Format And Type Fonts


 

 CCHHEEMMIICCAALL  EENNGGIINNEEEERRIINNGG  TTRRAANNSSAACCTTIIOONNSS  
 

VOL. 45, 2015 

A publication of 

 
The Italian Association 

of Chemical Engineering 

www.aidic.it/cet 
Guest Editors: Petar Sabev Varbanov, Jiří Jaromír Klemeš, Sharifah Rafidah Wan Alwi, Jun Yow Yong, Xia Liu  

Copyright © 2015, AIDIC Servizi S.r.l., 

ISBN 978-88-95608-36-5; ISSN 2283-9216 DOI: 10.3303/CET1545006 

 

Please cite this article as: Čuček L., Kravanja Z., 2015, A procedure for the retrofitting of large-scale heat exchanger 

networks for fixed and flexible designsesentation of the grid diagram for heat exchanger networks, Chemical Engineering 

Transactions, 45, 31-36  DOI:10.3303/CET1545006 

31 

A Procedure for the Retrofitting of Large-Scale Heat 

Exchanger Networks for Fixed and Flexible Designs 

Lidija Čuček, Zdravko Kravanja 

Faculty of Chemistry and Chemical Engineering, University of Maribor, Smetanova ulica 17, 2000 Maribor, Slovenia 

zdravko.kravanja@um.si 

This contribution presents a novel three-step procedure incorporating software tools TransGen and 

HENSYN for the retrofitting of heat exchanger networks (HENs). This procedure is used for improving Heat 

Integration (HI) within individual plants and Total Sites (TSs) under fixed or steady-state conditions and 

under varying operational conditions. The software tool TransGen is used during the first targeting step 

and the second identification step for the most promising modifications, and HENSYN is used for the final 

synthesis of the retrofitted HEN. The targeting step is based on a combined Mathematical 

Programming/Pinch Analysis (MP/PA) approach, whilst the identification and the synthesis steps on the 

MP approach. The main objective of both tools TransGen and HENSYN is maximisation of the economic 

performance (annual incremental profit) of plants and/or TSs by minimisation of energy consumption and 

cost, and maximisation of economically-viable production of energy, such as production of hot water for 

district heating. Both tools consider the trade-offs between operating and investment cost but HENSYN 

considers it in more details. An illustrative example shows the capabilities of the developed software tools 

and the three-step procedure. 

1. Introduction  

Increased global competition and environmental awareness have forced industrial companies to improve 

the performances of their processes and also TSs and Locally Integrated Energy Sectors (Perry et al., 

2008). HI, TS Integration (TSI) and waste heat utilisation within industrial processes and TSs are efficient 

ways of conserving energy and achieving emission reductions. Several approaches have been developed 

for this purpose. In general they are divided into an approach based on physical insights – PA, and an 

approach based on mathematical programming – MP (Klemeš and Kravanja, 2013). Both approaches 

have their own advantages and drawbacks. In order to optimise the energy consumption within large-scale 

process plants and TSs under varying operating conditions the MP and PA approaches should be 

combined and further extended. Over recent years combined MP/PA approaches for overcoming the 

drawbacks of both approaches have also been developed. 

A novel three-step procedure incorporating software tools called TransGen and HENSYN has been 

developed for HI and TS optimisation of energy and cost in order to be applied for energy targeting and 

retrofitting of existing large-scale industrial plants and TSs. The software tool TransGen is intended for 

targeting during the first step and for identification of the most optimal retrofitting solutions during the 

second step. The targeting step is based on a combined MP/PA, and the identification step on MP and on 

the superstructure mixed-integer linear programming (MILP) approach. The identification step through pre-

screening of modification alternatives also narrows the original search space and guides the optimisation 

towards obtaining (near) global as well as feasible solutions regarding retrofitting.  

Further, during the third-step, the software tool HENSYN is used for more detailed HEN synthesis on the 

reduced space of alternatives. Only those data related to selected modifications during the second step 

are used during the third step. The third step also enables easy obtaining the final structure regarding the 

retrofitted HEN including positions and temperatures of heat exchangers within the network. 



 

 

32 

 
The three-step procedure and the methodology for the TSI within existing TSs is presented in the following 

and demonstrated on an illustrative example of a TS consisting of two process plants under varying 

operating conditions (modified from Čuček and Kravanja, 2014). 

2. Problem Description 

In order to perform the retrofitting of the HENs within large-scale industrial process plants and TSs, the 

combined MP/PA approach was applied, and the following three-step procedure was developed and 

applied: 

i) Targeting step based on MP/PA. By comparing the solutions obtained for target and existing designs, 

the potential is identified for HI, TSI and waste heat utilisation. The targeting step is performed using the 

software tool TransGen;  

ii) Identification step with forbidding unfeasible matches. This step is based on MP, on an extension of the 

expanded transhipment model developed by Papoulias and Grossmann (1983). This step enables the 

obtaining of the most optimal retrofitting modifications regarding energy consumption reduction and 

intermediate utilities production in regards to trade-offs between operating and investment costs. It 

should be noted that several loops could be required in order that the obtained results are (near) 

optimal, verified and feasible. Second step is also performed using TransGen;  

iii) Synthesis step based on MP on the stage-wise superstructure (Yee and Grossmann, 1991) extended 

for retrofits, different exchanger types (Soršak and Kravanja, 2004) and the dynamics of the operations. 

It enables the obtaining of detailed HEN retrofit from those modifications identified during the second 

step. The third step is performed using software tool HENSYN. 

2.1 Methodology 
The following methodology is used for HI and TSI using the novel three-step procedure and software tools 

TransGen and HENSYN, see also Figure 1: 

i) Data acquisition and  incorporation of data within TransGen; 

ii) Targeting using software tool TransGen and a comparison with the existing energy consumption; 

iii) Identification of the most profitable modifications accompanied by pre-screening of the alternatives and 

forbidding the infeasible matches; 

iv) Verifications of the obtained solutions from step iii) and returning to step iii) as long as all proposed heat 

exchange matches are acceptable. Several loops could be required to obtain a near optimal, verified 

and feasible solution; 

v) Preparation of the data for selected modifications and the synthesis of the retrofitted HEN using 

software HENSYN. 

Data acquisition

Targeting  Potential for HI and TSI

Pre-screening and selection of 

modifications 

Verifications of modifications 

Preparation of data and synthesis of 

retrofitted HEN

feasible

infeasible

 

Figure 1: Methodology used for Heat and Total Site Integration 

2.2 Software Tool TransGen 
The software tool TransGen is composed of two parts: i) the data-independent algorithm and ii) the data. 

In order to perform any study, only the data should be changed appropriately. 

Software tool TransGen is used for:  

i) Targeting and comparison of the target and existing designs in order to obtain the potential for HI, TSI 

and intermediate utility production; 

ii) Identification of modifications by considering forbidding the infeasible matches; 

iii) Preparation of the data for selected modifications to be used in software tool HENSYN. 



 

 

33 

TransGen is based on combined MP/PA, and the graphical and numerical output is obtained. Graphical 

output could be in the form of Composite Curves (CCs), Balanced CCs, Grand Composite Curves (GCCs) 

and Total Site Profiles (TSPs). It could be produced for targeted, existing and modified designs, and for 

both nominal designs under fixed conditions, and under uncertain conditions for flexible designs. 

Numerical output provides economic parameters, e.g. incremental profit, investment cost, pipeline cost, 

and payback time, utility consumption reduction, modifications of existing heat exchange units, proposed 

new matches, etc. The features of software tool TransGen are presented in Čuček and Kravanja (2014), 

and those for improved version of TransGen in Čuček and Kravanja (2015). 

2.3 Software Tool HENSYN 
Software tool HENSYN is used for the detailed synthesis of the retrofitted HENs. HENSYN is also 

composed of two parts: i) the algorithm and ii) the data. Only the data file should be changed for the 

specific analysis.  

HENSYN is based on MP and the following results could be obtained for synthesised HENs, such as the 

economic data, areas of proposed heat exchangers, temperatures at all the stages, locations of proposed 

modified heat exchangers and other results. By having such results, retrofitted HEN could be drawn. 

The main features of software tool HENSYN are: 

i) Performing the detailed synthesis regarding the trade-offs between investment and operating cost on a 

smaller number of solutions. Only those modifications identified using TransGen are optimised within 

HENSYN; 

ii) Synthesis of modified HENs including positions and temperatures of heat exchanger units within 

networks under fixed and flexible designs. A multi-period model has been developed for obtaining a 

flexible design; 

iii) Accounting for different heat exchanger types. In the current version of HENSYN there are four optional 

heat exchanger types which can be selected in the proposed HEN design, such as double pipe, plate 

and frame, fixed plate shell and tube and shell and tube with U-tubes. Other types of heat exchangers 

could also be added. 

The methodology used for HI and TSI using the three-step procedure and softwares TransGen and HEN 

SYN is illustrated in the following on a simple illustrative example. 

3. Illustrative example 

3.1 Description of Illustrative Example 

A simple illustrative example of TS consisting of two units A and B (Perry et al., 2008) with a small number 

of streams was selected to demonstrate the methodology and applicability of the developed procedure and 

software tools TransGen and HENSYN. Such an illustrative example was selected in order to be 

representative and easily verified. The case study under flexible varying operational conditions was 

analysed.  

It was assumed that units A and B are located 1.5 km apart. All process streams were assumed to be 

liquids and for simplicity it was also assumed that the pressures of the process streams are at 1 bar.  

The data for the plants A and B were taken from Perry et al. (2008), and were slightly modified to account 

for varying operating conditions (see Čuček and Kravanja, 2015). The HEN for TS under varying operating 

conditions is shown in Figure 2, where the case 1 is represented as “A1”, case 2 as “A2” and case 3 as 

“A3”. 

3.2 Targeting and Identifying the Potential for Heat and Total Site Integration 

The first step was to perform targeting. By comparing the target and existing designs, it was possible to 

identify the potentials for improvements within process plants and TSs in terms of energy consumption 

reduction. The main results obtained for the target and existing designs are shown in Table 1.  

Table 1:  Main results regarding existing and target designs 

 Existing design Target design 

Average energy cost 3.4 M€/y 0.9 M€/y 

Average energy consumption 21.4 MW 4.7 MW 

Hot utility consumption   

Case 1 | Case 2 | Case 3 10.1 MW | 9.6 MW | 15.4 MW 1.7 MW | 3.7 MW | 6.7 MW 

Cold utility consumption   

Case 1 | Case 2 | Case 3 9.8 MW | 9.6 MW | 9.8 MW 0.1 MW | 2 MW | 0.2 MW 

 



 

 

34 

 
Further, graphical outputs for target and existing designs were obtained, such as GCCs, CCs, Balanced 

CCs and TSPs. Figure 3 shows an example of graphical output under varying conditions. It shows TSPs 

for existing designs with the potential for hot water production. As there are varying conditions, the 

solutions are also varying, and TSPs show the ranges of the solutions. 

H1

A1: 170 °C

A2: 167 °C

A3: 172 °C

A1: 25 °C

A2: 25 °C

A3: 18 °C

C1

A1: 1.5 MW

A2: 1.36 MW

A3: 1.39 MW

A1: 100 °C

A2: 109 °C

A3: 89 °C

1

A1: 143 °C

A2: 144.48 °C

A3: 152.11 °C

A1: 30 °C

A2: 41 °C

A3: 36 °C

C3

A1: 3.5 MW

A2: 3.58 MW

A3: 4.3 MW

A1: 53.3 °C

A2: 65 °C

A3: 61 °C

2

A1: 80 °C

A2: 84.98 °C

A3: 90.59 °C

170 °C

170 °C

3
A1: 1.75 MW

A1: 65 °C

19 °C

15 °C

3 A2: 0.66 MW
A3: 0.67 MW

A2: 74 °C

A3: 81 °C
H3

A1: 200 °C

A2: 196 °C

A3: 202 °C

A1: 75 °C

A2: 79 °C

A3: 60 °C

C7

A1: 7 MW

A2: 6.54 MW

A3: 7.75 MW

A1: 150 °C

A2: 154 °C

A3: 147 °C

6

A1: 116 °C

A2: 112.52 °C

A3: 98.36 °C

A1: 20 °C

A2: 20 °C

A3: 21 °C

C4

A1: 3 MW

A2: 3 MW

A3: 1.75 MW

A1: 80 °C

A2: 80.1 °C

A3: 58.3 °C

7

A1: 80 °C

A2: 74 °C

A3: 75 °C

H4

A1: 150 °C

A2: 165 °C

A3: 147 °C

A1: 2.15 MW

A2: 3.2 MW

A3: 2.58 MW

A1: 104.3 °C

A2: 107.5 °C

A3: 93 °C

8

A1: 120.4 °C

A2: 121 °C

A3: 109 °C

A1: 1 MW

A2: 1 MW

A3: 1.35 MW

A1: 100 °C

A2: 101 °C

A3: 87 °C

10

A1: 106.65 °C

A2: 106.45 °C

A3: 89.19 °C

C5

A1: 100 °C

A2: 101 °C

A3: 89 °C

9 A1: 7.85 MW
A2: 9.1 MW

A3: 14.47 MW

A1: 120 °C

A2: 126 °C

A3: 121 °C

A1: 60 °C

A2: 61 °C

A3: 53 °C

C6

A1: 0.5 MW

A2: 0.66 MW

A3: 0.33 MW

A1: 85 °C

A2: 86 °C

A3: 69 °C

11

A1: 99.78 °C

A2: 97.23 °C

A3: 84.4 °C

12

13 A1: 4.35 MW
A2: 4.2 MW

A3: 3.22 MW

A1: 40 °C

A2: 39 °C

A3: 37 °C

A1: 0.5 MW

A2: 0.45 MW

A3: 0.52 MW

A1: 110 °C

A2: 103 °C

A3: 95 °C

A1: 150 °C

A2: 154 °C

A3: 145 °C

A1: 70 °C

A2: 67 °C

A3: 74 °C

C2

A1: 1.05 MW

A2: 0.92 MW

A3: 0.83 MW

A1: 100 °C

A2: 98 °C

A3: 104 °C

4

A1: 134.6 °C

A2: 139 °C

A3: 134.2 °C

5 A1:5.43 MW
A2: 4.72 MW

A3: 5.95 MW

A1: 55 °C

A2: 62 °C

A3: 57 °C
H2

19 °C

15 °C

170 °C

170 °C

170 °C

170 °C

19 °C

15 °C

 

Figure 2: HEN for plants A and B under varying operating conditions 

 

Figure 3: Total Site Profiles for existing design operating under varying conditions with the potential for hot 

water production 

3.3 Identification of Modifications 
As from the targeting stage significant potential for energy consumption reduction was shown, the next 

step was the pre-screening and selections of the modifications. It was also assumed that 2 MW of hot 

water should be produced. The analyses regarding the most optimal modifications by considering the 

trade-offs between energy and investment cost were thus performed. All the modifications were compared 

to the existing design, and thus the results directly showed possible improvements in terms of energy 

consumption and cost.  

The retrofit modifications were analysed only for a TS under varying operating conditions. However the 

same analysis could be performed at the plant level, and under steady-state operating conditions.  

Analysis was performed for a certain number of the most optimal new heat exchange matches, e.g. for 5 

new matches. The main results regarding modified designs with 5 modifications are shown in Table 2, see 

also the contribution by Čuček and Kravanja (2015). 



 

 

35 

It could be seen again that all the results were obtained within ranges. Most of the new heat exchange 

matches were selected at TS level, whilst 2 MW of hot water was produced from hot stream 5 at plant A. 

The next step was verifications of the modifications and repeating this step until all the proposed heat 

exchange matches were acceptable for the leaders and engineers of the plant and/or TS. The 

identification step could be repeated several times by excluding infeasible (and difficult) modifications, 

before the final solution could be obtained. In this illustrative example it was assumed that all the 

modifications were acceptable and thus this step was not repeated. In the case where there would be 

infeasible modifications proposed, those infeasible (and difficult) matches would have been defined as 

forbidden. 

Table 2: Main results regarding modified design with 5 modifications  

 Modified design with 5 modifications  

Average profit 1.25 M€/y 

Average energy savings  7.96 M€/y 

Retrofit area investment 0.72 M€ 

Pipe investment 0.83 M€ 

Payback time 0.94 y 

Modified heat exchange units (fraction of 

rearranged heat exchange unit‟s energy) 

2 (29 – 40 %), 5 (87 – 96 %), 6 (30 – 43 %), 9 (32 – 43 %), 

13 (27 – 60 %) 

New heat exchange matches at process level H6 – C9 (1.9 – 3.3 MW), H5 – hot water (2 MW) 

New heat exchange matches at TS level H2 – C9 (1 – 1.7 MW), H5 – C6 (2.2 – 3.7 MW), H13 – C2 

(1.1 – 1.9 MW) 

Hot utility consumption (absolute and relative reduction) 

Case 1 | Case 2 | Case 3  6.7 MW (3.4 MW – 33 %) | 6.7 MW (2.9 MW – 30 %) | 

10.6 MW (4.9 MW – 31.6 %) 

Cold utility consumption (absolute and relative reduction) 

Case 1 | Case 2 | Case 3 3.9 MW (5.9 MW – 60 %) | 4.3 MW (5.3 MW – 55 %) |   

2.2 MW (7.6 MW – 77 %) 

 

3.4 Synthesis of Retrofitted Heat Exchanger Network 
Finally, the detailed synthesis of retrofitted HEN with the appropriate trade-offs between investment and 

operating cost was performed using the software tool HENSYN. The synthesis was performed only for 

those selected retrofitting modifications during the second identification step. The retrofitted HEN could be 

easily extracted from the HENSYN solution output. The main results from the detailed analysis are 

presented in Table 3, and compared with those results obtained from TransGen. 

Table 3: Main results obtained from TransGen and HEN SYN 

 Modified design with 5 modifications 

 TransGen HENSYN 

Average profit 1.25 M€/y 1.3 M€/y 

Retrofit area investment 0.72 M€ 0.79 M€ 

Pipe investment 0.83 M€ 0.83 M€ 

 

Figure 4 shows the retrofitted HEN design when performing 5 of the most profitable modifications at the TS 

level under varying operating conditions. HEN also includes all the temperatures, exchanged heat, and 

areas and types of selected heat exchangers. Two heat exchanger types were selected, fixed plate shell 

and tube (“F”) and shell and tube with U tubes (“U”). This figure is available in colour online. Existing heat 

exchangers that now had the smaller heat exchanger area are highlighted in the online version in grey, 

and the new ones in yellow. The heaters are highlighted in red and the coolers in blue. It can be seen that 

there are 5 new heat exchangers, as was set. From those five, three are exchanging heat between the 

plants A and B, one within plant B, and one is used for hot water production (previously heat exchanger 5). 

Hot water production was completely satisfied. 

4. Conclusions 

This contribution presented and demonstrated a three-step procedure and software tools TransGen and 

HEN SYN for the retrofitting of existing process plants and TSs under varying conditions. It should be 



 

 

36 

 
noted that the analysis could be even more easily performed for plants and TSs under steady-state 

conditions.  

Both software tools are based on MP, whilst TransGen uses also insights from PA, and the whole 

procedure is automated. Both software tools are written in data-independent ways and could be easily 

applied to any process plant or TS under steady-state and varying conditions with any number of heat 

exchange units just by changing the data. The proposed modifications which consider trade-offs between 

investment and operating cost, could be obtained in „real‟ time. Problems of any size could be handled 

ranging from small processes up to complex industrial TSs, see also the contribution by Čuček et al. 

(2015) where the procedure is illustrated on the existing refinery‟s TS with more than 100 HE units. 

A1: 40 °C

A2: 40 °C

A3: 40 °C

A1: 30 °C

A2: 41 °C

A3: 36 °C

A1: 75 °C

A2: 79 °C

A3: 60 °C

A1: 104.3 °C

A2: 107.5 °C

A3: 93 °C

A1: 80 °C

A2: 84.98 °C

A3: 90.59 °C

A1: 55 °C

A2: 62 °C

A3: 57 °C

A1: 116  °C

A2: 112.52 °C

A3: 98.36 °C

A1: 40 °C

A2: 39 °C

A3: 37 °C

H13

Hot 

water

H2

H5

H6

C2

C6

C9

A1: 60 °C

A2: 60 °C

A3: 60 °C

A1: 143 °C

A2: 144.48 °C

A3: 152.11 °C

A1: 134.6 °C

A2: 139 °C

A3: 134.2 °C

A1: 200 °C

A2: 196 °C

A3: 202 °C

A1: 99.78 °C

A2: 97.23 °C

A3: 84.4 °C

A1: 54.07 °C

A2: 65.76 °C

A3: 62.12 °C

A1: 120 °C

A2: 126 °C

A3: 121 °C

A1: 152.89 °C

A2: 156.48 °C

A3: 151.14 °C

A1: 150 °C

A2: 154 °C

A3: 147 °C

A1: 53.3 °C

A2: 65 °C

A3: 61 °C

F – fixed plate shell and tube

U – shell and tube with U tubes

A1: 2.64 MW

A2: 2.46 MW

A3: 0.61 MW

A1: 0.53 MW

A2: 0.66 MW

A3: 1.42 MW

A1: 3.74 MW

A2: 5.75 MW

A3: 10.52 MW

A1: 76.29 °C

A2: 73.12 °C

A3: 46 °C

A1: 62.83 °C

A2: 72.84 °C

A3: 75.48 °C

Hot water production satisfied

A1: 2 MW

A2: 2 MW

A3: 2 MW

F,

A=146.9 m
2

A1: 1.71 MW

A2: 1.74 MW

A3: 2.61 MW

HE between 

units

F,

A=270.4 m
2

A1: 41.38 °C

A2: 52.67 °C

A3: 51.18 °C

A1: 1.9 MW

A2: 1.96 MW

A3: 1.88 MW

F, 

A=167.2 m
2

A1: 114.3 °C

A2: 117.45 °C

A3: 117.53 °C

A1: 2.89 MW

A2: 2.05 MW

A3: 2.52 MW

HE between 

units

F,

A=258.4 m
2

A1: 92.17 °C

A2: 105.5 °C

A3: 101.4 °C

A1: 106 °C

A2: 102.52 °C

A3: 88.36 °C

F,

A=94.8 m
2

A1: 1.59 MW

A2: 1.63 MW

A3: 2.42 MW

HE between 

units

A1: 4.38 MW

A2: 4.71 MW

A3: 5.59 MW

F,

A=735.2 m
2

A1: 107.45 °C

A2: 110.73 °C

A3: 97.48 °C

A1: 2.62 MW

A2: 1.84 MW

A3: 2.16 MW

U,

A=89.3 m
2

A1: 168.52 °C

A2: 172.58 °C

A3: 173.14 °C

A1: 112.62 °C

A2: 114.43 °C

A3: 101.48 °C

Preheated from, 10 % losses are assumed when TS Integration 

Savings of 

MPS

A1: 4.22 MW

A2: 3.46 MW

A3: 4.57 MW

Savings of 

CW

A1: 6.6 MW

A2: 5.79 MW

A3: 7.13 MW

 

Figure 4: HEN of TS operating under varying conditions with 5 modifications 

Acknowledgments 

The authors acknowledge the financial support from EC FP7 project ENER/FP7/296003/EFENIS „Efficient 

Energy Integrated Solutions for Manufacturing Industries – EFENIS‟, and from the Slovenian Research 

Agency (programs P2-0032 and P2-0377).  

References 

Čuček L., Kravanja Z., 2014, Efficient Transshipment-Based Framework for Energy Targeting and 

Retrofitting Industrial Total Sites. Chemical Engineering Transactions, 39, 1813-1818 

Čuček L., Kravanja Z., 2015, Retrofitting of Large-Scale HENs within Total Sites under Uncertainty by 

considering Trade-Offs between Investment and Operating Costs. Chemical Engineering Transactions, 

45 

Čuček L., Mantelli V., Yong J.Y., Varbanov P.S., Klemeš J.J., Kravanja Z., 2015, A Procedure for the 

Retrofitting of Large-Scale Heat Exchanger Networks for Fixed and Flexible Designs applied to Existing 

Refinery Total Site. Chemical Engineering Transactions, 45 

Klemeš J.J., Kravanja Z., 2013, Forty years of Heat Integration: Pinch Analysis (PA) and Mathematical 

Programming (MP). Current Opinion in Chemical Engineering, 2, 461-474 

Papoulias S. A., Grossmann I. E., 1983, A structural optimization approach in process synthesis – II: Heat 

recovery networks, Computers & Chemical Engineering, 7(6), 707-721 

Perry S., Klemeš J., Bulatov I., 2008, Integrating waste and renewable energy to reduce the carbon 

footprint of locally integrated energy sectors. Energy, 33, 1489-1497 

Soršak A., Kravanja Z., 2004, MINLP retrofit of heat exchanger networks comprising different exchanger 

types. Computers & Chemical Engineering, 28, 235-251 

Yee T. F., Grossmann I. E., 1991, A screening and optimization approach for the retrofit of heat-exchanger 

networks, Industrial & Engineering Chemistry Research, 30(1), 146-162