recovery in existing HENs with the above comprehensive strategies. The new retrofit approach addresses exchanger detailed performances (heat transfer coefficient and pressure drop) related to exchanger geometry, identifies suitable 4 ACS Paragon Plus Environment

Page 5 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


exchangers for intensification, and finds out the augmentation level of intensification in retrofitted HENs. Moreover, the developed software program can help users easily to achieve significant energy saving in existing HENs through all the main steps of the proposed retrofit procedure. The use of int-HEAT® in an industrial case demonstrates the validity and efficiency of the proposed method and software. The approach of retrofitting HENs with heat transfer intensification addresses the improvement of heat transfer in intensified exchangers whose heat transfer coefficients can increase several times greater with implementing intensification devices or changing exchanger geometry. Since the new retrofit method considers the exchanger performances varying with the changes of exchange geometry, stream properties and intensification devices, it can avoid the impracticality of the existing retrofit approaches which fix the heat transfer coefficients for all exchangers even some of them need to be changed after retrofit. In order to present the new retrofit approach clearly, the rest of the paper is structured as follows. The models used to predict exchanger performance for HEN retrofitting are briefly introduced in Section 2, followed by the descriptions of the optimization method for finding appropriate retrofitted HENs in Sections 3. Based on the methods of exchanger modelling and HEN optimization, a new software tool (int-HEAT®) and a HEN retrofit approach are developed in Section 4. In Sections 5 and 6, an industrial study is carried out to demonstrate the validity and efficiency of the proposed approach.

2. MODELS FOR PREDICTING EXCHANGER PERFORMANCE

The HEN retrofit presented in this paper considers the characteristics of real-world engineering problems, such as variable heat transfer coefficients, pressure drop constraints, varying stream thermal properties and detailed geometry of heat exchangers. Thus, it is necessary to know the performance change of each exchanger in a HEN after the retrofit. 5 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 43

2.1 Model of heat exchanger Shell-and-tube is the most widely used type of heat exchangers in the process industries. Thus the heat recovery in the paper is based on heat transfer through shell-and-tube exchangers. The heat transfer coefficient and pressure drop of each individual heat exchanger in a network are calculated by the model proposed by Wang et al. 2, which has been demonstrated to be reliable for estimating shell-and-tube exchanger performances. Figure 1 illustrates the geometric details of shell-and-tube heat exchangers and the definition of various parameters required in the calculations of heat transfer and pressure drops. The fully detailed models can be found in the work reported by Wang et al. 2. Eqs 1-9 show the main equations to calculate the heat transfer and pressure drops in tube side and shell side.

Tube-side heat transfer coefficient (hi): hi = (k i / Di ) × Nu i

(1)

where hi is tube-side heat transfer coefficient, ki is tube-side fluid thermal conductivity, Di is tube inner diameter, and Nui is tube-side Nusselt number.

Tube-side pressure drop (△Pi):

n p f i Lρi vi2 ∆Pfi = 2 g c Di

(2)

0.5α r ρ i vi2 gc

(3)

∆Pi = ∆Pfi + ∆Pr + ∆Pni

(4)

∆Pr =

where △Pfi is tube-side pressure drop due to friction loss, np is the number of tube passes, fi is Darcy friction factor, L is exchanger length, ρi is tube-side fluid density, vi is tube-side fluid velocity, gc is unit conversion factor which is equal to 1.0 kg·m/(N·s2), △Pr is tube-side 6 ACS Paragon Plus Environment

Page 7 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


pressure drop related to the tube entrance, exit and return losses, αr is the number of velocity heads allocated for minor losses, and △Pni is pressure drop in tube-side nozzles.

Shell-side heat transfer coefficient (h0):

Fs Fp FL k02 / 3 (C p 0 µ0 ) h0 = D0

1/ 3

(5)

where h0 is shell-side heat transfer coefficient, Fz, Fp and FL are the factors to calculate shell-side heat transfer coefficient, k0 is shell-side fluid thermal conductivity, Cp0 is shell-side fluid specific heat capacity, µ0 is shell-side fluid viscosity, and D0 is tube outer diameter.

Shell-side pressure drop (△P0):

∆Pfb , 20% Bc =

f 0 Ds ρ 0 v 2p 0

(6)

2 g c De

∆Pf 0, 20% Bc = (nb − 1)∆Pfb , 20% Bc + Rs ∆Pfb , 20% Bc

(7)

∆Pf 0 = ∆Pf 0, 20% Bc (Bc / 20% )

(8)

∆P0 = ∆Pf 0 + ∆Pn 0

(9)

np0

where △Pfb,20%Bc is shell-side pressure drop in one central baffle spacing when baffle cut is 20%, f0 is shell-side friction factor, De is equivalent diameter, vp0 is of shell-side fluid velocity, △Pf0,20%Bc is shell-side pressure drop in straight section of shell with 20% baffle cut, nb is the number of baffles, Rs is the correction factor for unequal baffle spacing,

△Pf0 is

shell-side pressure drop in straight section, and △Pn0 is pressure drop in shell-side nozzles.

2.2 Models of heat transfer intensification



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 43

Heat transfer intensification is the study of enhancing or improving heat transfer performance, which generally means an increase of heat transfer coefficient. Thermal and hydraulic calculations on heat transfer intensification can provide a better understanding of the addressed intensification techniques, and enable designers to identify suitable intensification techniques to determine the best economic trade-offs between energy saving and investment.20 Conventional intensification techniques considered in HEN retrofit problems include tube-side intensification (internal tube fins, twisted-tape inserts, coiled-wire inserts, and hiTRAN®), and shell-side intensification (external tube fins, and helical baffles). The functions and calculations of the aforementioned techniques are briefly introduced in this section. More details of intensification modelling have been addressed in the work presented by Pan et al. 20

Internal tube fins: Internally finned tube is one method that has been used for passive heat transfer intensification. In general, internal fins increase the heat transfer area in the tube side but do not affect the development of the flow pattern. Figure 2(a) describes the geometric variables of internal fins: fin height (e), helix angle (α), fin pitch (p), the number of fins (Nf) and average fin width (t). Eq 10 presents the general form of the correlation that predicts the enhancement ratio of the Nusselt number in turbulent flow (Nuif,t), where lcsw is modified characteristic length for swirling flows, and func(geometry) is the geometry function of internal fins for calculating Nusselt number .21

l =  csw Nu  Di

Nu if ,t

   

−1

2

  0.25πDi2   2  0.25πDi − N f ⋅ e ⋅ t   

0.8

func( geometry )

(10)


Page 9 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


The fanning friction factor of internal fins in turbulent flow (fif,t) has been also correlated the pressure drop data into a Blasius form:21

l =  csw f  Di

f if ,t

   

−1.25

  0.25πDi2   2  0.25πDi − N f ⋅ e ⋅ t   

1.75

(11)

Twisted-tape inserts: Twisted-tape inserts are swirl-flow devices that create rotating or secondary flow along the tube length. They consist of a thin strip of twisted metal with usually the same width as the tube inner diameter. The geometric features characterizing this type of insert are twist pitch (H), tape thickness (δ), and tube inside diameter (Di), as shown in Figure 2(b). Eq 12 presents the Nusselt number of twisted-tape inserts in turbulent flow (Nut,t): 22 0.8

Nu t ,t = 0.023 Re 0.8

2δ     π  π + 2 − Di    Pr 0.4  4δ   4δ  π −   π− Di   Di 

     

0.2

(1 + 0.769 / y ) × φ

(12)

where y is dimensionless twist ratio, and φ is the correction factor for fluid property variation.

The correlation to predict the fanning friction factor of twisted-tape inserts in the turbulent region (ft,t) is shown in eq 13. 22

f t ,t

   π  0.0791   = 0.25   δ 4 Re  π −  Di  

1.75

2δ  π + 2− Di   4δ  π− Di 

     

1.25

 2.752  1 + 1.29  y  

(13)



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 43

Coiled-wire inserts: Coiled-wire inserts are tube inserts consisting of a helical coiled spring can induce a swirl effect and hasten the transition from laminar to turbulent flow. The geometry of helically coiled-wire is characterized by helical pitch (p), wire diameter (w) and inserts thickness (t), as described in Figure 2(c). Eqs 14 and 15 describe the Nusselt number (Nuc,t) and fanning friction factor (fc,t) for coiled-wire inserts within turbulent region.23

 p Nu c ,t = 0.132  Di

f c ,t

 p = 9.35   w

   

−0.372

Re 0.72 Pr 0.37

(14)

−1.16

Re −0.217

(15)

hiTRAN®: Tube insert, hiTRAN® tube inserts (produced by Cal Gavin Ltd) consist of a unique wire frame matrix with different densities, are used to improve tube-side heat transfer coefficient for the laminar region, as shown in Figure 2(d). In the laminar region, the fluid nearest the wall is subjected to frictional drag, which has the effect of slowing down the fluid at the wall. This laminar boundary layer can significantly reduce the tube side heat transfer coefficient and consequently, the performance of the heat exchanger. By using hiTRAN® in the tube, the laminar boundary layer will be disrupted, which creates additional fluid shear and mixing, thus minimizing the effects of frictional drag. Heat transfer coefficient and pressure drop of hiTRAN® can be obtained from software hiTRAN.SP® (supported by Cal Gavin Ltd).24

External tube fins


Page 11 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


The externally finned tube used in shell-and-tube heat exchangers is often referred to as radial low-fin tube. It is made by an extrusion process in which the wall of a plain tube is depressed by stacks of disks. The undisturbed portion of the tube wall between the disks forms the fins. There are three different tube diameters which are arranged inside-to-outside, finned tube inner diameter (Di), root tube diameter (Dr), and finned tube outer diameter (Do). Figure 3(a) presents the schematic diagram of external fins, where τ is fin thickness, and b is fin height. Shell-side heat transfer coefficient of external tube fins (hs,ef) can be determined as:25

hs ,ef = jH × Prs

1/ 3

(k

s

/ De ,s )

 B   0.08 Re s0.6821 + 0.7 Re s0.1772 jH = 0.51 + D s  

(

(16)

) (17)

where, Prs is shell-side Prandtl number, ks is shell-side thermal conductivity, B is baffles spacing, and Ds is shell side diameter.

Shell side friction factor of external tube fins (fs,ef) is:25

f s ,ef = 144[ f1 − 1.25(1 − B / Ds )( f1 − f 2 )]

(18)

Helical baffles: Helical baffles are usually made by joining four elliptical sector-shaped plates in succession and then arranging them in a pseudo-helical manner. Each baffle occupies one-quarter of the cross section of the heat exchanger, and is angled to the axis of the heat exchanger. Figure 3(b) shows the schematic diagram of helical baffles, where βs is helical angle, and B is helical baffle spacing. According to Zhang et al., 26 the shell-side heat transfer 11 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 43

coefficient of helical baffles (hs,hb) depends on the value of the associated Nusselt number (Nus):

hs ,hb = (k s / D0 )Nu s

(19)

where, ks is shell-side fluid thermal conductivity, and D0 is tube outer diameter.

To calculate the shell-side friction factor, a correlation has been proposed to predict shell side friction factor of helical baffles (fs,hb), based on the shell-side fluid Reynolds number (Res).26

D

f s ,hb = C fs Re s fs

(20)

where Cfs and Dfs are the parameters for calculating friction factor of helical baffles.

3. SIMULATED ANNEALING OPTIMIZATION METHOD FOR HEN RETROFIT PROBLEMS

Stochastic methods have been used widely for large industrial optimization problems over the past decades. Simulated annealing (SA) and genetic algorithms (GA) are often viewed as quite separate, competing paradigms in the field of modern heuristics. SA is a generic probabilistic metaheuristic for the global optimization problem of locating a good approximation to the global optimum of a given function in a large search space. It is often used when the search space is discrete. The goal of SA is to bring a system, from an arbitrary initial state, to a state with the minimum possible energy. While, GA is a search heuristic that mimics the process of natural selection, which generates solutions to optimization problems using techniques inspired by natural evolution, such as inheritance, mutation, selection, and 12 ACS Paragon Plus Environment

Page 13 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


crossover. In the theoretical comparison of using SA and GA for retrofitting HENs, the key difference between SA and GA is that: SA creates a new solution by modifying only one solution with a local move (the moves can be described as moving exchangers, intensifying heat transfer, adding area, and installing new exchangers), and a good solution is searched with these random moves; but GA creates solutions by combining two existing good solutions (including all retrofit options), making the assumption that such combination (children) meaningfully share the properties of their parents, so that a child of two good solutions is probably better than a random solution. However, it is very common that two initial good solutions are not easily to be found for a particular industrial problem, and then GA will not provide an advantage over SA. Moreover, the random moves in SA can be used to control retrofit options directly, showing its practical advantages in HEN retrofit. Thus, SA optimization method is adopted in this paper, which is mainly based on the work proposed by Wang et al.27

3.1 Simulated annealing algorithm The basic iteration of SA algorithm includes: at each step, the SA heuristic considers some neighbouring state of the current state, and probabilistically decides between moving the system to neighbouring state or staying in the current state. These probabilities ultimately lead the system to move to states of lower energy. Typically this step is repeated until the system reaches a state that is good enough for the application, or until a given computation budget has been exhausted. Figure 4 presents the SA algorithm proposed in the work of Wang et al.27 The algorithm begins with an initial annealing temperature obtained from an estimated trial solution, then a new solution is created with a random change and confirmed to be better than the current solution repeatedly until the annealing temperature is met. Once the best solution in the current annealing temperature condition is determined, the annealing temperature is reduced, and the procedure of finding a better solution is executed 13 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 43

again. Finally, when the annealing temperature is close to the final annealing temperature, the whole optimization progress can stop, and a final optimal solution is provided. In SA optimization, initial annealing temperature and Markov chain length are used to control the accuracy and time of calculation, and a cooling schedule has been proposed to control the temperature cooling speed, which can be found in details in work of Kirkpatrick et al.28

3.2 Simulated annealing moves Several typical retrofit strategies, including exchanger relocations, heat transfer intensification, additional area, new exchanger installation, and stream repiping, are considered for HEN retrofitting in this paper. These strategies are described as simulated annealing moves in the SA optimization procedure, where each move generates a network with a small random difference from the current network. The probability of each move can be different according to the influence on the performance of the network. However, the assignment of probabilities to moves is highly problem specific. In the proposed optimization approach, a heat exchanger network is represented by unique nodes on each stream, and linear equation-based models are solved simultaneously to calculate node temperatures. The proposed optimization procedure addresses the simulated annealing moves, which include continuous moves (e.g. modifying heat duties, changing heat transfer coefficients, and changing splitting flow fractions), and discrete structural moves (e.g. adding heat exchangers, deleting heat exchangers, adding bypasses, adding stream splits, deleting bypasses, deleting stream splits, resequencing exchangers and repiping exchangers). It is noted that the optimization search space is related to the considered simulated annealing moves. Running all the addressed SA moves might lead to significant search space increase, and the computational time of the optimization approach will be too long to find an optimal solution. This does not normally happen in practice, as it increases the 14 ACS Paragon Plus Environment

Page 15 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


difficulties of implementing retrofit in HEN, which are commonly restricted by topology, safety and maintenance constraints, and requires high capital costs for civil works. In industrial applications, retrofitting problems are usually specified, and the proposed SA optimization approach can identify optimal solutions under different SA move specifications.

3.3 Objective function The proposed SA optimization approach aims to find an optimal solution for a particular retrofit scenario where certain energy saving or net present value can be achieved. To calculate retrofit profit, the cost of implementing retrofit strategies and the profit of reducing energy consumption should be determined together. Firstly, eq 21 presents the energy saving (QS) achieved in the retrofitted HEN, where EXhu and EXcu are the set of all hot and cold utility exchangers, HFCPex and CFCPex are heat-flow capacities (the multiplication between heat capacity and flow-rate) of hot stream and cold stream in exchanger ex, and CTI’ex/CTIex and HTI’ex/HTIex are the initial/new inlet temperatures of cold stream and hot stream in exchanger ex before/after retrofit.

QS =

∑ [CFCP × (CTI ′ ex

ex∈EX hu

ex

− CTI ex )] +

∑ [HFCP × (HTI ex

ex

− HTI ex′ )]

(21)

ex∈EX cu

Consequently, the profit of energy saving (PES) can be computed as follows:

  PES = YR× CCU × ∑[CFCPex × (CTI ex′ − CTIex )] + CHU× ∑[HFCPex × (HTIex − HTI ex′ )] ex∈EXhu ex∈EXcu  

(22)

where CCU is a yearly cost parameter per cold-utility-duty unit, HCU is a yearly cost parameter per hot-utility-duty unit, and YR is the expected project lifetime. 15 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 43

Based on the cost of implementing retrofit strategies (COSTR) and the profit of energy saving (PES), the retrofit profit before tax accumulated over the expected lifetime (RP) of the retrofitted HEN is obtained:

RP = PES − COSTR

(23)

Two objectives are considered in this paper: the minimization of utility consumption (i.e. maximum energy saving), and the maximization of retrofit profit. In the proposed SA optimization procedure, a new network can be accepted if its objective value improves and meets the acceptance criterions in a simulated annealing move. After a numbers of moves, an optimal network is achieved with sufficient iterations.

4. NEW SOFTWARE PACKAGE AND RETROFIT METHOD

In this section, the detailed exchanger models and optimization approach (proposed in Sections 2 and 3) are utilized to developed a new software package, called as int-HEAT®, which can be used to improve the performances of heat recovery systems in the process industries. Based on the proposed approach and software tool, a systematic retrofit method is then presented, including identifying bottlenecked parts in heat exchanger networks, screening different retrofit options, evaluating impacts of heat transfer intensification, and providing the most appropriate use of heat transfer enhancement in the context of heat exchanger networks.


Page 17 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


4.1 A new software tool (int-HEAT®) int-HEAT® can be used to optimize the choice of utilities for an individual process. Once options have been explored using targets, int-HEAT® can design the appropriate heat exchanger network design automatically for the choice of utilities and level of heat recovery made. Both retrofit and new design can be carried out interactively or automatically. If design is carried out automatically, the designer maintains control over network complexity. The characteristics of the developed software tool are shown as follows:

• Energy targets: int-HEAT® sets energy targets and optimizes the selection of utilities for individual processes. The tools include the composite curves, the grand composite curve, and the problem table. These tools allow the designer to predict hot and cold utility targets for individual processes. int-HEAT® automatically places the optimal mix of utilities against the grand composite curve. In addition to providing energy-based targets for the process, the program can also target for the surface area of the heat exchangers and the minimum number of heat exchanger units and shells. Combining these targets allows total cost targets to be predicted ahead of design. • Interactive design: An interactive design grid allows the process designer to manipulate the design of existing and new heat exchanger networks manually. • Automatic design of new heat exchanger networks: New design can be carried out automatically. To avoid automatic design creating structures that are over-complex, int-HEAT® allows the designer to keep control over the resulting network complexity. • Automatic retrofit of existing heat exchanger networks: int-HEAT® offers flexibility in the approach to heat exchanger network retrofit. Once the approach identifies bottlenecks in the existing network that limit energy recovery, structural



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 43

changes can then be made to overcome the bottlenecks. The structural changes made are resequencing (change of location of an existing exchanger on the same streams), repiping (change of location allowing the streams to change), adding a new exchanger or introducing a stream split. int-HEAT® finds the best structural modifications, but leaving the designer under control of the number of modifications. An alternative approach allows multiple modifications to be identified according to the retrofit economic criterion. • Simulation: For a given network structure, int-HEAT® will calculate the intermediate network temperatures and heat exchanger performances using simple heat exchanger models. The software has different simulation modes that dependent on the data specified and the options selected. Moreover, exchanger performances (such as heat transfer coefficients and pressure drops) can be predicted based on detailed heat exchanger models if sufficient exchanger geometry details are provided. • Optimization: int-HEAT® can automatically adjust the degrees of freedom in the network to achieve minimum total annualized cost for both retrofit and new design. The optimization reduces the network cost by trading off the utility cost against capital cost.

4.2 A new design method for retrofitting HENs

The objective of the new retrofit method is to find an optimal solution to achieve energy recovery in existing HENs with suitable retrofit strategies. Based on the overview of the proposed models and optimization approach in Sections 2 and 3, the new retrofit procedure for improving heat recovery in existing HENs can be described as:


Page 19 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Step 1: Retrofit an existing HEN based on the proposed SA optimization approach to reduce its total energy consumption, which usually includes network structural changes, new exchanger installations, stream repiping, and new stream splitting. Step 2: Identify suitable exchangers for retrofit based on the optimal solution obtained in Step 1 (e.g. if the required heat transfer coefficient of an exchanger is larger than its designed heat transfer coefficient in the retrofitted solution, this exchanger is considered to be the candidate for retrofit). Step 3: Identify retrofit strategies (intensification, additional area, or both) in the selected exchangers (the trade-off between the costs of implementing intensification technologies and adding exchanger area is considered). Step 4: Evaluate detailed performances (heat transfer coefficients and pressure drops) of the selected exchangers based on the given exchanger geometry information and the exchanger detailed models. Step 5: Select suitable intensification techniques (subject to maximum pressure drop allowances and the technology costs) on the enhanced exchangers to achieve substantial economic and environmental benefits.

5. AN INDUSTRIAL PROBLEM

The problem addressed in this paper is an existing preheat train for a crude oil distillation column in a refinery plant. Figure 5 describes the network structure, which includes 31 heat exchangers and 14 processing streams (3 cold streams and 11 hot streams). The stream data and initial exchanger data can be found in Table 1 and Table 2 (the comprehensive data of HEN can be found in Table S1 in the SUPPORTING INFORMATION for interested readers). The minimum approach temperature (∆Tmin) before and after heat transfer



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 43

intensification are 19 ºC and 5 ºC. Different to the cases reported by Pan et al.,11-13, this case considers both the designed and required heat transfer coefficients of exchangers, which is very common in a practical situation as the heat transfer coefficient of an exchanger (DUex) is usually designed to be larger than its required value (RUex) due to the avoidance of significant fouling problems and process disturbances, as shown in Table 2. In this paper, the aim is to achieve substantial energy saving in heat exchanger networks without any topology modifications. The advantages of HEN retrofit problems without topology modifications have been demonstrated by Wang et al.,2 that is, the heat transfer capacity for the given heat exchanger can be increased without changing its physical size, the capital cost is lower as no repiping and civil works is required, and less production loss occurs since the implementation of enhancement is a relative simple task and can be easily achieved within a normal maintenance period.

6. RETROFITTING THE INDUSTRIAL HEN

In this section, the developed software tool (int-HEAT®) including the packages of exchanger detailed models and the SA optimization approach is utilized to retrofit the industrial scale HEN. To solve such a complicated retrofit problem, the design method proposed in Section 4.2 is presented through all the main steps as follows:

6.1 Retrofitting the whole HEN to reduce energy consumption

6.1.1 Building the existing HEN with int-HEAT® To retrofit the existing HEN, its network structure must be built with int-HEAT® at first (See Figure S1 in the SUPPORTING INFORMATION for interested readers), and then the stream data (temperatures, CPs, and duties) and exchanger data (duties, temperatures) presented in Tables 1 and 2 are input in the existing HEN. More information can be found in 20 ACS Paragon Plus Environment

Page 21 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure S2 in the SUPPORTING INFORMATION for interested readers, including supply temperature (TS), target temperature (TT), duty (DH), heat capacity (CP), and the method used to calculate exchanger area and temperatures. 6.1.2 Setting optimization strategies for the SA approach in int-HEAT® Before the use of the proposed SA optimization approach, several characteristic parameters of the approach must be specified based on Section 3 (See Figure S3 in the SUPPORTING INFORMATION for interested readers). The objective of the SA optimization is to minimize utility with 5 ºC of the minimum approach temperature. Annealing temperatures, Markov chain length, and iteration times are used to control the accuracy and time of calculation, and the cooling schedule is used to control the temperature cooling speed. In move probabilities, only the increase of exchanger duty is considered during the retrofit procedure, since this paper aims to achieve substantial energy saving in the addressed HEN with less exchanger changes under constant network structure, as stated in Section 5. Finally, an optimal solution is obtained.

6.1.3 Optimal solution based on the SA approach The original HEN is now retrofitted to increase energy saving by using the SA optimization approach in int-HEAT®. The optimal retrofit solution is shown in Figure 6. The main data of the retrofitted solution also can be found in Table 2 (the comprehensive information of retrofit can be found in Table S2 in the SUPPORTING INFORMATION for interested readers), where Exchangers 2, 7, 8 and 9 are removed, and the total hot utility (HU) consumption is reduced from 54633.2 kW to 49115.9 kW (10.1% energy saving) leading to $551730 of energy cost reduction per year based on the utility cost of 100 $/kW·year.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 43

6.2 Identifying suitable exchangers for retrofit and retrofit strategies (intensification, additional exchanger area, or both) in the selected exchangers

As presented in Table 2, the required heat transfer coefficients of four exchangers (Exchangers 3, 13, 21 and 24) are larger than their original designed heat transfer coefficients, and lower heat transfer coefficients are required in the rest exchangers. Thus, only these four exchangers are considered for retrofit in the optimal HEN. It also can be found that, significant increase of heat transfer coefficients occur in some selected exchangers, such as Exchanger 13 (+111.1% of heat transfer coefficient), and Exchanger 21 (+156.9% of heat transfer coefficient). It is very difficult to implement intensification techniques in these two exchangers to reach such high heat transfer coefficients and whilst satisfy pressure drop restrictions. Since Exchangers 13 and 21 are not very large size exchangers, adding exchanger area are considered for the retrofit, e.g. a exchanger with larger area (150 m2) and intensified heat transfer (0.65331 kW/m2• ºC of heat transfer coefficient) is installed to replace Exchangers 13, and another exchanger with larger area (155 m2) is used to replace Exchangers 21. As Exchangers 3 and 24 require not very high increase in heat transfer coefficients, only heat transfer intensification is implemented in these two exchangers.

6.3 Evaluating exchanger performances and selecting suitable intensification techniques on the enhanced exchangers

The exchanger detailed model used in int-HEAT® can calculate heat transfer coefficients and pressure drops, as presented in Table 3, including plain (original) tube-side and shell-side pressure drops, plain (original) tube-side and shell-side heat transfer coefficients, original overall heat transfer coefficients, and minimum required heat transfer coefficients


Page 23 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


with intensification (the stream properties and geometry details of these four exchangers used for the calculations can be found in Table S3 and Figure S4 in the SUPPORTING INFORMATION for interested readers). Table 3 shows that, Exchangers 3, 13 and 24 can be enhanced on the tube side to increase their overall heat transfer coefficients to the required values. As the exchangers are existing, the analysis is on a retrofit basis no changes to the geometry were made. hiTRAN.dll developed by Cal Gavin Ltd24 can provide new or retrofit heat exchanger designs with hiTRAN®. Exchangers 3, 13 and 24 are looked at in turn using this dll in conjunction with int-HEAT® software (See Figure S5 in the SUPPORTING INFORMATION). After the simulation, it is found that the ideal candidates for hiTRAN® are Exchangers 3 and 13. This is in agreement with Table 3 which gives details of the heat transfer coefficient for both the tube and shell sides. Exchanger 24 is shell-side controlled with the shell-side coefficient 834.6 kW/m2· ºC less than the tube side coefficient. If the tube-side coefficient is increased on this exchanger then the performance will improve slightly but not significantly as the shell side is the limiting factor for this case and the improvement of the exchanger is shell-side dependant. The tube-side and shell-side coefficients on exchanger 13 are fairly equal compared to exchanger 3 which is heavily tube-side controlled with the shell-side coefficient being 1425.5 kW/m2· ºC larger than the tube side (1151.5 kW/m2· ºC). Table 3 gives details on the plain tube heat transfer coefficient and the enhanced heat transfer coefficient needed. After the analysis, it is found that by using hiTRAN® in Exchanger 3, the tube-side heat transfer coefficient increased by 746.8 kW/m2· ºC and as a result the overall heat transfer coefficient of the exchanger is improved by 124% with pressure drop limitations for the exchanger satisfied. Likewise, the required tube-side coefficients of Exchangers 13 and 24 can be also met with the implementation of hiTRAN®.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 24 of 43

6.4 Retrofit summary

In the optimal retrofit solution, Exchangers 2, 7, 8 and 9 are removed. Since the exchanger area of Exchangers 13 and 21 is not very large, the new exchangers are installed in the retrofitted HEN, e.g. a exchanger with larger area (150 m2) and hiTRAN® (0.65331 kW/m2· ºC of heat transfer coefficient) is installed for Exchangers 13, and another exchanger with larger area (155 m2) is installed for Exchangers 21. As Exchangers 3 and 24 require not very high heat transfer coefficients, only tube-side heat transfer intensification (hiTRAN®) is implemented in these exchangers. The comparison between the original and retrofitted HENs is shown in Table 4, where the hot utility consumption is reduced from 54633.2 kW to 49115.9 leading to $551730 of energy cost reduction per year based on the utility cost of 100 $/kW·year, the total network operating cost decreases as four exchangers are removed, and the retrofit cost is not very high as not network topology modifications is required and only four exchangers are needed for retrofitting. The profits from energy saving and operating cost decreasing, and the cost of implementing of the retrofit techniques lead to $499738.6 of retrofit profit in one year. Moreover, the CO2 emission in the retrofitted HEN in one year decreases 26661.2 ton based on the assumption of 153.23 (kg CO2)/ (GJ utility consumption), namely lower environmental impact. It also can be noted that with the implementation of intensification in the selected exchangers, the cost implications involved are significantly less to the capital costs that would need to be implemented in order to improve the exchangers. Not only would the exchangers need to be bigger but in some cases re-working of pipe work would need to take place due to limitation on plot space which would incur further costs. With regards to the utilities of both streams involved in heating or cooling, these streams to the required temperatures needed for the duty would decrease with intensification as the overall heat 24 ACS Paragon Plus Environment

Page 25 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


transfer coefficient is increased. Therefore by reducing the stream temperatures for the duty of the exchanger, less work is needed for the streams, which means fewer utilities, reducing the running cost of the exchanger further and also being environmentally friendly.

7. CONCLUSIONS The systematic application of conventional retrofit strategies in the process industries is a complicated and challenge task, as large scale problems with large degrees of freedom will be difficult to solve when the trade-off among retrofit costs, energy saving and environmental effect is considered. The existing mathematical programming methodologies based on large-size complex MINLP or MIP-NLP models, are often not readily applicable for retrofit, because there are numerical difficulties related to nonlinearity, problem size and computational time for the optimization of industrial-scale design problems. In this paper, a new retrofit design method is proposed to increase energy saving in the existing HENs, and it is presented to retrofit an existing HEN in a refinery plant with the use of the developed software tool (int-HEAT®). The results of the case study demonstrate the practical applicability of the proposed method and the developed software, and the benefits of implementing intensification techniques in industry.

ASSOCIATED CONTENT Supporting Information The supporting information includes exchanger details in the industrial case (original and retrofitted HEN), and the illustration of using int-HEAT®. This information is available free of charge via the Internet at http://pubs.acs.org/.

AUTHOR INFORMATION Corresponding Author 25 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 26 of 43

*Tel: +44 161 3064390. Email address: [email protected] Notes The authors declare no competing financial interest.

ACKNOWLEDGMENTS Financial support from FP7-SME-2010-1 (262205 Intensified Heat Transfer Technologies for Enhanced Heat Recovery), and ENERGY.2011.5.1-1 (282789 Design Technologies for Multi-scale Innovation and Integration in Post-Combustion CO2 Capture: From Molecules to Unit Operations and Integrated Plants) are gratefully acknowledged.

NOMENCATURE

φ

correction factor for fluid property variation of twisted-tape inserts

△P0

shell-side pressure drop

△Pf0

shell-side pressure drop in straight section

△Pf0,20%Bc

shell-side pressure drop in straight section of shell with 20% baffle cut

△Pfb,20%Bc

shell-side pressure drop in one central baffle spacing when baffle cut is 20%

△Pfi

tube-side pressure drop due to friction loss

△P i

tube-side pressure drop

△Pn0

pressure drop in shell-side nozzles

△Pni

pressure drop in tube-side nozzles

△Pr

tube-side pressure drop related to the tube entrance, exit and return losses

µ0

shell-side fluid viscosity

B

baffles spacing

CCU

a yearly cost parameter per cold-utility-duty unit

CFCPex

heat-flow capacities (the multiplication between heat capacity and flow-rate) of cold stream in exchanger ex

Cfs

parameter for calculating friction factor of helical baffles

COSTR

cost of implementing retrofit strategies 26 ACS Paragon Plus Environment

Page 27 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Cp0

shell-side fluid specific heat capacity

CTI’ex

initial inlet temperatures of cold stream in exchanger ex before retrofit

CTIex

new inlet temperatures of cold stream in exchanger ex after retrofit

D0

tube outer diameter

De

shell equivalent diameter

Dfs

parameter for calculating friction factor of helical baffles

Di

tube inner diameter

Dr

root tube diameter of externally finned tube

Ds

shell side diameter

e

fin height of internal fins

EXcu

set of all cold utility exchangers

EXhu

set of all hot utility exchangers

f0

shell-side friction factor

fc,t

fanning friction factor for coiled-wire inserts within turbulent region

fi

Darcy friction factor

fif,t

fanning friction factor of internal fins in turbulent flow

FL

factor to calculate shell-side heat transfer coefficient

Fp


fs,ef

shell side friction factor of external tube fins

fs,hb

shell side friction factor of helical baffles

ft,t

fanning friction factor of twisted-tape inserts in the turbulent region

func(geometry) geometry function of internal fins for calculating Nusselt number Nuif,t Fz


gc

unit conversion factor, equal to 1.0 kg·m/(N·s2)

H

twist pitch of twisted-tape inserts

h0

shell-side heat transfer coefficient

HCU

a yearly cost parameter per hot-utility-duty unit

HFCPex

heat-flow capacities (the multiplication between heat capacity and flow-rate) of hot stream in exchanger ex

hi

tube-side heat transfer coefficient

hs,ef

shell-side heat transfer coefficient of external tube fins

hs,hb

shell-side heat transfer coefficient of helical baffles

HTI’ex

initial inlet temperatures of hot stream in exchanger ex before retrofit

HTIex

new inlet temperatures of hot stream in exchanger ex after retrofit 27 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 28 of 43

k0

shell-side fluid thermal conductivity

ki

tube-side fluid thermal conductivity

ks

shell-side thermal conductivity

L

exchanger length

lcsw

modified characteristic length for swirling flows of internal fins

nb

number of baffles

Nf

number of internal fins

np

number of tube passes

Nuc,t

Nusselt number for coiled-wire inserts within turbulent region

Nui

tube-side Nusselt number

Nuif,t

enhancement ratio of the Nusselt number of internal fins in turbulent flow

Nus

Nusselt number of shell-side heat transfer coefficient of helical baffles

Nut,t

Nusselt number of twisted-tape inserts in turbulent flow

p

fin pitch of internal fins

p

helical pitch of twisted-tape inserts

PES

profit of energy saving

Prs

shell-side Prandtl number

QS

energy saving in a HEN

Res

shell-side fluid Reynolds number of helical baffles

RP

retrofit profit before tax accumulated over the expected lifetime

Rs

correction factor for unequal baffle spacing

t

average fin width of internal fins

t

inserts thickness twisted-tape inserts

vi

tube-side fluid velocity

vp0

shell-side fluid velocity

w

wire diameter twisted-tape inserts

y

dimensionless twist ratio of twisted-tape inserts

YR

expected project lifetime

α

helix angle of internal fins

αr

number of velocity heads allocated for minor losses

βs

helical angle of helical baffles

δ

tape thickness of twisted-tape inserts

ρi

tube-side fluid density


Page 29 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


REFERENCES (1) Tjoe, T. N.; Linnhoff, B. Using Pinch technology for process retrofit. Chem. Eng. 1986, 28, 47-60. (2) Wang, Y.; Pan, M.; Bulatov, I.; Smith, R.; Kim, J. K. Application of intensified heat transfer for the retrofit of heat exchanger network. Appl. Energy. 2012, 89, 45-59. (3) Ciric, A.R.; Floudas, C.A. A retrofit approach for heat exchanger networks. Comput. Chem. Eng. 1989, 13(6), 703-715,. (4) Sorsak, A.; Kravanja, Z. MINLP retrofit of heat exchanger networks comprising different exchanger types. Comput. Chem. Eng. 2004, 28, 235-251. (5) Rezaei, E.; Shafiei, S. Heat exchanger networks retrofit by coupling genetic algorithm with NLP and ILP methods. Comput. Chem. Eng. 2009, 33, 1451-1459. (6) Kovac, Kralj, A.; Glavibvc, P. Retrofit of complex and energy intensive processes. Comput. Chem. Eng. 1997, 21, S517-S522. (7) Polley, G.T.; Reyes Athie, C.M. Gough M. Use of heat transfer enhancement in process integration. Heat Recovery Systems and CHP. 1992, 12(3), 191-202. (8) Zhu, X.; Zanfir, M.; Klemes, J. Heat transfer enhancement for heat exchanger network retrofit. Heat Transfer Eng. 2000, 21(2), 7-18. (9) Pan, M.; Bulatov, I.; Smith, R.; Kim, J.K. Novel optimization method for retrofitting heat exchanger networks with intensified heat transfer. Comput. Aided Chem. Eng. 2011, 29, 1864-1868. (10) Pan, M.; Bulatov, I.; Smith, R.; Kim, J.K. Improving energy recovery in heat exchanger network with intensified tube-side heat transfer. Chem. Eng. Trans. 2011, 25, 375-380. (11) Pan, M.; Bulatov, I.; Smith, R.; Kim, J.K. Novel MILP-based iterative method for the retrofit of heat exchanger networks with intensified heat transfer. Comput. Chem. Eng.

2012, 42, 263-276. (12) Pan, M.; Smith, R.; Bulatov, I. A novel optimization approach of improving energy recovery in retrofitting heat exchanger network with exchanger details. Energy. 2013, 57, 208-221. (13) Pan, M.; Bulatov, I.; Smith, R.; Kim, J.K. Optimisation for the retrofit of large scale heat exchanger networks with comprising different intensified heat transfer techniques. Appl. Therm. Eng. 2013, 53, 373-386. (14) Pan, M.; Bulatov, I.; Smith, R. Novel MILP-based optimization method for retrofitting heat exchanger networks. Comput. Aided Chem. Eng. 2012, 30, 567-571.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 30 of 43

(15) Pan, M.; Bulatov, I.; Smith, R. Novel MILP-based optimization method for heat exchanger network retrofit considering stream splitting. Comput. Aided Chem. Eng. 2012, 31, 395-399. (16) Pan, M.; Bulatov, I.; Smith, R. New MILP-based iterative approach for retrofitting heat exchanger networks with conventional network structure modifications. Chem. Eng. Sci.

2013, 104, 498-524. (17) Pan, M.; Bulatov, I.; Smith, R. Intensifying heat transfer for retrofitting heat exchanger networks with topology modifications. Comput. Aided Chem. Eng. 2013, 32, 307-312. (18) Pan, M.; Bulatov, I.; Smith, R. Retrofit procedure for intensifying heat transfer in heat exchanger networks prone to fouling deposition. Chem. Eng. Trans. 2012, 29, -1428. (19) Pan, M.; Bulatov, I.; Smith, R. Exploiting tube inserts to intensify heat transfer for the retrofit of heat exchanger networks with considering fouling mitigation. Ind. Eng. Chem. Res. 2013, 52(8), 2925-2943. (20) Pan, M..; Jamaliniya, S.; Smith, R.; Bulatov, I.; Gough, M.; Higley, T.; Droegemueller, P. New insights to implement heat transfer intensification for shell and tube heat exchangers. Energy, 2013, 57, 208-221. (21) Jensen, M.; Vlakancic, A. Technical note - experimental investigation of turbulent heat transfer and fluid flow in internally finned tubes. Int. J. Heat and Mass Transfer. 1999, 42, 1343-1351. (22) Manglik, R.; Bergles, A. heat transfer and pressure drop correlations for twisted-tape inserts in isothermal tubes. Part II: transition and turbulent flows. ASME J. of Heat Transfer. 1993, 115,890-896. (23) Garcia, A.; Vicente, P.; Viedma, A. Experimental study of heat transfer enhancement with wire coil inserts in laminar-transition-turbulent regimes at different prandtl numbers. Int. J. Heat and Mass Transfer. 2005, 48, 4640-4651. (24) Cal Gavin Ltd. hiTRAN® thermal systems: changing fluid dynamics and harnessing the benefits. Tech. report of Cal Gavin Ltd. 2009. (25) Serth, R.W. Design of shell-and-tube heat exchangers, Handbook of process heat transfer principles and applications. Elsevier: Oxford, 2007. (26) Zhang, J.F.; Li, B.; Huang, W.J.; Lei, Y.G.; He, Y.L.; Tao, W.Q. Experimental performance comparison of shell-side heat transfer for shell-and-tube heat exchangers with middle-overlapped helical baffles and segmental baffles. Chem. Eng. Sci. 2009, 64, 1643-1653.


Page 31 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(27) Wang, Y.; Smith, R.; Kim, J. K. Heat exchanger network retrofit optimization involving heat transfer enhancement. Appl. Therm. Eng. 2012, 43, 7-13. (28) Kirkpatrick, S.; Gelatt, C.D.; Vecchi, M.P. Optimization by simulated annealing, Sci.

1983, 220, 671-680.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 32 of 43

List of tables Table 1. Stream details in case studies. Table 2. Exchanger data in the original and retrofitted HENs. Table 3. Heat transfer coefficients, pressure drop restrictions and exchanger area of Exchangers 3, 13, 21 and 24 for the retrofit.

Table 4. Comparison between the original and retrofitted HENs.


Page 33 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48


Table 1. Stream details in case studies. Stream C1 C2 C3 H1 H2 H3 H4 H5 H6 H7 H8 H9 FCP (kW/ ºC) 322.5 358.5 474.0 14.2 181.5 113.0 99.9 22.2 39.5 28.0 176.0 24.5 Tin (ºC) 33.51 91.34 151.05 335.40 253.20 293.70 212.44 212.68 174.40 364.26 290.38 284.20 Tout (ºC) 95.59 157.27 351.93 69.44 116.05 130.00 156.05 61.77 43.33 65.56 210.90 65.56 Duty (kW) 20020.8 23635.9 95217.1 3776.6 24892.7 18498.1 5633.4 3352.4 5177.3 8363.6 13988.5 5356.7

H10 25.0 240.07 57.78 4557.3

H11 HU CU 69.6 93.0 9652.5 178.70 1500 12.45 69.30 7614.2

FCP: heat-flow capacities (the multiplication between heat capacity and flow-rate); Tin, Tout: stream inlet and outlet temperatures in HEN.

33

ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

Page 34 of 43

Table 2. Exchanger data in the original the retrofitted HENs. EXs 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31

LMTDex (ºC) 51.660 116.416 74.024 54.291 125.521 45.500 50.939 56.573 39.535 90.200 64.218 52.070 87.473 86.996 74.344 51.308 82.246 31.809 83.862 31.243 19.665 34.741 42.634 57.110 202.682 44.923 29.175 35.065 57.142 891.221 114.751

EXAex (m2) 227.487 15.166 303.316 227.487 130.000 845.000 26.000 39.000 72.228 361.140 80.000 50.000 100.000 111.436 55.718 167.154 269.997 280.000 179.998 1799.980 60.000 120.000 53.999 1100.000 86.658 389.961 320.000 1326.000 312.000 234.000 78.000

Original HEN RUex DUex (kW/m2· ºC) (kW/m2· ºC) 0.09622 0.18763 0.07917 0.15182 0.17193 0.33527 0.21715 0.42348 0.34523 0.67320 0.16065 0.31326 0.07734 0.15357 0.25905 0.50424 0.26102 0.50828 0.08696 0.16973 0.22220 0.47937 0.44196 1.49152 0.23655 0.46428 0.23580 0.45998 0.13678 0.26513 0.88462 1.72500 0.14455 0.28188 0.09640 0.19495 0.12031 0.23451 0.20777 0.40515 0.25105 0.81579 0.23234 0.50335 0.74903 1.46059 0.13571 0. 23500 0.11872 0.23148 0.61488 1.19902 0.08360 0.17196 0.25601 0.49922 0.18106 0.35306 0.26197 0.51085 0.52322 1.02114

Duty (kW) 1130.8 139.8 3860.3 2682.0 5633.4 6176.5 102.4 571.6 745.4 2832.8 1141.5 1150.6 2069.2 2285.9 566.6 7586.7 3210.1 858.6 1816.1 11684.2 296.2 968.6 1724.4 8525.4 2085.1 10771.6 780.5 11903.4 3228.0 54633.2 4683.1

LMTDex (ºC) 36.66

EXAex (m2) 227.487

41.15 36.19 123.35 56.71

303.316 227.487 130.000 845.000

86.07 74.06 25.04 36.72 63.74 68.95 49.90 57.84 16.52 79.07 14.98 11.29 21.87 27.55 39.51 201.77 29.50 14.34 19.41 35.39 919.99 109.11

361.140 80.000 50.000 150.000 (+50%) 111.436 55.718 167.154 269.997 280.000 179.998 1799.980 155.000 (+158%) 120.000 53.999 1100.000 86.658 389.961 320.000 1326.000 312.000 234.000 78.000

Retrofitted HEN RUex DUex (kW/m2· ºC) (kW/m2· ºC) 0.06087 0.18763 Removed 0.41482 0.41482 (+23.7%) 0.36082 0.42348 0.35122 0.67320 0.07338 0.31326 Removed Removed Removed 0.08118 0.16973 0.45225 0.47937 1.26771 1.49152 0.65331 0.65331 (+41.6%) 0.09414 0.45998 0.09947 0.26513 0.89700 1.72500 0.21736 0.28188 0.19229 0.19495 0.10523 0.23451 0.38767 0.40515 0.81116 0.81579 0.47874 0.50335 1.36735 1.46059 0.27748 0.27748 (+18.1%) 0.08485 0.23148 0.95755 1.19902 0.17036 0.17196 0.48395 0.49922 0.30036 0.35306 0.22815 0.51085 0.27981 1.02114

Duty (kW) 507.6 5177.5 2970.7 5632.1 3517.0

2523.0 2679.6 1587.4 3598.7 668.6 382.1 7482.4 3394.8 889.2 1497.7 10450.5 1419.4 1256.6 2033.9 12059.6 1483.7 11015.5 781.9 12457.9 3316.5 49115.9 2381.3

Exs: exchangers; LMTDex: logarithmic mean temperature difference in exchanger ex; EXAex: designed area of exchanger ex; RUex: required heat transfer coefficient of exchanger ex; DUex: designed heat transfer coefficient of exchanger ex.

34


Page 35 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48


Table 3. Heat transfer coefficients, pressure drop restrictions and exchanger area of Exchangers 3, 13, 21 and 24 for the retrofit. Heat Exchanger 3 Tube side Original heat transfer coefficient hi (kW/m2· ºC) Minimum required heat transfer coefficient hie (kW/m2· ºC) Original pressure drop ∆Pi (kPa) Maximum pressure drop ∆Pie (kPa) Shell side Original heat transfer coefficient hs (kW/m2· ºC) Original pressure drop ∆Ps (kPa)

Heat Exchanger 13 Heat Exchanger 21

Heat Exchanger 24

549.6 746.8 (Enhanced) 2.79 100

1151.5 3001.3 (Enhanced) 7.17 100

5728.9 5728.9 57.00 100

1049.4 2730.2 (Enhanced) 41.37 500

2603.1 10.08

1425.5 5.63

1846.6 32.86

834.6 22.50

0.81579 0.81579

0. 23500 0.27748 (Enhanced)

Overall performance Original overall heat transfer coefficient U (kW/m2· ºC) 0.33527 0.46428 Minimum required heat transfer coefficient Ue (kW/m2· ºC) 0.41482 (Enhanced) 0.65331 (Enhanced)

35



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 36 of 43

Table 4. Comparison between the original and retrofitted HENs. Original HEN Retrofitted HEN 54633.2 49115.9 16969.6 11616.0 4710051.7 4161590.5 (-11.6%)

Hot Utility (kW) Cold Utility (kW) Total Utility Cost ($/yr) Exchangers Total Network Operating Cost ($/yr) Retrofit Cost ($) Retrofit Profit in One Year ($) CO2 Emission in One Year (ton), (153.23 kg CO2/GJ Total Utility)

31 1984275.8

27 1928998.4 (-2.8%)

0 0

104000.0 499738.6

264002.0

237340.8 (-10.1%)


Page 37 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


List of figures Figure 1. Geometry specifics of shell-and-tube heat exchanger. Figure 2. The illustrations of tube-side intensification techniques. Figure 3. The illustrations of shell-side intensification techniques. Figure 4. Procedure of simulated annealing algorithm. Figure 5. An industrial HEN. Figure 6. The optimal retrofitted HEN using the SA optimization approach in int-HEAT®.



Shell-side Tube-side Inlet Inlet

nb = 6

90 º tube pattern

A

B

Dno

Shell

Shell

L Bout

Baffle cut Bc Tube bundle

Dni

Baffle

PT Lsb

Ds

Tube

Dni

Dno

Tube pass 1 2

Bin

A’

D0

Di

Baffle nt = 47

……

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

Page 38 of 43

np

Tube-side Shell-side Outlet Outlet (a) shell-and-tube heat exchanger

(b) A-A’ (tube bundle)

(c) A-A’ (baffle cut & tube pass)

L: tube length, B: baffle spacing, Bin: inlet baffle spacing, Bout: outlet baffle spacing, nb: baffle number, Dno: shell-side nozzle diameter, Dni: tube-side nozzle diameter, PT: tube pitch, Ds: shell diameter, Di: tube inner diameter, D0: tube outer diameter, Lsb: Shell-bundle clearance, nt: tube number, Bc: baffle cut, np: tube pass Figure 1. Geometry specifics of shell-and-tube heat exchanger

38


Page 39 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


A

Tube

α

p

Internal fins

Internal fins

e

t

A

Section A-A (a) Internal fins

Twisted-tape inserts

A H

δ A Di Section A-A

Tube (b) Twisted-tape inserts

Coiled-wire inserts w w

p t

Tube (c) Coiled-wire inserts

(d) hiTRAN®

Figure 2. The illustrations of tube-side intensification techniques.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

b

Page 40 of 43

Dr Di

τ

D0 (a) external tube fins

Continuous helical baffles

Tube

Noncontinuous helical baffles

Tube βs B

Continuous

Noncontinuous (b) Helical baffles

Figure 3. The illustrations of shell-side intensification techniques.


Page 41 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Estimate an initial trial solution

Set an initial annealing temperature (AT)

Repeat times (t) Is t > tmax ?

Yes

No Generate a new trial solution by making a random move

Evaluate objective function of the new trial solution

Yes

Is the new solution better than the current one?

Accept the move

No

Reject the move

Reduce annealing temperature (AT)

No

Is AT ≈ ATfinal ?

Yes Stop

Figure 4. Procedure of simulated annealing algorithm.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

12

13

6

Page 42 of 43

1

5

C1

3 21

20

18

17

C2 16

30

29

28

27 26 24

4

23

H1

22

C3

17

H2

15

24 20

H3

32 16

26

2

H4

5 22 12

H5

7

1

H6

3

8

9

H7

31

H8

29 28

H9

25

H10

4

18

H11

19 23

H12 HU

14

13

27

21

10 6

11

30 2 Hot stream:

H

Hot utility: HU

Cold stream:

C

25 32 19 15 14 Cold utility:

7

11 10

8

9

31

CU

CU

Figure 5. An industrial HEN. 42


Page 43 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48


Removed

Figure 6. The optimal retrofitted HEN using the SA optimization approach in int-HEAT®.


Efficient Retrofitting Approach for Improving Heat Recovery in Heat

Recommend Documents