chemical engineering research and design 8 7 ( 2 0 0 9 ) 1509–1519

Contents lists available at ScienceDirect

Chemical Engineering Research and Design

journal homepage: www.elsevier.com/locate/cherd

A new technique for simultaneous water and energy

minimisation in process plant

Zainuddin Abdul Manan ∗ , Swee Yin Tea, Sharifah Rafidah Wan Alwi
Process Systems Engineering Centre, Faculty of Chemical and Natural Resources Engineering, Univesiti Teknologi Malaysia, 81310 UTM
Skudai, Johor, Malaysia

a b s t r a c t

This paper presents a new technique for simultaneous minimisation of water and energy in process plants through
a combination of numerical and graphical tools. The technique consists of three steps, namely, setting the minimum
water and wastewater targets; design of minimum water utilisation network, and finally, heat recovery network
design. This technique offers two key advantages over current state-of-the art techniques. Firstly, it is applicable to
mass transfer based and non-mass transfer based water-using operations. Secondly, it introduces a new graphical
visualisation tool through a plot of temperature versus stream flowrate, termed as heat surplus diagram to guide
water and energy reduction simultaneously. The heat surplus diagram provide insights on the energy demand as
well as on stream matching scenarios during design of a maximum water and energy recovery network. A case study
on a paper mill plant demonstrates that significant reductions in water and energy consumption can be achieved
using this approach.
© 2009 The Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.

Keywords: Water minimisation; Energy minimisation; Pinch analysis; Process integration; Heat surplus diagram;
Heuristics

1. Introduction and computational effectiveness particularly when handling

complex water distribution systems involving multiple con-
Water conservation efforts have intensified over the last taminants. It is however less popular among engineering
decade due to rising costs of raw water and wastewater treat- practitioners mainly due to the difficulty to master the
ment, as well as the emphasis on pollution prevention. Even technique and to set up the problem models. In addition,
though water minimisation has been the subject of exten- mathematical approach typically provides designers with less
sive study for water-intensive industry, research concerning control over the solution space as well as little insights on
simultaneous recovery of heat from recycled process water water network design. In contrast, conceptual techniques are
has lagged behind. typically easier to master and apply and most vital as visu-
The importance of simultaneous minimisation of energy alisation tools for network targeting and design. It however
and water was first addressed by Savelski and Bagejewicz has major limitations in terms of computational effectiveness
(1997). Later studies have been based on conceptual (Savulescu and in dealing with problem dimensionality and often can-
and Smith, 1998; Zheng et al., 2003; Savulescu et al., 2002, not guarantee a global optimal solution. The two approaches
2005a,b; Feng et al., 2008) and mathematical modeling tech- are complimentary and are widely used to provide better
niques (Du et al., 2004; Liao et al., 2008; Dong et al., 2008; engineering understanding through visualisation (conceptual
Bogataj and Bagajewicz, 2008; Leewongtanawit and Kim, 2008; approaches) and to handle complex problems (mathematical
Kim et al., 2009; Xiao Feng et al., in press). modeling).
Mathematical modeling approach offers the advantage This paper presents a new conceptual technique for simul-
of accuracy and global optimality, problem dimensionality taneous reduction of water and energy in process plants which

∗
Corresponding author. Tel.: +60 7 5535609; fax: +60 7 5581463.
E-mail address: [email protected] (Z.A. Manan).
Received 14 January 2009; Received in revised form 24 April 2009; Accepted 29 April 2009
0263-8762/$ – see front matter © 2009 The Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.
doi:10.1016/j.cherd.2009.04.013
1510 chemical engineering research and design 8 7 ( 2 0 0 9 ) 1509–1519

2. Methodology
Nomenclature

C0 equipment cost of heat exchanger (USD) 2.1. Step 1: water and wastewater targeting
S0 heat transfer area
m contaminant loads (kg/h) Savulescu and Smith introduced a technique for simultaneous
Tlm Log mean temperature energy and water minimisation that focuses on heat recov-
A heat exchanger area (m2 ) ery (Savulescu and Smith, 1998; Savulescu et al., 2005a,b).
BC base cost (USD) Reduction of cooling water is achieved indirectly when the
C concentration (ppm) temperature of wastewater is reduced. In fact, for a water-
Cpinch pinch concentration (ppm) intensive industry that utilises water in processes on top
D demand/sink of those required by a cooling water circuit, more savings
DIP de-inking plant and benefits can be achieved when water reduction becomes
MF module factor the main consideration followed by heat recovery, as demon-
MNI Malaysian Newsprint Industry strated in this work. Various water targeting approaches are
MPF material and pressure factor available to generate the maximum water recovery targets
ppm parts per million applicable for both MTB and NMTB systems such as source
Q heat duty (MW) sink composite curves (El-Halwagi et al., 2003; Prakash and
RCF recycle fiber plant Shenoy, 2005) and water cascade analysis (WCA) (Manan et
S source al., 2004). In this work, the WCA is used to target the mini-
S heat transfer area of a particular heat mum water and wastewater targets due to its versatility and
exchanger capability to yield accurate results.
SWE simultaneous water and energy
TDS total dissolved solids 2.2. Step 2: water network design
TSS total suspended solid
UF updated factor Once the water and wastewater targets are established, water
WCA water cascade analysis utilisation network is designed using the improved cleanest
to cleanest rule by Wan Alwi and Manan (2008) that takes
into consideration both ‘flowrate deficit cases’ and ‘mass load
deficit cases’. In order to achieve the fresh water and wastewater
combines numerical and graphical visualisation tools. This
flowrate targets, it is necessary to observe the pinch division
approach offers two key advantages over current state-of-
and some well-established water network design rules which
the art techniques. Firstly, it is applicable to mass transfer
specify that water sources above the pinch may not be fed
based (MTB) and non-mass transfer based (NMTB) water-using
to demands below the pinch, and may not be mixed with
operations (i.e. global water operations) not considered in pre-
sources that are below the pinch concentration. Exceptions
vious studies (Savelski and Bagejewicz, 1997; Savulescu and
apply for the sources at the pinch concentration where part
Smith, 1998; Savulescu et al., 2002, 2005a,b; Zheng et al., 2003;
of these sources can be used to satisfy demands below the
Feng et al., 2008). Note that a resilient analysis tool should
pinch.
be able to handle global water operations which include the
During water network design and retrofit, it is quite com-
MTB and NMTB operations. An MTB operation involves the
mon to encounter more than one heat recovery match options.
preferential transfer of species from a rich stream to a lean
To simultaneously consider energy reduction during water
stream (a mass separating agent, MSA). A typical example of
network design, it is best to observe Heuristic SWE as follows:
such operation is ammonia used as an absorbent to recover
H2 S from a sour gas stream. On the other hand, the NMTB
operation covers function of water other than as a mass Heuristic SWE: Favour matches with closer heat load.
separating agent, including water fed as a raw material, or
being withdrawn as a product or by-product in a chemical Matching the units or streams with closer heat loads can
reaction. keep the number of heat exchanger units to a minimum.
The second contribution of this technique is introduction
of the heat surplus diagram (HSD) which is a new graphical 2.3. Step 3: heat recovery network design
visualisation tool used to guide water and energy reduction.
The heat surplus diagram, which is a plot of temperature (T) Heat recovery can be maximised and the minimum energy
versus water stream heat capacity flowrate (FCP, assumed as consumption targets can be achieved by considering direct
constant) is used together with design heuristics to provide and indirect heat transfer. The optimal water network mix-
insights on the minimum energy demand as well as on stream ing scheme is first identified to cater for direct heat transfer
matching scenarios during design of a maximum water and through isothermal mixing. This is done using the heat sur-
energy recovery network. plus diagram as described next. Then, a heat recovery system
Section 2 describes the methodology for simultaneous is designed using heat pinch design method to consider indi-
water and energy (SWE) reduction which consists of three rect heat transfer via heat exchangers.
steps, namely, setting the minimum water and wastewa- The source and demand streams are plotted on a new
ter targets, design of minimum water utilisation network graphical tool introduced in this work called “heat surplus
and heat recovery network design. The effectiveness of diagram” in order to explore the potential water and energy
the approach for energy and water reduction is demon- savings. Heat surplus diagram is a plot of stream tempera-
strated via a case study on a paper mill plant (Section ture versus water flowrate to identify the optimal isothermal
3). mixing scheme for a water utilisation network.
 chemical engineering research and design 8 7 ( 2 0 0 9 ) 1509–1519 1511

Fig. 1 – Heat surplus diagram (above-Cpinch ).

Heat surplus diagrams above and below the pinch concen-

trations are plotted separately. Figs. 1 and 2 show the demand
and source streams plotted on heat surplus diagrams above
and below the pinch concentration. The bracketed figures next
to each demand and source stream indicate the contaminant
loads, m.
Figs. 1 and 2 show regions where the source streams are
above (and below) the demand streams, respectively. Fig. 3
shows the existing heat surpluses (+) and deficits (−) on the
heat surplus diagram for regions above and below-Cpinch . If
the source is above the demand composite, a heat surplus
(+) exists (i.e. cold utility is needed to cool down the source
stream to its targeted temperature before being matched with
a demand stream). If source stream is below the demand
Fig. 3 – Existing heat surplus (+) and heat deficit (−) on heat
stream, a heat deficit (−) exists (i.e. hot utility is required to
surplus diagram: (a) heat surplus diagram (above-Cpinch ). (b)
heat up the source stream to its targeted temperature before
Heat surplus diagram (below-Cpinch ).
being matched with demand stream). The amounts of heat
surplus or heat deficit in each region are determined by calcu-
lating the area enclosed by each rectangle, multiplied by water Step 1. Note that the problem addressed here involves only
heat capacity which is assumed constant (see the shaded water streams which include stream mixing and splitting,
area in Fig. 3). Referring to the same graphs, the overall util- unlike the case of classical heat integration problem where
ity targets can be obtained by getting the difference between a mixture of fluids may exist, and only indirect heat exchange
heat surplus and heat deficit for cases above and below the is possible.
pinch regions separately. These quantities represent the heat The procedure for SWE reduction is summarised in Fig. 4.
requirements corresponding to the targeted fresh water con-
sumption. 3. Application of SWE minimisation
Finally, heat recovery can be increased further via indirect technique on a paper mill plant (Manan et al.,
heat transfer using established pinch analysis technique for 2007).
heat exchanger network design. The energy target is the mini-
mum energy consumption for the water network that achieves Many studies on water and energy minimisation have been
the targeted minimum fresh water and wastewater flowrate in done for paper mill using either graphical approach or math-
ematical programming methods (Tripathi, 1996; De Beer et
al., 1998; Parthasarathy and Krishnagopalan, 2001; Thompson
et al., 2001; Bengtsson et al., 2002a,b; Manninen et al., 2002).
This case study involves a paper mill plant in Malaysia. Water
is used in abundance for processes such as stock dilution,
washing, cleaning, steam production, etc. On the other hand,
thermal energy is used for drying processes and to generate
electricity for the plant.

3.1. Data extraction

Data extraction involved process line-tracing, establishing

process material and energy balances and specifying the limit-
ing water data in terms of water sources (outlet streams with
potential to be recycled) and water demands (inlet streams
Fig. 2 – Heat surplus diagram (below-Cpinch ). representing process water requirements). Tables 1 and 2 are
1512 chemical engineering research and design 8 7 ( 2 0 0 9 ) 1509–1519

Fig. 4 – Three-step procedure for SWE minimisation.

Table 1 – Demand streams limiting data for MNI paper mill case study.
Demand Type F (t/h) Cmax,in (ppm) Mass load of contaminant, m (kg/h) T (◦ C)

D1 Pressing 1 126 20 2.52 50

D2 Forming 1 54 20 1.08 50
D3 Chemical Preparation 36 20 0.72 81
D4 Clarified Water 1 202 20 4.03 38
D5 Deculator 1 18 20 0.36 38
D6 Pressing 2 169 100 16.92 43
D7 Clarified Water 2 1130 150 169.56 46
D8 Pressing 3 677 160 108.29 43
D9 Forming 2 155 160 24.77 43
D10 DIP 396 160 63.36 46
D11 Deculator 2 104 160 16.70 46
D12 Deculator 3 68 250 17.10 40

the extracted stream data for the paper mill process which 13.3% wastewater flowrates from the existing network’s water
include twelve demands and five sources. The data extracted consumption. Next, a maximum water recovery network was
include water flowrate, contaminant concentration and tem- designed to achieve the water targets.
perature data for source and demand streams.
Total suspended solids (TSS) measured in parts per million 3.3. Step 2: water utilisation network design
(ppm) was selected as the dominant water quality parameter
and the limiting contaminant for water re-use in this work. A water utilisation network was designed for the case study
The TSS lumps together various types of contaminants sus- using the improved cleanest to cleanest rule from Wan
pended in water into a pseudo single-contaminant system. Alwi and Manan (2008). The demands and sources listed in
Fig. 5 shows the existing process flow diagram of MNI water Tables 1 and 2 are divided into regions above and below the
utilisation network. Table 3 shows the existing heaters and pinch concentration of (Cpinch ) 150 ppm. Table 4 shows the list
cooler duties of the paper mill water utilisation network. of demands and sources for regions above and below Cpinch .
Next, the source-demand matching process was carried
3.2. Step 1: water and wastewater targeting out. Since all the matches represent the ‘flowrate deficit case’,
demands at the lowest concentration were matched with the
The first step of the SWE minimisation technique is to estab- sources at the lowest concentration. Note that S2 and S3 were
lish the fresh water and wastewater consumption targets selected to be matched with D7 since both sources existed at
using the water cascade analysis (WCA) technique by Manan the same limiting concentration. Looking at the temperatures,
et al. (2004). The WCA technique yields the minimum fresh however, S2 existed at 40 ◦ C and S3 at 46 ◦ C. According to the
water and wastewater flowrate targets of 377.5 and 377.5 t/h, new SWE heuristic, matches with closer temperature are to be
respectively and a pinch concentration at 150 ppm. The tar- preferred. In line with this heuristic, S3 was matched with D7
gets represent potential reductions of 13.3% freshwater and in order to simultaneously improve heat transfer and energy

Table 2 – Source streams limiting data for MNI paper mill case study.
Source Type F (t/h) Cmax,out (ppm) Mass load of contaminant, m (kg/h) T (◦ C)

S1 Saveall Disc Filter 2 169 100 16.92 43

S2 DAF 92 436 150 65.34 48
S3 Saveall Disc Filter 3 1130 150 169.56 46
S4 Clarified Water Tower 832 160 133.06 46
S5 Saveall Disc Filter 1 68 250 17.10 40
 chemical engineering research and design 8 7 ( 2 0 0 9 ) 1509–1519 1513

Fig. 5 – Existing water utilisation network for MNI paper mill.

3.4. Step 3: heat recovery

Table 3 – Heat exchanger duties in the water utilisation
network for MNI paper mill.
Sections 3.4.1 and 3.4.2 describe the procedures for heat recov-
Type QH (MW)
ery through direct and indirect heat transfer, respectively.
A Flash steam heating 2.68
B Condensed moisture heating 0.59
3.4.1. Direct heat transfer
C Steam 2.14
Using the procedure described in Section 2.3, the heat surplus
D Steam 2.73
diagram for the regions above and below the pinch concentra-
Total 8.14
tion (Cpinch ) were constructed once the water network design
was completed. Figs. 8–10 show the heat surplus diagram for
Type QC (MW)
the regions above and below the pinch. The overall minimum
E Cooling water 2.37 utility targets were obtained by calculating the shaded area
F Cooling tower 6.61 under each rectangle. Note that freshwater is supplied at 30 ◦ C
Total 8.98 while the target temperature for the wastewater discharged to
cooling tower is 35 ◦ C.
efficiency. Figs. 6 and 7 summarise the matches for the paper
mill water network for the regions above and below the pinch. 3.4.2. Indirect heat transfer through HEN design
The fresh water and of wastewater flowrates calculated dur- Once the isothermal mixing scenarios of the water utilisation
ing network design agree with the minimum fresh water and network system was established, the heat recovery network
wastewater targets predicted using WCA in Step 1, indicating was designed using pinch analysis technique. The stream data
that the maximum water recovery network design has been for heat exchanger was extracted from the heat surplus dia-
achieved. gram. Table 5 shows the stream data for the paper mill heat
recovery system extracted from Figs. 8–10. Note from Table 5
Table 4 – Demands and sources for regions above and that most of the hot streams’ supply temperatures are lower
below-Cpinch . than the cold streams’ target temperatures. Since heat from
hot stream cannot be supplied to the cold streams, heat recov-
Above pinch F (t/h) C (ppm) m (kg/h)
ery through indirect heat transfer cannot be implemented for
D1 126.0 20 2.52 this example.
D2 54.0 20 1.08
D3 36.0 20 0.72
D4 202.0 20 4.03
Table 5 – Hot and cold streams for further energy
D5 18.0 20 0.36
recovery in the MNI paper mill.
D6 169.0 100 16.92
D7 1130.4 150 169.56 Match Type F (t/h) Ts (◦ C) Tt (◦ C) Q (MW)
S1 169.0 100 16.92
FW-D1 Cold 100.8 38 50 1.53
S2 435.9 150 65.34
S1-D1 Cold 25.2 43 50 0.21
S3 113.4 150 169.56
FW-D2 Cold 43.1 38 50 0.65
Below pinch F (t/h) C (ppm) m (kg/h) S1-D2 Cold 10.8 43 50 0.09
FW-D3 Cold 28.8 38 81 1.48
D8 676.8 160 108.29 S1-D3 Cold 7.2 43 81 0.32
D9 154.8 160 24.77 FW-D6 Cold 25.1 38 43 0.18
D10 396.0 160 63.36 S1-D4 Hot 40.3 43 37 0.28
D11 104.0 160 16.64 S2-D6 Hot 58.4 48 43 0.34
D12 68.0 250 17.10 S4-D8 Hot 676.8 46 43 2.37
S4 1331.6 160 213.06 S4-D9 Hot 154.8 46 43 0.54
S5 68.4 250 17.10 S2-WW Hot 377.5 48 35 5.73
1514 chemical engineering research and design 8 7 ( 2 0 0 9 ) 1509–1519

Fig. 6 – Water network design that achieves the minimum fresh water consumption and wastewater generation for region
above Cpinch .

Fig. 7 – Water network design that achieves the minimum fresh water consumption and wastewater generation for region
below Cpinch .

3.4.3. Non-isothermal mixing Note that, for the same duty supplied to Heaters A and B,
Since none of the indirect heat transfer techniques can be the fresh water temperature rises from 30 to 38 ◦ C. This would
applied for heat recovery in this work, non-isothermal mixing reduce the hot utility of fresh water in the subsequent pro-
process was used. cesses. Referring to Table 6, the duty of Heater D is 2.73 MW.

Fig. 8 – Heat surplus diagram above Cpinch , part 1.

 chemical engineering research and design 8 7 ( 2 0 0 9 ) 1509–1519 1515

Fig. 9 – Heat surplus diagram above Cpinch , part 2.

Fig. 10 – Heat surplus diagram below Cpinch .

Through non-isothermal mixing, S1 which was used to sat- was added here to achieve the task. The temperature of the
isfy the demands D1, D2 and D3 at a total flowrate of 43.2 t/h mixture became 45 ◦ C. To reduce the mixture temperature to
(25.2 t/h + 10.8 t/h + 7.2 t/h) was mixed with fresh water sent to 37 ◦ C, 0.78 MW heat duty was removed from Cooler I.
D1 and D2 at a total flowrate of 143.9 t/h (100.8 t/h + 43.1 t/h). For the matches that needed to be cooled down like S4-D8
The temperature of the mixture is 39 ◦ C. In order to increase and S4-D9, new coolers were added in order to satisfy the tem-
the mixture temperature to the target temperature for D1 and perature requirement of each demand. Portions of S4 which
D2 at 50 ◦ C, the duty of Heater D had to be raised to 2.77 MW. were sent to D8 (676.8 t/h) and D9 (154.8 t/h) were cooled using
Heater C has a duty of 2.14 MW. 7.2 t/h of S1 which was previ- Cooler G with a total duty of 2.91 MW. Note that the cold utility
ously heated to 50 ◦ C was then sent to D3. S1 was mixed with for wastewater was also reduced due to wastewater flowrate
28.8 t/h of fresh water before being sent to Heater C for further reduction. 5.73 MW of cold utility for Cooler F was needed to
temperature increase. It was calculated that only 1.71 MW was cool down 377.5 t/h of wastewater to the targeted temperature
needed for Heater C to increase the water temperature to 81 ◦ C of 35 ◦ C. Tables 6 and 7 show the new duties for the existing
compared to 2.14 MW before isothermal mixing. and the newly added heaters and coolers. Fig. 11 shows the
In order to achieve the target temperature of D4, Cooler I final minimum water and energy network.
was added to reduce the temperature of S1 from 43 to 37 ◦ C.
25.1 t/h fresh water and 58.4 t/h of S2 was sent to D6. Fresh 3.4.4. Summary of the energy recovery
water and S2 were mixed before sending to D6. In this way, Table 8 summarises the savings from heat recovery from the
only cold utility was needed to cool down the mixed streams application of SWE minimisation technique. Note that the cold
in order to satisfy the heat load requirement of D6. Cooler H utility is increased by 5.5% in the effort to reduce the hot utility.

Table 6 – Hot duty of heat exchangers before and after the application of SWE minimisation.
Heat exchanger Existing water utilisation After SWE minimisation % difference, Q (MW)
network, QH (MW) technique, QH (MW)

A 2.68 2.68 –
B 0.59 0.59 –
C 2.14 1.71 −20.0%
D 2.73 2.77 +1.4%

Total 8.14 7.75 −4.8%

1516 chemical engineering research and design 8 7 ( 2 0 0 9 ) 1509–1519

Table 7 – Cold duty of heat exchangers before and after the application of SWE minimisation.
Heat exchanger Existing water utilisation After SWE minimisation % difference, Q (MW)
network, QC (MW) technique, QC (MW)

E 2.37 – –
F 6.61 5.73 −13.3%
G – 2.91 –
H – 0.78 –
I – 0.05 –

Total 8.98 9.47 +5.0%

Fig. 11 – Final water utilisation network for MNI paper mill.

Table 8 – Comparison of utility usage for MNI paper mill case study before and after implementation of SWE
minimisation technique.
Without using With NEW technique % of utility reduction
NEW technique
Direct heat transfer Indirect heat transfer

Hot utility (MW) 8.14 8.14 7.75 4.8%

Cold utility (MW) 8.98 8.98 9.47 −5.0%

4. Results and discussion results show that the new water utilisation network has
potential to save 13.4% of both fresh water and wastewater
4.1. Utility cost savings costs.
(b) Energy cost saving.
The previous sections reported the reductions in water and In the paper mill water utilisation network, the main
energy consumption achieved using the new SWE minimi- energy costs comprised of the cost live steam, electric
sation technique. In this section, the utility cost savings are
calculated. The paper mill operates 24 h a day and 330 days a
year. Table 9 shows the cost of water, wastewater and energy Table 10 – Fresh water and wastewater cost comparison
for the mill. (with and without using SWE technique).
Before After using
(a) Water and wastewater cost savings. using SWE SWE
Table 10 compares the fresh water and wastewater costs
Fresh water (t/h) 436 377.5
of the existing and the new water utilisation network
Cost of fresh water (USD/year) 156,591 135,580
designed using the SWE minimisation technique. The Savings of fresh water (USD/year) 0 21,010
% fresh water savings 0 13.4

Table 9 – Utilities cost. Before After using

Utility Cost using SWE SWE

Water USD 0.045 per m3 Wastewater (t/h) 436 377.5

Wastewater treatment USD 0.312 per m3 Cost of wastewater (USD/year) 1,078,949 934,182
Hot utilities USD 0.362 per kW Savings of wastewater (USD/year) 0 101,461
Cold utilities USD 0.003 per kW % wastewater savings 0 523,552
 chemical engineering research and design 8 7 ( 2 0 0 9 ) 1509–1519 1517

Table 11 – Load of exchangers before and after SWE Table 13 – Base cost for heat exchangers.
minimisation.
Types
Existing water utilisation New SWE minimisation C0 = USD 300
network, QH (MW) technique, QH (MW) S0 = 5.5 ft2
˛ 0.024
A 2.68 2.68
UF(2002) 3.62
B 0.59 0.59
MPF 0.80
C 2.14 1.71
MF 1.83
D 2.73 2.77

Total 8.14 7.75

4.2. Capital investment
Existing water utilisation New SWE minimisation
network, QC (MW) technique, QH (MW)
The following equation was used for sizing shell and tube heat
E 2.37 – exchangers (Biegler et al., 1997):
F 6.61 5.73
H – 2.91
I – 0.78 Q = UA Tlm (1)
J – 0.05

Total 8.98 9.47 Q is the heat duty derived from energy balance, A is the
required area, and Tlm is the log mean temperature (see Eq.
(2))

power (to drive pumps) and cooling water cost. Table 11 [(T1 − t2 ) − (T2 − t1 )]
Tlm = (2)
shows the duties of heat exchangers for cases before and ln ((T1 − t2 )/(T2 − t1 ))
after applying SWE technique. Here, low pressure steam
was used as hot utility and cooling water as cold utility. The overall heat transfer coefficient was estimated as
The cost of each utility is shown in Table 9. As mentioned U = 1000 Btu/ft2 h ◦ F or 5678 W/m2 K.
in Section 3, the hot utility supply for Heater A and Heater The base cost for each heat exchanger is obtained using the
B came from the recycled heat of flash steam or condensed following equations:
moisture. Therefore, the duties for these two heaters were
not included in the overall water network hot utility cost.  S a
Note that, while the hot utility was reduced from 8.14 to BC = C0 (3)
S0
7.75 MW, the cold utility increased by about 2.86 MW for
the network (see Table 11). Table 12 shows the total cost C0 is the equipment cost of heat exchanger, S0 is the heat
of the hot and cold utilities before and after applying the transfer area, ˛ is the exponent of the equation. S represents
SWE technique. Note that the cold utility had increased the heat transfer area of a particular heat exchanger.
by 5.5% in the effort to reduce the hot utility. When this is The updated bare module cost can be obtained from Eq.
translated into cost savings, the modifications resulted in (4) once the update factor (UF), material and pressure factor
an annual cold utility increment of about USD 0.12 million. (MPF), and module factor (MF) are determined.
However, this loss was compensated for the significant
annual savings of hot utility of about USD 1.13 million.
updated bare module cost = UF (BC) (MPF + MF − 1) (4)

Table 13 shows the base cost of heat exchangers.

Tables 14 and 15 provides the design data and capital cost data
Table 12 – Energy cost comparison (with and without
using SWE technique). for the heat exchangers.
Referring to Section 3.3 coolers were added to the exist-
Existing water New SWE
ing water utilisation network. The coolers have duties ranging
utilisation minimisation
network technique from 0.05 to 2.91 MW. For the existing exchangers, no extra
heat load was added to any of the exchangers. Therefore,
Hot utility (MW) 4.87 4.48
the capital cost investment only involved the newly added
Cost of hot utility 14,012,542 12,890,387
exchangers. Table 15 shows the total capital investment for the
(USD/year)
Saving of hot utility 0 1,122,154 water utilisation network. An estimated investment of about
(USD/year) USD 7190 was required for the water utilisation network.
% utility savings 0 8.0

Existing water New SWE

Table 14 – Tlm , Q and UA for the newly added heaters
utilisation minimisation
and coolers.
network technique
Matches T1 T2 t1 t2 Tlm Q (MW) UA
Cold utility (MW) 8.98 9.47 (MW/◦ C)
Cost of cold utility 229,393 241,910
(USD/year) D 80 70 39 50 29.3 0.04 0.001
Saving of cold utility 0 −12,517 H 46 43 30 33 12.8 2.91 0.228
(USD/year) I 45 37 30 31.7 9.3 0.78 0.084
% cold utility savings 0 −5.5 J 43 37 30 33 7.6 0.05 0.007
1518 chemical engineering research and design 8 7 ( 2 0 0 9 ) 1509–1519

Table 15 – Updated Bare module cost for each heater or cooler.

Match UA (W/K) U (W/m2 K) A (m2 ) Base cost (USD) Updated bare module cost (USD)
−7
D 0.001 5678 2.4 × 10 211 1247.9
H 227,952 5678 40 333 1966
I 83,656 5678 15 325 1919
J 6,548 5678 1 306 1805

Biegler, Lorenz T., Grossman, Ignacdio E. and Westerberg, Authur

Table 16 – Performance comparison for the paper mill
W., (1997). Systematic Methods of Chemical Process Design.
water utilisation network before and after applying SWE
(Prentice Hall Itl).
minimisation technique.
Bogataj, M. and Bagajewicz, M.J., 2008, Synthesis of
Cost (USD/year) Existing water New SWE non-isothermal heat integrated water networks in chemical
utilisation minimisation processes. Computers and Chemical Engineering, 32:
network technique 3130–3142.
De Beer, J., Worrel, E. and Blok, K., 1998, Long-term
Fresh water 156,591 135,580
energy-efficiency improvements in the paper and board
Wastewater 1,078,949 934,182
industry. Energy, 23(1): 21–42.
Hot utility 14,012,542 12,890,387
Dong, H.G., Lin, C.Y. and Chang, C.T., 2008, Simultaneous
Cold utility 229,393 241,910
optimization approach for integrated water-allocation and
Total 15,477,475 14,202,059 heat-exchange networks. Chemical Engineering Science, 63:
Annual utility savings – 1,275,415 3664–3678.
Du, J., Meng, X., Du, H., Yu, H., Fan, X. and Yao, P., 2004, Optimal
design of water network with energy integration in process
4.3. Payback period industries. Chinese Journal of Chemical Engineering, 12:
247–255.
Table 16 compares the utility consumption for the network El-Halwagi, M.M., Gabriel, F. and Harrel, D., 2003, Rigorous
graphical targeting for resource conservation via material
before and after applying the SWE technique. The payback
reuse/recycle networks. Industrial and Engineering Chemistry
period for investment was calculated using Eq. (5). Research, 42: 4319–4328.
Feng, X., Li, Y. and Yu, X., 2008, Improving energy performance of
capital investment (USD) water allocation networks through appropriate stream
payback period = (5)
utility savings (USD/year) merging. Chinese Journal of Chemical Engineering, 16:
480–484.
The annual utility savings was calculated at USD 1.1 mil- Kim, J.Y., Kim, J.K., Kim, J.H., Yoo, C.K. and Moon, I., 2009, A
simultaneous optimization approach for the design of
lions. The capital investment was estimated at USD 7190. This
wastewater and heat exchange networks based on cost
investment could be recovered within a very short payback
estimation. Journal of Cleaner Production, 17: 162–171.
period of 0.006 year (less than a month). The analysis shows Leewongtanawit, B. and Kim, J.K., 2008, Synthesis and
that the project is a worthwhile investment for the relevant optimisation of heat-integrated multiple-contaminant water
company. systems. Chemical Engineering and Processing, 27: 670–694.
Liao, Z., Wu, J., Jiang, B., Wang, J. and Yang, Y., 2008, Design
energy efficient utilization systems allowing operation split.
5. Conclusion
Chinese Journal of Chemical Engineering, 1: 16–20.
Manan, Z.A., Foo, C.Y. and Tan, Y.L., 2004, Targeting the minimum
A new technique for simultaneous water and energy (SWE) water and wastewater flowrate using the new water cascade
reduction in process plant has been introduced. The new analysis technique. AIChe Journal, 50(12): 3169–3183.
technique combines numerical and graphical methods that Manan, Z.A., Tan, Y.L., Foo, C.Y. and Tea, S.Y., 2007, Application of
is applicable to global water operations. The SWE procedure water cascade analysis technique for water minimisation in a
consists of water and wastewater targeting, water utilisation Paper Mill Plant. International Journal of Environment and
Pollution, 29(1–3): 2007.
network design and energy recovery design. A plot of temper-
Manninen, J., Puumalainen, T. and Talja, R., 2002, Energy aspects
ature versus stream flowrate, termed as heat surplus diagram,
in paper mills utilizing future technology. Applied Thermal
is a novel graphical tool developed in this research to guide Engineering, 22: 929–937.
water and energy minimisation simultaneously. Application Parthasarathy, G. and Krishnagopalan, G., 2001, Systematic
of the new SWE reduction technique on a paper mill plant reallocation of aqueous resources using mass integration in a
demonstrates that USD 1.1 million reductions in water and typical pulp mill. Advances in Environmental Research, 5:
energy consumption can be achieved with an investment cost 61–79.
Prakash, R. and Shenoy, U.V., 2005, Targeting and design of water
of USD 7190, giving a payback of less than a month using this
networks for fixed flowrate and fixed contaminant load
approach.
operations. Chemical Engineering Science, 60(1): 255–268.
Savelski, M.J. and Bagejewicz, M.J., 1997, Design and retrofit of
References water utilisation systems in refineries and process plants, In
AIChe Annual Meeting Los Angeles, 188,
Savulescu, L.E. and Smith, R., 1998, Simultaneous Energy and
Bengtsson, C., Nordman, R. and Berntsson, T., 2002a,
Water Minimisation. AIChe Annual Meeting, November 15–20,
Co-ordination pf pinch technology and the MIND method –
Miami Beach, Florida, Unpublished.
applied to a Swedish Board Mill. Applied Thermal
Savulescu, L., Sorin, M. and Smith, R., 2002, Direct and indirect
Engineering, 22: 133–144.
heat transfer in water network systems. Applied Thermal
Bengtsson, C., Nordman, R. and Berntsson, T., 2002b, Utilization
Engineering, 22: 981–988.
of excess heat in the pulp and paper industry – a case study of
Savulescu, L., Kim, J.K. and Smith, R., 2005, Studies on
technical and economic opportunities. Applied Thermal
simultaneous energy and water minimisation—part 1:
Engineering, 22: 1069–1081.
 chemical engineering research and design 8 7 ( 2 0 0 9 ) 1509–1519 1519

systems with no water re-use. Chemical Engineering Science, Wan Alwi, S.R. and Manan, Z.A., 2008, Generic graphical
60: 3279–3290. technique for simultaneous targeting and design of water
Savulescu, L., Kim, J.K. and Smith, R., 2005, Studies on networks. Industrial & Engineering Chemistry Research, 47(8):
simultaneous energy and water minimisation—part 2: 2762–2777.
systems with maximum re-use of water. Chemical Xiao Feng, X., Li, Y. and Shen, R., in press, A new approach to
Engineering Science, 60: 3291–3308. design energy efficient water allocation networks. Applied
Thompson, G., Swain, J. and Kay, M., 2001, The Treatment of Pulp Thermal Engineering.
and Paper Mill Effluent; A Review. Bioresource Technology, 77: Zheng, X.S., Feng, X. and Cao, D.L., 2003, Design water allocation
275–286. network with minimum freshwater and energy consumption.
Tripathi, P., 1996, Mass exchange integration maximizes water Computer Aided Chemical Engineering, 15: 388–393.
recycling at a paper mill. Chemical Engineering, 103(11): 87–90.