Analysis and revamping of heat exchanger networks …...to design the heat exchanger network. Papoulias and Grossmann (1983) developed a mathematical model for the same purpose. Both

ORIGINAL PAPER

Analysis and revamping of heat exchanger networks for crude oilrefineries using temperature driving force graphical technique

Dina A. Kamel1 • Mamdouh A. Gadalla1 • Fatma H. Ashour2

Received: 5 January 2017 / Accepted: 24 July 2017 / Published online: 31 July 2017

� Springer-Verlag GmbH Germany 2017

Abstract The refining industry is an energy-intensive

industry; most of the energy is consumed in heating and

cooling requirements. Revamping or retrofitting of existing

chemical engineering plants is an appropriate opportunity

for improving and enhancing the existing heat exchanger

network’s (HEN) design and performance. Revamping is

commonly used to modify the existing process for many

objectives, such as saving energy, reducing the environ-

mental emissions and increasing the productivity of the

plant. This paper presents a new graphical approach for the

analysis and revamping of existing HENs based on pinch

analysis rules. The HEN is represented on a simple graph,

where the cold stream temperatures are plotted on the

X-axis, while the temperature driving forces for each

exchanger are plotted on the Y-axis. This graphical tech-

nique describes the energy analysis problems in terms of

temperature driving force (TDF) inside the heat exchanger,

which is an important factor in the revamping process as

the difference in that temperature driving force is involved

in the calculation of the area of heat exchangers, conse-

quently affecting the cost. Also, each exchanger is repre-

sented in this graph as a straight line whose slope is related

to the heat capacities and length is related to the heat flow.

The TDF graphical approach is applied to an existing HEN

in an Egyptian crude oil refinery. The detailed steps for the

graphical analysis and revamping are applied on the HEN

for the objective of energy savings, and obtained results

showed savings of approximately 10.5% in the energy

demand with minor structural modifications and equivalent

energy cost savings of 360,000 $/year.

Keywords Graphical revamping � Temperature driving

force � Heat exchanger networks � Pinch analysis � Energysavings

List of symbols

Cpc Specific heat of cold stream (kJ/kg �C)Cph Specific heat of hot stream (kJ/kg �C)CPc Heat capacity flow for cold stream (kJ/s �C)CPh Heat capacity flow for hot stream (kJ/s �C)L Length of exchanger line (m)

mc Mass flow rate of cold stream (kg/s)

mh Mass flow rate of hot stream (kg/s)

PA Pump-around

H Hot stream

C Cold stream

Q Heat duty or flow

Tc Temperature of process cold stream in general; s

for inlet and t for outlet (�C)Th Temperature of process hot stream in general; s

for inlet and t for outlet (�C)Ths Temperature of inlet hot stream to exchangers

(�C)Tht Temperature of outlet hot stream from exchangers

(�C)Tcs Temperature of inlet cold stream to exchangers

(�C)Tct Temperature of outlet cold stream from

exchangers (�C)Thp Hot pinch temperature (�C)Thup Hot utility pinch temperature (�C)Tcp Cold pinch temperature (�C)

& Dina A. Kamel

[email protected]

1 Department of Chemical Engineering, The British University

in Egypt, El-Shorouk City, Cairo 11837, Egypt

2 Department of Chemical Engineering, Cairo University,

Giza 12613, Egypt

123

Clean Techn Environ Policy (2018) 20:243–258

https://doi.org/10.1007/s10098-017-1403-4

http://crossmark.crossref.org/dialog/?doi=10.1007/s10098-017-1403-4&domain=pdf

http://crossmark.crossref.org/dialog/?doi=10.1007/s10098-017-1403-4&domain=pdf

https://doi.org/10.1007/s10098-017-1403-4

Tcup Cold utility pinch temperature (�C)DThe Hot end temperature driving force (�C, K)DTce Cold end temperature driving force (�C, K)DT Temperature driving force of exchanger (�C, K)DTmin Minimum temperature approach difference (�C,

K)

DTc Temperature difference of cold stream (�C, K)DTh Temperature difference of hot stream (�C, K)Tci Intermediate cold temperature (�C)

Introduction

Chemical plants, especially crude oil refineries, consume

huge amounts of energy for the purpose of heating of the

raw crude oil. Revamping or retrofitting of existing

chemical engineering plants is an appropriate opportunity

for improving and enhancing the existing heat exchanger

network’s (HEN) design and performance. The crude oil is

processed in several units to achieve its final products as

kerosene, gasoline and other important products. At the

first stage of the processing, it enters an atmospheric dis-

tillation tower at a relatively high temperature of approx-

imately 360 �C, and this temperature may vary according

to the refinery’s conditions.

Raw crude oil is heated initially in a preheat train and

then enters the furnace, where most of the energy is con-

sumed. In order to decrease the energy consumption in the

furnace and consequently decrease the overall energy

consumption in the refinery, it is required to maximize the

temperature of the crude oil before entering the furnace by

increasing the efficiency of the preheat train. The preheat

train is a typical heat exchanger network (HEN), where

several heat exchangers are connected together for the aim

of heat transfer. Using energy integration techniques to

make use of the available hot and cold streams within the

plant can lead to substantial reduction in energy require-

ments and reduce utility cost of the process (Akpa and

Okoroma 2012).

Process integration is an approach that considers the

process as one unit and considers the interaction between

the process units rather than optimizing them separately.

Energy integration, in process, is done by using the heat

from streams or units that are required to be cooled to heat

other streams or units that requires to be heated (Papoulias

and Grossmann 1983). Process integration is widely used

for bench marking the process performance and for gen-

erating process improvements (El-Halwagi 2017). Such

integration takes place in the preheat train for the crude oil

to reduce the energy consumption on the furnace.

Pinch analysis is a very important method of process

integration; it was developed in late 1970s by Hohman

(1971), Linhoff and Flower (1978), Linnhoff and

Hindmarsh (1983) and Umeda (1983). The pinch analysis

aims to identify the opportunities of heat recovery between

different sources and sinks in the plant. The heat source is a

hot stream that produces heat and at the same time is

required to be cooled, and on the other hand the heat sink is

a cold stream that requires to be heated.

Hohman (1971) developed a graphical technique to

obtain the minimum energy requirements for a process and

to design the heat exchanger network. Papoulias and

Grossmann (1983) developed a mathematical model for the

same purpose. Both models were used to build up a tool

that is used nowadays in the analysis and the design of heat

exchanger networks (HENs) in complex plants.

Smith (2005) and Klemes (2013) provided the complete

steps for the design of the composite curves. They obtained

the maximum achieved energy recovery for a given mini-

mum temperature difference (DTmin) for a certain process,

based on pinch analysis rules.

Following the pinch analysis rules and applying them in

the revamping process of the existing HEN can be very

profitable and can save up to 35% of the energy consumed

(Natural Resources 2003). Retrofit design has to reuse

efficiently the existing equipment with minimum addition

of new units to improve the overall performance.

For the last 40 years many researches have worked on

developing the techniques of revamping of existing HENs.

Asante and Zhu (1997) were able to establish a new

approach for revamping of HENs known as the network

pinch approach. In this technique, the main restrictions for

efficient heat integration were recognized and accordingly

the proper modifications were recommended for the

improvement in the heat integration within the existing

network. This technique was modified by Bakhtiari and

Bedard (2013) to take into consideration the effect of

stream splitting and segmentation within the network.

Zhang et al. (2013) developed a semi-rigorous distilla-

tion model, along with a heat exchanger model. The two

models were simultaneously used to reduce both the

environmental emissions and the energy consumption for

the existing crude oil distillation system. Their work aimed

to obtain the optimum design of the existing HEN with

respect to the operating conditions, thus reducing the fuel

consumption in the furnace, and consequently, reducing the

CO2 emissions. Ochoa-Estopier et al. (2014) reviewed the

use of reduced models (i.e., simplified and statistical

models) in the optimization of heat-integrated crude oil

distillation systems. Applying a reduced model, namely

artificial neural networks (ANN), to an existing crude oil

distillation system indicated that ANN distillation model

predictions were in very good agreement with those of

rigorous models. The ANN model performed better in

terms of computation time and robustness than the rigorous

distillation model.

244 D. A. Kamel et al.

123

It is both difficult and tedious to use composite curves to

obtain the positions of the inefficiencies graphically. The

network pinch exchangers and exchangers with negative

minimum temperature difference could not be identified

graphically within the existing networks for the purpose of

revamping since the composite curves do not supply or

indicate the individual temperature of hot and cold streams,

the driving force between hot and cold streams and the

relative area of each heat exchanger (Zubairu et al. 2015).

The common method for representing the existing heat

exchanger networks is the grid diagrams. These diagrams

provide a simple representation for the existing network

regarding the inlet and outlet streams and connections, but

it cannot predict the presence of a network pinch within the

design, especially in very complex networks (Zubairu et al.

2015).

Recently, Gadalla (2015) introduced a new graphical

approach for the revamping of HENs. The HEN was rep-

resented in a (T–T) diagram by plotting the temperatures of

hot streams entering and leaving an exchanger unit versus

those corresponding temperatures of cold streams. Thus,

the Y-axis represents the temperatures of all hot streams

around exchangers, whereas the X-axis represents the

corresponding cold stream temperatures. This approach

represents each heat exchanger graphically and provides

the positions of the network inefficiencies. On the other

hand, the work does not consider neither the driving force

nor the area of each heat exchanger in the analysis.

The aim of this work is then to present a new graphical

approach for the revamping of existing heat exchanger

networks in chemical plants. The approach considers the

temperature driving force (TDF) in every exchanger which

is plotted on the Y-axis against the cold stream tempera-

tures plotted on the X-axis. The new approach provides a

visual representation for each heat exchanger, position of

inefficiencies (as network pinch exchangers, misallocated

exchangers and coolers) and temperature driving forces in

each exchanger. This will consequently provide a visual

indicative for the area of heat transfer. This approach is

applied on analysis and revamping of existing crude oil

preheat train to reduce the overall energy consumption.

New graphical approach for HEN analysis

The new graphical approach uses the basic principles of

pinch analysis. Due to the graphical nature of this

approach, only two axes of temperatures are employed to

describe the heat recovery system. For a given exchanger

unit, the driving force at both ends is plotted against the

corresponding temperature of the cold stream (entering or

leaving). Figure 1 presents the temperature profile through

a counter current heat exchanger. As shown in Fig. 1, the

driving force inside each heat exchanger has different

values in both ends and can be explained as follows: The

hot end driving force (DThe) is the difference between the

cold stream outlet temperature (Tct) and the hot stream inlet

temperature (Ths), while the cold end driving force (DTce) isthe difference between the cold stream inlet temperature

(Tcs) and the hot stream outlet temperature (Tht).

Going further, Fig. 2 depicts an X–Y diagram for the

new plot. The coordinates that identify each point on the

graph are:

• The x-coordinate is the cold stream temperature at one

end.

• The y-coordinate is the driving force across that end.

In the new analysis, exchangers are represented as

straight lines, assuming constant heat capacities for process

streams. The starting point of each exchanger is the hot end

driving force, while the end point is the cold end driving

force.

Given an exchanger line, the values of the y-coordinate

give the real values of the temperature driving force.

Fig. 1 Temperature profile through a counter current heat exchanger

Fig. 2 Graphical representation of a single heat exchanger

Analysis and revamping of heat exchanger networks for crude oil refineries using temperature… 245

123

Therefore, a horizontal line implies an equal driving force

across the exchanger unit.

Features of the new graphical approach for HEN

1. The cold stream lines are represented as vertical lines

starting from Tc at the X-axis, while the hot stream

lines are represented as inclined straight lines starting

from DT = Th on the Y-axis (Tc = zero) and ending at

Tc = Th on the X-axis (DT = zero) as shown in Fig. 3.

2. The cold pinch temperature obtained graphically from

the composite curves or by Aspen energy analyzer

(2011) is represented as a vertical line at Tc = Tcp.

3. The hot pinch temperature obtained graphically from

the composite curves or by Aspen energy analyzer

(2011) is represented as an inclined line connected

between the points (0, Thp) and (Thp, 0). All hot streams

are parallel to the hot pinch line, while all cold streams

are parallel to the cold pinch line.

4. DTmin is represented as horizontal line starting at

DT = DTmin.

5. Each point in the graph is the locus of the temperature

driving force (TDF), the hot temperature and the cold

temperature as shown in Fig. 4. Accordingly, all of the

mentioned temperatures can be easily determined

visually for each exchanger within the graph without

further calculations. As shown in Fig. 4, at both ends

of an exchanger dashed lines are extended parallel to

the hot pinch line to yield the hot supply and target

temperatures. Vertical lines are extended to yield the

cold supply and target temperatures, and horizontal

lines are extended to yield the temperature driving

force at both ends of the exchanger.

6. Assuming constant heat capacities for process streams,

each exchanger can be represented by a straight line.

This straight line is drawn between (Tcs, DTce) and (Tct,DThe) as shown in Fig. 2.

7. The slope of the exchanger is determined from Eq. (1)

as the ratio between the difference in the driving force

and the difference in the cold temperature for a certain

stream which is involved in the heat transfer process.

S ¼ DThe�DTceDTc

ð1Þ

Both the hot end and the cold end temperature driving

forces are represented in Eqs. (2) and (3), respectively,

DThe ¼ Ths�Tct ð2ÞDTce ¼ Tht�Tcs ð3Þ

By substitution in Eq. (1), the result will be Eq. (4).

S ¼ DTh�DTcDTc

ð4Þ

where DTh is the difference between the hot supply and

target temperatures, and DTc is the difference between the

cold supply and target temperatures.

The duty of the heat exchanger can be calculated by

either Eq. (5) or (6);

Q ¼ Cph � DTh ð5ÞQ ¼ Cpc � DTc ð6Þ

Since the duties calculated from the above equations are

equal for a certain exchanger, therefore Eq. (7) represents a

relation between the heat capacity flow and the temperature

difference;

CpcCph

¼ DThDTc

ð7Þ

Finally, the slope can be related to the ratio of heat flow

as seen from Eq. (8); thus by knowing the heat capacity

flows for certain hot and cold streams, a heat exchanger can

be plotted within the available heat recovery area using the

slope.Fig. 3 Graphical representations of the main lines in the new

approach

Fig. 4 Locus of every point in the graph


123

S ¼ CpcCph

�1 ð8Þ

The length of the exchanger straight line, L, in Fig. 2,

can be determined as follows:

L ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

DThe�DTceð Þ2þðDTcÞ2q

ð9Þ

From Eqs. (2) and (3) we get:

L ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

ðDThÞ2�2 DTh � DTcð Þ þ 2ðDTcÞ2q

ð10Þ

From Eqs. (5) and (6) we get:

L ¼ Q

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

1

Cph

� �2

� 2

Cph � Cpcþ 2

Cpc

� �2s

ð11Þ

Since the heat capacity flows for certain hot and cold

streams are assumed to be constant, the length of the heat

exchanger line will be proportional to its duty.

L a Q ð12Þ

8. The new graphical representation is divided into five

regions according to the pinch analysis principals as

shown in Fig. 5:

• Region 1 is below the hot pinch temperature and

below the cold pinch temperature; therefore, it is

the optimum region for heat exchangers and

coolers.

• Region 2 is above the hot pinch temperature and

below the cold pinch temperature; therefore, it is

the non-optimum region for heat exchangers as it

opposes the pinch rules (i.e, crossing the pinch).

• Region 3 is above the hot pinch temperature and

above the cold pinch temperature; therefore, it is

the optimum region for heat exchangers and

heaters.

• Region 4 is below the DT = zero line, so the

presence of any heat exchanger in this region is

infeasible (Smith 2005).

• Region 5 is below the hot pinch temperature and

above the cold pinch temperature; therefore, it is

the non-optimum region for heat integration.

9. The exchanger location in the graph gives a good

indication to its area; as the exchanger approaches the

DTmin line, its area increases as its driving force

decreases.

The aim of the retrofit process for the HEN in crude oil

refineries is to maximize the exit temperature of the cold

stream by efficiently using the existing available hot

streams in the process.

Accordingly, the retrofitted exchanger in the new

graphical representation will move to the right within the

same hot stream supply and target temperatures, so con-

sequently it will move down toward the DTmin line. So its

driving force will decrease and the area of heat transfer will

increase. In Fig. 6, exchanger 1 is retrofitted and hence

moves to the right as its cold stream inlet temperature is

raised to (Tcs2). The hot stream inlet and outlet tempera-

tures are fixed to maintain fixed duty within the same

exchanger.

For a fixed duty, the area and the logarithmic mean

temperature difference (LMTD) are inversely proportional.

So, as the LMTD decreases, the area increases as repre-

sented in Eq. (13). This is obvious from the graph; the area

of exchanger 1 is lower than that of exchanger 2 since

exchanger 2 is closer to the DTmin line (i.e., smaller

driving force).

Q ¼ U � A� LMTD ð13Þ

Fig. 5 Graphical representation of feasible regions for heat

integration

Fig. 6 Relation between the area of a heat exchanger and its position

in the graph


123

where Q = same duty of the exchanger; U = heat transfer

coefficient; A = area of the exchanger; and

LMTD = logarithmic mean temperature difference.

10. One of the features of the new graphical approach is

the instantaneous spotting of the misallocated heat

exchangers and the presence of network pinch, as

shown in Fig. 7. Both exchangers E1 and E2 are

misallocated since neither E1 nor E2 exists in

highlighted region 1 or 3, while exchanger E3 is a

network pinch exchanger as E3 is touching the DTmin

line. A network pinch is the position in the network

where the driving force inside the exchanger equals

to DTmin. It is known that the network pinch limits

the heat recovery for the existing HEN. So obviously

there are no further revamping opportunities for this

particular exchanger.

11. Exchangers including condensing steam or exhibit-

ing reboilers are represented differently than men-

tioned above. For condensing streams, such as

saturated steam, exchangers including such streams

will be presented as lines coinciding on the hot

streams since the stream inlet and outlet tempera-

tures are the same (see Fig. 7). Such exchangers will

heat some cold streams from their supply to target

temperatures through the latent heat of condensation

of saturated steam. For these exchangers, Eq. (1)

will result in a value of one for the exchanger line

slopes. Exchangers including evaporated steam on

the other hand are represented as vertical lines

coinciding with the cold stream temperatures. For

this case, the stream supply and target temperatures

are equal (see Fig. 7). Therefore, the slope of lines

representing these exchangers is infinity (Eq. 1).

12. Graphical identification of utility paths within the

network, where heat duty can be transferred across

the utility paths to minimize the utility consumption

is yet another feature of the new graphical approach.

These paths can be determined graphically by

plotting the hot temperature line of the stream

entering the cooler to identify whether there is an

exchanger that exists directly before the cooler by

having a common hot temperature, then tracing the

cold stream graphically using its cold temperatures

(Tc) (vertical dashed lines), until it eventually

reaches a heater. As shown in Fig. 8, exchanger E1

lies in a utility path with the cooler C1 and the heater

H1. Both exchanger E1 and cooler C1 have a

common hot temperature, and the cold stream

existing from exchanger E1 passes through a heater.

On the other hand, if the cold stream is traced

graphically and did not reach a heater, then it cannot

be considered within a utility path. In general,

exchangers connecting cold streams/hot streams with

utilities are among utility paths. For example, in

Fig. 8 exchanger E2 is not within a utility path

because it only connects a cold stream with a heater

with the absence of connecting a hot stream with a

cooler. Unlike-wise E1 connects a hot stream with a

cooler and also merges a cold stream with a heater.

Also, the loops can be observed from the graph, where

heat can be shifted around the loop to relax the temperature

driving forces. The change in heat duties around the loop

maintains the network heat balance and the stream supply

and target temperatures (Smith 2005). However, around the

loop the temperatures change and hence the temperature

differences of the exchangers in the loop change as well in

addition to their duties. Figure 9 presents two loops to

illustrate the concept, and the first loop lies between

exchangers E9 and E11, where both exchangers have a

common hot temperature and at the same time exist within

Fig. 7 Different positions for heat exchangers Fig. 8 Graphical representation of utility paths


123

the same cold stream. Similarly, the second loop exists

between exchangers E8 and E10. A loop of exchangers can

be identified from the TDF as: (1) exchanger units are

bounded between at least one hot stream and one cold

stream. (2) On hot stream side, the exchangers share one

common temperature (or more) that is intermediate

between the hot supply temperature and one intermediate

hot temperature. On cold side, exchangers share common

temperatures with themselves or with other exchangers not

in the same loop. It must be noted that the common tem-

perature must lie on one side, either hot or cold stream. In

some cases, the common temperatures lie on the hot and

cold streams at the same time. In general, loops are iden-

tified by inspecting graphically the inclined temperatures

and vertical temperatures around hot streams and cold

streams.

Identifying both utility paths and exchangers’ loops is

important for two reasons. The first reason is that the utility

consumption can be minimized through removing/shifting

energies from cold/hot utility to existing exchangers via

utility paths. The second is that additional areas can be also

minimized by relaxing heat between exchangers in loops.

13. The optimum position for steam generation is the

highlighted area between the process pinch and the

utility pinch shown in Fig. 10. This area can be

visualized from the graph to select the appropriate

streams that can be utilized in the process of steam

generation. As shown in Fig. 10, excess heat should

be removed from the process by generating steam in

the highlighted area between the process pinch and

the utility pinch. Below the utility pinch temperature,

excess heat has to be removed using cooling water.

The vertical part of the heat exchanger represents the

latent heat, while the inclined part represents the heat

transfer between the cold stream and the steam.

The new approach will allow the graphical revamping of

a HEN by visual identification of the misallocated coolers,

exchangers, paths and loops for the existing HEN, sug-

gesting appropriate steps for the revamping procedure.

Illustrative example: energy analysis of crude oil

plant

The objective of this example is to demonstrate the use of

the TDF graphical approach to analyze the performance of

the existing network graphically regarding the energy

efficiency. The data of this crude oil refinery are obtained

from the literature (Rossiter 2010); this case is selected to

show the applicability of the new approach in the analysis

of HEN.

Process description

The process base case is a 90,000-bbl/d crude distillation

unit (CDU) that includes both atmospheric and vacuum

towers. Figure 11 gives details of the crude preheat train.

The atmospheric and vacuum towers have pump-around

circuits; these pump-arounds provide a mechanism for

removing heat from the towers at intermediate temperature

levels, rather than taking all of the distillation heat out in

the overheads. Most of the pump-around heat is used to

preheat the crude feed, and excess heat from the pump-

arounds is either used to generate steam or rejected to

cooling water.

The existing energy demand in this process is

200 MBtu/h for the hot utility consumption and

188.2 MBtu/h for cold utility consumption. Also, the

energy targets for the hot and cold utilities are 166.2 and

154.4 MBtu/h, respectively. Therefore, the potential

reduction of energy consumption is 33.8 MBtu, which is

equivalent to 17%.

Fig. 9 Graphical representation of loops

Fig. 10 Steam generation region


123

Graphical analysis of the existing crude oil plant

The new TDF graphical representation is employed to

describe and analyze the process base case of the crude oil

plant in light with the previously mentioned features. By

using the data from Fig. 11 of the inlet and exit tempera-

tures for each exchanger, the TDF graph presented in

Fig. 12 is plotted. As shown in Fig. 12, the network has 9

process to process heat exchangers and 2 steam generation

exchangers (E10, E11); the network has 6 coolers and 1

heater (furnace).

The analysis of the HEN can be easily visualized from

Fig. 12, where exchanger E10 is obviously generating

steam across the pinch; it exists in region 2, above the hot

pinch temperature and below the cold pinch temperature,

which is non-optimum region for heat integration. The

amount of heat crossing the pinch is calculated from

Eq. (5), where DT is the difference between the hot supply

Fig. 11 HEN details (Rossiter 2010)

Fig. 12 Graphical analysis of

the preheat train using TDF

approach


123

temperature and the hot pinch temperature and equals to

23.9 MBtu/h.

Similarly, part of exchanger E11 is crossing the pinch;

the vertical part above the hot pinch temperature, the

amount of energy crossing the pinch is calculated from

Eq. (5), where DT is the difference between the hot supply

temperature and the hot pinch temperature and equals to

4.9 MBtu/h.

Part of exchanger E5 is crossing the pinch, and the

amount of energy crossing the pinch is calculated from

Eq. (5) and equals to 2.86 Mbtu/h.

Finally, exchanger E4 is crossing the pinch by a small

amount about 1 Mbtu/h.

To conclude the analysis of this case, the total amount of

energy crossing the pinch is 33.8 MBtu/h, which is the

exact difference between the target and the existing hot

demand.

Another feature that can be highlighted in this case is the

steam generation region which lies between the utility

pinch temperature and the process pinch, as represented in

Fig. 12. As noticed exchanger E4, part of exchangers E3

and E5 lies within thin region. For optimal operations,

there duties should be utilized for steam generation rather

than process to process heat transfer.

The analysis obtained by the TDF graphical approach

completely agrees with the analysis previously presented in

the literature.

Revamping of heat exchanger networks HENs

The new approach together with its valuable features will

allow the graphical revamping of a HEN by visual identi-

fication of the misallocated coolers, exchangers, paths and

loops for the existing HEN, suggesting appropriate steps

for the revamping procedure. A systematic procedure for

revamping existing heat exchanger networks train is pre-

sented in the form of algorithm in Fig. (13). The procedure

starts by plotting the TDF for a given HEN and then ana-

lyzing the performance of the existing HEN against the

features previously defined. The analysis step will result in

the identification of energy inefficiencies present in the

existing HEN. The procedure also involves a step of net-

work modifications or relocation/addition of exchangers

and additional area calculations. The procedure follows up

to determine the amount of energy savings, total capital

costs required for revamping.

Case study: energy analysis and revampingof HEN

Existing crude unit description

In the existing crude unit presented in Fig. 14, the crude oil

is heated from 25 to 261 �C by exchanging heat with

Fig. 13 Algorithm for the complete steps of graphical revamping


123

process hot streams in two preheat trains. The first train

consists of 5 heat exchangers after which the temperature

of the crude reaches 116 �C. Then the crude passes througha desalter to remove inorganic salts, impurities and soluble

metals. The desalted crude then enters the second train

which consists of 6 heat exchangers. The temperature of

the exiting stream from the furnace is about 355 �C. Fueloil or fuel gas, depending on the refinery availability, is

used as the energy source. All the heat needed for the

separation is provided in the furnace, so no reboiler is

present in the main column.

The high temperature difference between the inlet and

the outlet streams of the furnace and the high flow rate of

the crude processed make the main crude furnace one of

the highest energy consumers of the whole refinery. The

duty of the main furnace is 52.25 MMkcal/h (60.7 MW).

Therefore, the cost of this unit is a significant part of the

overall production costs.

Moreover, a secondary furnace is used to heat the main

steam line. This furnace consumes 5.25 MMkcal/h

(6.1 MW), so the overall energy consumption for heating

purposes is 66.8 MW.

The cost of energy used in the furnace can be calculated

from the data available from the refinery. Each 1 MMkcal/

h costs 4982 $/month; therefore, the total cost of the energy

used in the two furnaces is 3,437,580 $/year. This value

will be used later in calculations of energy savings.

As shown in Fig. 15, the crude oil stream is heated in 11

heat exchangers against different streams. The product

streams of the atmospheric distillation are in exchangers

E1, E4 and E7. The streams cooled in the pump-arounds to

be used as cold reflux for the atmospheric distillation col-

umn in exchangers E5, E6 and E7. The streams cooled in

the pump-arounds to be used as cold reflux for the atmo-

spheric distillation column in exchangers E2, E8, E10 and

E11. The reboiler of the kerosene side stripper is used to

cool a product stream in exchanger E14 and, finally, with

the reflux from the condenser in exchanger E3.

Graphical energy analysis of EORU heat exchanger

network

According to the algorithm presented in Fig. 13, the first

step in the analysis process is to represent the network on

the TDF graph. Using the data presented in Table 1 for the

supply and target temperatures, the graph shown in Fig. 16

is plotted between the driving force for heat transfer on

Y-axis and the temperatures of the cold streams on X-axis.

The energy targets, the hot pinch temperature and the

cold pinch temperature are calculated using Aspen energy

analyzer (2011). Based on a minimum temperature differ-

ence of 20 �C, the target hot utility required is 54.96 MW

and the target cold utility required is zero. Also the hot

pinch and the cold pinch temperatures are found to be 40

and 20 �C, respectively.The existing hot duty required for the heat exchanger

network is calculated either by using Eq. (13) or graphi-

cally using the relationship between the length of the heater

and the duty in Eq. (11) which is 66.8 MW. Thus, an

excess of fuel oil of approximately 21.5% is consumed.

Similarly, the cold duty required for this HEN is

11.97 MW.

Fig. 14 Crude oil distillation

configuration


123

The following step is to analyze the existing network

graphically based on the presented features in ‘‘Features of

the new graphical approach for HEN’’ section. From

Fig. 16, the 4 coolers are obviously misallocated because

all the coolers are present above the hot pinch line and

below the cold pinch line. This is translated to region 2 in

Fig. 15 HEN configuration

Table 1 Changes in the

exchangers after revampingExchanger Ths (�C) Tht (�C) Tcs (�C) Tct (�C) Duty (MW) Area (m2)

E1 127.4 56.4 20 33.3 4.44 236.21

E10 127.4 40 20 36.5 5.4712 420.6

E4 323.45 197.8 86 103.4 5.741 129.17

E40 323.45 109.2 89.2 119.2 9.93 450

E13 346.4 55 20 94.1 2.235 73.15

E130 346.4 40 20 97 2.347 93.25

Fig. 16 TDF of the existing

HEN


123

Fig. 5. The optimum position of the coolers in any HEN is

located within region 1, between the supply and target

temperature of the cold utility.

As noticed from the graphical representation of the HEN

in Fig. 16, all the heat exchangers exist in the third region

which fulfills the pinch rules. There are no network pinch

exchangers in this process, although exchanger E14 is very

near to the DTmin line since its hot end temperature dif-

ference equals to 21 �C and its cold end temperature dif-

ference equals to 22 �C.

Graphical conceptual revamping

This HEN has 3 utility paths illustrated in Fig. (16). The

first path connects between exchanger E1, cooler C4 and

the main crude furnace. The second path connects between

exchanger E4, cooler C1 and main crude furnace. The last

path connects between exchanger E13, cooler C3 and

secondary furnace.

The revamping of this particular HEN will be focused

on moving the misallocated coolers to an appropriate

position. This will be achieved by applying the two

graphical revamping methodologies presented in the algo-

rithm in Fig. 13; the first case is HEN revamping without

structural modification for cooler existing in utility paths to

decrease the duty consumed by the coolers and conse-

quently decrease the overall energy consumption in the

HEN.

The second case is HEN revamping by adding a com-

pletely new heat exchanger to the network which will

require some structural modifications. This exchanger will

be added to replace some or all of the duty of the cooler in

case of the absence of a utility path.

Revamping of coolers existing within a utility path

This revamping procedure can only be used in the presence

of a utility path, so it will be applied on the three existing

utility paths in the HEN. The order of the heat exchangers

in the HEN is considered while revamping, so the path of

exchanger E1 is considered first, then the path of exchanger

E4 and finally the path of exchanger E13.

The first utility path between exchanger E1, cooler C4

and the main crude furnace is shown in Fig. 17. Cooler C4

is cooling the crude oil in exchanger E1 from 56.4 to

40 �C, with a duty of 1.15 MW.

The minimum target temperature that the hot stream can

reach is obtained graphically at the intersection between

the cold supply temperature (20 �C) and the DTmin line that

is 40 �C. This point represents the new starting point of the

modified exchanger is shown in Fig. 17. After that, the

modified exchanger is plotted using the starting point and

the slope till it intersects the Ths line (127.4 �C line).

The new cold target temperature is 36.5 �C instead of

33.3 �C, while the intermediate cold temperature is

23.5 �C, as shown in Fig. 18. All the duty of cooler C4 is

recovered by the modified exchanger E10, the modified

exchanger is closer to the DTmin line, and this incurs an

additional area of 184.3 m2. The cost of the additional area

is calculated from Eq. (14) for year 2003 (Gadalla 2003) as

follows:

Heat exchanger area cost $ð Þ¼ 1530� additional areasð Þ0:63

ð14Þ

The cost calculated from the equation is multiplied by

the cost index for 2014 (Chemical engineering magazine

2014), so the cost of additional area is found to be $40,936

For the second utility path, between exchange E4, cooler

C1 and the main crude furnace, the cooler C1 is cooling the

same hot stream existing in exchanger E4 from

Ths = 197.8 �C to Tht = 40 �C with a duty of 5.93 MW, as

Fig. 17 Graphical determination of the new starting point of

modified E1

0

20

40

60

80

100

120

0 5 10 15 20 25 30 35 40

∆T

Tc

E 1

C4

Cold pinch- Tcs

∆T min

Tct =33.3 ˚C

New starting point

Fig. 18 Graphical presentation of the modified exchanger E1


123

shown in Fig. 19. Although all the duty of cooler C1 can be

conceptually recovered in the exchanger E4, only a fraction

of this duty is investigated graphically by following the

same steps used in the first path.

The new cold target temperature is 119.1 �C instead of

106.4 �C; the intermediate cold temperature is 98.5 �C. Itis obvious from Fig. 20 that not all the duty of the cooler

can be transferred to the exchanger since the maximum hot

target temperature is 109.2 �C which cannot comply with a

DTmin of 20 �C.Therefore, the maximum duty that can be transferred

from the cooler C1 to the exchanger E4 can be easily

calculated either from Eq. (13) or graphically using

Eq. (11). The transferred duty equals to 4.189 MW, and

this duty will be used in heating the cold stream to a higher

temperature. So the new exchanger E40 will be able to heat

the crude oil from 89.2 to 119.2 �C with a duty of

9.93 MW.

The additional area is calculated to be 320.9 m2 with a

cost of $83,586.

Similarly, for the third and last path, the same graphical

procedure is applied between E13 and C3, as shown in

Fig. 21. The maximum hot target temperature is 55 �C.The new cold target temperature obtained graphically

from Fig. 22 is 97.8 �C instead of 94.1 �C, while the

intermediate cold temperature is 24.5 �C. The recovered

duty can be calculated as above and equals to 0.113 MW.

The additional area is 20 m2 and costs $10,132.

Table 1 summarizes the changes in E1, E4 and E13,

with regard to the terminal temperatures, the duties and the

areas.

After applying these additional area modifications, the

overall energy recovery is 5.33 MW.


modified E4

Fig. 20 Graphical presentation of the modified exchanger E4 and

modified C1


modified E13

Fig. 22 Graphical presentation of the modified exchanger E13


123

Revamping with structural modifications

This revamping technique is applied on the cooler C2

which has a duty of 4.817 MW and does not exist within a

utility path. Therefore, a new exchanger will be introduced

to the HEN in order to recover part or all of the cooler’s

duty.

The position of the new exchanger in the HEN is

investigated graphically. Since the new exchanger is

intended to replace the cooler C2 (which is placed on the

brine to crude stream), the supply and target hot tempera-

ture boundaries for the new exchanger are hot supply

temperature (Ths) and hot target temperature (Tht) of 127.5

and 43.9 �C, respectively.As for the cold streams, it is more convenient to add the

new exchanger to the main steam line to reduce the duty of

the existing furnace without numerous modifications in the

existing HEN.

The new heat exchanger has been introduced after

exchanger E13 with a cold supply temperature equal to the

cold target temperature of exchanger E130 which is 97 �C.The new heat exchanger is drawn using the starting

point at the intersection between the hot target temperature

(Tht = 117 �C) and the cold supply temperature

(Tcs = 97 �C). The slope is calculated from Eq. (8) where

Cpc = 0.883 MW/�C, Cph = 0.156 MW/�C. The line is

extended till it intersects with either the maximum cold

target temperature or the hot supply temperature, which-

ever comes first.

Figure 23 shows that the new exchanger E16 will

intersect with the hot supply temperature (127.5 �C) first.The duty recovered by the new exchanger is only

1.68 MW, due to the DTmin constraints. The area of the

new exchanger is 251 m2 and will cost $101,381.

It can be observed that the new exchanger unit will incur

additional costs relatively higher than those required by the

additional areas. This means the modification of the new

exchanger may not be attractive from the point of view of

operation implications.

0

50

100

150

200

250

0 20 40 60 80 100 120 140

∆T

TC

98.8 ˚C97 CCold pinch

127.5˚C

117˚C

43˚C

Modifiedcooler C2

New heat exchanger

E13’

110 ˚C

Heat recovery area

Fig. 23 Graphical representation of the new exchanger after E13

Fig. 24 TDF for the modified

HEN


123

Results

Figure 24 shows the TDF graphical representation of the

modified HEN. The final HEN requires the presence of 2

coolers only with duties of 6.81 MW instead of 4 coolers

with duties of 11.9 MW. The revamping of coolers using

the utility paths technique shifted all exchangers after E1 to

the right, so the temperature of crude oil before the main

furnace is increased from 260.9 to 270.3 �C. A new

exchanger E16 is added after exchanger E13 to recover part

of the duty of cooler C2.

Table 2 reveals significant savings in energy demands

and fuel oil consumption, versus the cost of the modifica-

tions and payback period. It is clearly seen that the pro-

posed modifications by the new graphical representation

improve the heat integration of the existing refinery.

Conclusions

A new graphical technique has been proposed for the

revamping and analysis of the existing heat exchanger

networks. This graphical technique is based on pinch

analysis principles. The existing HEN is represented

graphically where the driving force of the exchanger is

plotted on the Y-axis and the cold temperature of the

streams is plotted on the X-axis. The graph describes

energy analysis problems in terms of the driving force

inside the heat exchanger. The driving force is an important

factor in the revamping process as the difference in that

driving force is involved in the calculation of the area of

heat exchangers, consequently affecting the cost. The new

technique completely follows the pinch analysis rule, with

the added advantage of being simple since each exchanger

is represented by a straight line. This visual technique is

indicative as both the length of the line and its slope have

important physical meanings. Each exchanger position in

the HEN provides its relative area; also the feasible regions

for heat integration are easily identified visually.

A case study showed the application of the graphical

technique in the revamping of a heat exchangers’ network,

by both structural and non-structural methods, and the

results are presented in Table 2.

The potential for further energy recovery is still possible

provided that the existing structure of the HEN is to be

changed. This will incur substantial modification costs that

may lead to an uneconomical solution. However, the dis-

tillation process is still a key factor for further savings; this

study is underway and will be considered in further

publications.

References

Akpa JG, Okoroma J (2012) Pinch analysis of heat exchanger

networks in the crude distillation unit of Port-Harcourt refinery.

J Emerg Trends Eng Appl Sci (JETEAS) 3(3):475–484

Asante NDK, Zhu XX (1997) An automated and interactive approach

for heat exchanger network retrofit. Trans IChemE 7(3):275–349

Aspen energy analyser (2011) Aspen technology, V.7.3. USA:

University program

Bakhtiari B, Bedard S (2013) Retrofitting heat exchanger networks

using a modified network pinch approach. Appl Therm Eng

51(1):973–979

Chemical engineering magazine (2014) 121(12):36. www.chemengon

line.com

El-Halwagi M (2017) A return on investment metric for incorporating

sustainability in process integration and improvement projects.

Clean Technol Environ Policy 19(2):611–617

Gadalla M (2003) Retrofit design of heat integrated crude oil

distillation systems. UMIST, Manchester

Gadalla M (2015) A new graphical method for Pinch Analysis

applications: heat exchanger network retrofit and energy inte-

gration. Energy 81:159–174

Hohman EC (1971) Optimum networks of heat exchange. Ph.D.

thesis, University of Southern California, USA

Klemes J (2013) Handbook of process integration (PI). Minimisation

of energy and water use, waste and emissions, 1st edn.

Woodhead Publishing Limited, Cambridge

Linnhoff B, Flower J (1978) Synthesis of heat exchanger networks: I.

Systematic generation of energy optimal networks. Alche

24(4):633–642

Linnhoff B, Hindmarsh E (1983) The pinch design method for heat

exchanger networks. Chem Eng Sci 38(5):745–763

Natural Resources (2003) Pinch analysis: for the efficient use of

energy, water and hydrogen. CANMET,. Canada: ISBN: 0-662-

34964-4, Catalogue # M39-96/2003E

Ochoa-Estopier LM, Jobson M, Smith R (2014) The use of reduced

models for design and optimisation of heat-integrated crude oil

distillation systems. Energy 75:5–13

Papoulias SA, Grossmann I (1983) A structural optimization approach

in process synthesis—II: heat recovery networks. Comput Chem

Eng 7(6):707–721

Table 2 Summary of the results

Case Duty of the furnaces

(MW)

Add. area of HEN

(m2)

Energy cost ($/

year)

Energy saved

(%)

Cost of

modification ($)

Payback period

(year)

Base case 66.8 – 3,437,580 – – –

With additional area only 61.5 957.5 3,163,356 8 114,718 0.41

With add. area and new

exchanger

59.8 1208 3,077,435 10.5 224,105.8 0.62


123

http://www.chemengonline.com

http://www.chemengonline.com

Rossiter AP (2010) Improve energy efficiency via heat integration.

Chem Eng Progress 106:33–42

Smith R (2005) Chemical process design and integration. Wiley,

England

Umeda T (1983) Computer aided process synthesis. Comput Chem

Eng 7(4):279–309

Zhang N, Smith R, Bulatov I, Klemses JJ (2013) Sustaining high

energy efficiency in existing processes with advanced process

integration technology. Appl Energy 101:26–32

Zubairu B, Highina BK, Gutti B (2015) Minimizing hot and cold

utility requirements for vegetable oil refinery plant using pinch

Analysis. Int J Sci Eng Technol 4(5):298–301


123

Documents

Analysis and revamping of heat exchanger networks …...to design the heat exchanger network. Papoulias and Grossmann (1983) developed a mathematical model for the same purpose. Both