ARTICLE IN PRESS

G Model
CACE-4656; No. of Pages 8

Computers and Chemical Engineering xxx (2013) xxx–xxx

Contents lists available at SciVerse ScienceDirect

Computers and Chemical Engineering

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

Centralised utility system planning for a Total Site Heat Integration

network
Peng Yen Liew a , Sharifah Rafidah Wan Alwi a,∗ , Petar Sabev Varbanov b ,
Zainuddin Abdul Manan a , Jiří Jaromír Klemeš b
a
Process Systems Engineering Centre (PROSPECT), Faculty of Chemical Engineering, Universiti Teknologi Malaysia, 81310 UTM Johor Bahru, Johor, Malaysia
b
Centre for Process Integration and Intensification – CPI2 , Research Institute of Chemical and Process Engineering – MŰKKI, Faculty of IT, University of
Pannonia, Egyetemu. 10, H-8200 Veszprém, Hungary

a r t i c l e i n f o a b s t r a c t

Article history: Total Site Heat Integration (TSHI) is a technique of exchanging heat among multiple processes via a
Received 22 September 2012 centralised utility system. An analysis of the integrated multiple processes, also known as the Total
Received in revised form 6 February 2013 Site (TS) system sensitivity, is needed to characterise the effects of a plant maintenance shutdown, to
Accepted 13 February 2013
determine the operational changes needed for the utility production and to plan mitigation actions. This
Available online xxx
paper presents an improved Total Site Sensitivity Table (TSST) to be used as a systematic tool for this
purpose. The TSST can be used to consider various ‘what if’ scenarios. This tool can be used to determine
Keywords:
the optimum size of a utility generation system, to design the backup generators and piping needed in
Total Site Heat Integration
Pinch Analysis
the system and to assess the external utilities that might need to be bought and stored. The methodology
Total Site Sensitivity Table is demonstrated by using an Illustrated Case Study consisting of three processes. During the TS normal
Sensitivity analysis operation, the Total Site Problem Table Algorithm (TS-PTA) shows that the system requires 1065 kW of
Utility production planning High Pressure Steam and 645.5 kW of Medium Pressure Steam as the heating utility, while for the cooling
utility, 553.5 kW of Low Pressure Steam and 3085 kW of cooling water are required. The results of the
modified TSST proposed that a boiler and a cooling tower with the system design requiring a maximum
capacity of 2.172 MW of steam and 4.1865 MW of cooling water are needed to ensure an operational
flexibility between the three integrated processes.
© 2013 Elsevier Ltd. All rights reserved.

1. Introduction
Beginning in the 1970s, research on maximising heat recovery
using the Pinch Analysis technique has grown rapidly. Building
upon the analysis designed for a single plant heat recovery sys-
tem, previous works have also explored the potential of integrating
the heat requirements for multiple plants via a centralised utility
system, further reducing the utility requirements in industry and
producing integrated community systems. This analysis is known
as the Total Site Heat Integration (TSHI). The TSHI concept was first
introduced by Dhole and Linnhoff (1993). This integration method
offers a greater energy savings opportunity compared with the
focus on the single processes. The Site Sink-Source Profiles (SSSP)
are composite profiles derived from the combined Grand Compos-
ite Curves (GCCs) of the individual processes. SSSP could be divided
into site source and sink profile. The site source profile is the total
energy available in all processes, while the site sink profile shows
∗ Corresponding author. Tel.: +60 07 5535533; fax: +60 07 5581463.
E-mail addresses: shasha@cheme.utm.my, sr wanalwi@yahoo.com
(S.R. Wan Alwi).
the total energy requirement in the processes. These profiles were
introduced by Dhole and Linnhoff (1993) and Raissi (1994) to target
the overall energy requirements for the Total Site system. Klemeš,
Dhole, Raissi, Perry, and Puigjaner (1997) extended the targeting
method to include the Total Site Profile (TSP), the Site Composite
Curves (SCC) and the Site Utility Grand Composite Curve (SUGCC).
For the same concept, the TSP is synonymous with a SSSP, while the
SCC are constructed by performing Multiple Utility (MU) targets
on the TSP. The SUGCC is a form of the site composite, providing
valuable insights on the cogeneration potential for the Total Site
(TS). These tools have significantly impacted the development of
the TSHI technique.
To reduce the carbon footprint, Perry, Klemeš, and Bulatov
(2008) have conceptually extended the TSHI by integrating large
community servicing with corporate buildings as additional heat
sinks and renewable energy as heat sources. The inherent vari-
ability in the heat supply and demand increased the difficulty in
both handling and controlling the system. Varbanov and Klemeš
(2011) later introduced into the Total Site description the Time
Slices with a heat storage system for handling the variable sup-
ply and demand in the TSHI using the heat integration of the
individual batch processes. In the same paper, the Total Site Heat

0098-1354/$ – see front matter © 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.compchemeng.2013.02.007

Please cite this article in press as: Liew, P. Y., et al. Centralised utility system planning for a Total Site Heat Integration network. Computers and
Chemical Engineering (2013), http://dx.doi.org/10.1016/j.compchemeng.2013.02.007
 ARTICLE IN PRESS
G Model
CACE-4656; No. of Pages 8

2 P.Y. Liew et al. / Computers and Chemical Engineering xxx (2013) xxx–xxx

based on a numerical method known as the Total Site Problem

Nomenclature Table Algorithm (TS-PTA) to target the Total Site utility require-
ment. This method is easier to construct, giving faster and more
TS initial supply temperature (◦ C) accurate results compared to the graphical approach, which has a
TT final target temperature (◦ C) tendency to include a graphical error with the curve shifting. The
T shifted temperature (◦ C) Total Site Utility Distribution (TSUD) Table was also introduced
T double-shifted temperature (◦ C) to visualise the heat flows between the processes and the utility
CHP Combined Heat and Power system. A numerical tool, the Total Site Sensitivity Table (TSST),
CW cooling water to explore the site sensitivity was also proposed in the same
GCC Grand Composite Curve paper.
HPS high-pressure steam There are several mathematical models for the plant utility sys-
LIES Locally Integrated Energy Sector tem planning process that incorporate the TSHI theory. Mavromatis
LPS low-pressure steam and Kokossis (1998) introduced models for a steam system in the
mCp heat capacity flowrate (kW/◦ C) TSHI with two main objectives, the selection of the pressure steam
MU multiple utility levels and the determination of the operating unit configuration
MU-PTA Multiple Utility Problem Table Algorithm for the steam levels. A new methodology was developed by Halasz,
PTA Problem Table Algorithm Nagy, Ivicz, Friedler, and Fan (2002) for the optimal retrofit synthe-
Qcmin minimum cooling requirement (kW) sis and operation of the steam-supply system of a chemical complex
Qhmin minimum heating requirement (kW) with production capacities for a multitude of products that vary
SCC Site Composite Curve temporarily. Marechal and Kalitventzeff (2003) proposed optimi-
SDN Steam Distribution Network sation models for solving a multi-period problem that incorporated
SUGCC Site Utility Grand Composite Curve the selection and target for the optimum operational strategy for
SSSP Site Sink-Source Profile a utility system. A top-level analysis (Varbanov, Perry, Makwana,
TS Total Site Zhu, & Smith, 2004) is another mathematical modelling method-
TSP Total Site Profile ology that can be used for these concepts. This method allows for
TSHI Total Site Heat Integration “scoping”, i.e., selecting the site processes to target for the heat
TSST Total Site Sensitivity Table integration improvements. The current steam and power demands
TSUD Total Site Utility Distribution can be optimised and the potential benefit of reducing the steam
TS-HSC Total Site Heat Storage Cascade demand can be assessed. A set of curves for the steam marginal
TS-PTA Total Site Problem Table Algorithm prices can be produced for the system under consideration via a
DH Stream Heat Load (kW) top-level analysis. Chen and Lin (2011) proposed a systematic opti-
Tmin,pp minimum temperature difference between process misation approach to design a Steam Distribution Network (SDN)
stream (◦ C) of the steam systems to obtain an improved energy utilisation in
Tmin,up minimum temperature difference between utility the network. In this model, the operating conditions of the SDN
and process streams (◦ C) were treated as design variables to be optimised.
In another development on the TSHI, Bandyopadhyay, Varghese,
and Bansal (2010) proposed a simplified methodology to target
Cascade was introduced for visualising the heat flows across pro- the cogeneration potential based on the Salisbury (1942) approx-
cesses, the steam system and the heat storage system. Wan Alwi, imation. This method is simple and linear, using the rigorous
Liew, Varbanov, Manan, and Klemeš (2012) introduced a numeri- energy balance at the steam header. Kapil, Bulatov, Smith, and Kim
cal solution for the TSHI system to address the variable availability. (2012)introduced a new model based on an isentropic expansion.
This work presented a Total Site Heat Storage Cascade (TS-HSC) to These model results were favourable compared with the results
address the heat storage facilities required by the TS system. Nemet, from the detailed isentropic design methods. This method also
Klemeš, Varbanov, and Kravanja (2012) discussed the approaches included an optimisation study, which systematically determined
needed to maximise the use of the renewable energy sources with the levels of the steam mains, subjected to economic parameters
a fluctuating supply. They introduced a framework for the integra- and constraints.
tion of the chemical processes with solar energy, allowing a user to The TSST was first proposed by Liew et al. (2012). The tool could
determine the amount of potential solar thermal energy that could be used to systematically determine the minimum and maximum
be used within a process. boiler and cooling utility capacities at different operating condi-
Varbanov, Fodor, and Klemeš (2012) revisited the global mini- tions. However, the tool has a major assumption, which treated
mum temperature difference (Tmin ) used in the previous method. the hot utility and the cold utility as the same type of utility. For
This work suggested that the values for Tmin should be speci- example, the “Difference from normal operation” should not be
fied for each process on the site individually. They demonstrated calculated for the scenario when the Low Pressure Steam (LPS) con-
that the assumption of a global Tmin for the entire Total Site may sumption for the base case are located at different Pinch regions.
have been oversimplified, leading to inadequate results with the This is because the steam generated cannot be used to satisfy the
imprecise estimation of the overall Total Site heat recovery targets. cooling requirement at the same steam level. Furthermore, the tool
The modified Total Site targeting procedure proposed in the paper can be further developed to explore the potential of cascading the
allowed for more realistic heat recovery targets for the Total Sites excess energy at high temperature to satisfy the LPS requirement
to be targeted. during the process operational change scenario. In the method pro-
A graphical approach was used in the early stages of the TS tar- posed previously, the maximum and minimum boiler capacities are
geting, including the TSP, the SCC, the SUGCC construction (Klemeš also not determined in a systematic way via the TSST. The potential
et al., 1997), the TS targeting methodology for both industrial pro- worst case scenario was also not identified as there was no specific
cesses and renewable energy sources (Perry et al., 2008) and the ways to determine the numbers of “what if” scenarios.
TS targeting methodology with both process and utility specific In this paper, the limitations faced by the TSST methodology
Tmin (Varbanov et al., 2012). Liew, Wan Alwi, Varbanov, Manan, introduced by Liew et al. (2012) is addressed and the tool is
and Klemeš (2012) recently introduced an alternative approach improved to include the insights and guidelines generated from the

Please cite this article in press as: Liew, P. Y., et al. Centralised utility system planning for a Total Site Heat Integration network. Computers and
Chemical Engineering (2013), http://dx.doi.org/10.1016/j.compchemeng.2013.02.007
 ARTICLE IN PRESS
G Model
CACE-4656; No. of Pages 8
P.Y. Liew et al. / Computers and Chemical Engineering xxx (2013) xxx–xxx
TSHI utility production system planning process. This tool is able to
determine the worst case scenario, the maximum capacity and the
turn-down ratio for each type of utility supply for multiple scenar-
ios. This approach also allows for a more flexible plant operation
with a low Total Site sensitivity towards operational disturbances.
The effects of a process disruption within a Total Site can be ana-
lysed with this tool to allow contingency plans and actions to be
taken to avoid disruptions in the other processes that are integrated
in the system. The process disruptions may include shutdowns of
one or more plants, a scheduled maintenance, production changes
or emergency cases.
2. Methodology
The proposed methodology for the improved Total Site Sensi-
tivity Table (TSST) is defined in the following steps.
2.1. Step 1: stream data determination
The hot and cold stream data for the individual plants that would
be integrated in the Total Site system are determined. This pro-
cess includes the supply and target temperature data, the fluid flow
rate and the heat capacity data, information that is essential for a
Pinch Analysis study. The existing utility temperature levels from
the centralised utility system are also determined.
2.2. Step 2: TS utility requirement determination
Once the data are determined, the Total Site Heat Integration
analysis can be performed by using either the established graphical
or the numerical approach. The TSHI analysis involves the targeting
of the individual plant heat recovery potential and the determina-
tion of the loads and levels of utilities that can be generated below
the Pinch region (as TS heat sources) and used to satisfy the heat
requirements (TS sinks) above the Pinch region. The determination
of the steam requirements for these multiple utility levels can be
obtained by using either the Grand Composite Curves (GCC) with
the Total Site Profile (TSP) introduced by Klemeš et al. (1997) or the
Multiple Utilities Problem Table Algorithm (MU-PTA) by Liew et al.
(2012). The optimum utility levels are determined by varying the
steam saturation temperature or pressure on the established tools.
These TS sources and sinks streams are then composited and the TS
minimum multiple utility target after the TS heat recovery can be
determined using either the Site Composite Curve (SCC) with the
Site Utility Grand Composite Curve (SUGCC) (Klemeš et al., 1997)
or by a numerical approach through the Total Site Problem Table
Algorithm (TS-PTA) (Liew et al., 2012).
2.3. Step 3: Total Site Sensitivity Table (TSST) construction
The TSST is a useful tool, demonstrating the effect of plant shut-
downs or operability on the TS utility generation and requirements.
The following methodology components can be used to construct
the improved sensitivity analysis tool.
2.3.1. Determining the numbers of possible scenarios
The possible scenarios for a shutdown for either a single plant
or multiple plants can be estimated using Eq. (1). This equation
can accurately determine the number of scenarios (Sn ) needed to
be examined in the TSST as a function of the number of plants (n)
integrated in the TS. In this equation, the effect of a total plant shut-
down is assumed to be the consequences derived from the worst
case scenario. As an example, for three integrated plants, Plants A,
B and C, the seven potential scenarios in the TS analysis would be:
Please cite this article in press as: Liew, P. Y., et al. Centralised utility system planning for a Total Site Heat Integration network. Computers and
Chemical Engineering (2013), http://dx.doi.org/10.1016/j.compchemeng.2013.02.007
3
Scenario 1: All plants in full operation (base case)
Scenario 2: Plant A shutdown, Plants B and C in operation
Scenario 3: Plant B shutdown, Plants A and C in operation
Scenario 4: Plant C shutdown, Plants A and B in operation
Scenario 5: Plants A and B shutdown, Plant C in operation
Scenario 6: Plants A and C shutdown, Plant B in operation
Scenario 7: Plants B and C shutdown, Plant A in operation
1 3 5
Sn = n + n n = 2, 3, 4, 5 . . . (1)
6 6
where S, number of scenarios; n, number of plants/processes.
2.3.2. Obtaining the TS targets for all possible scenarios
Step 2 is repeated to obtain the utility requirement for the Total
Site system for the possible scenarios. To do this, the source and
the demand data of the plant being shut down are omitted in the
construction of the TSP or the TS-PTA.
2.3.3. Recording the TS target at utility levels with storage
facilities
The requirements for the different types of utilities during the
different scenarios are recorded in the TSST. The value for the hot
utility (MU above the TS Pinch point) is recorded without bracket,
while the value for the cold utility (MU below the TS Pinch Point)
is recorded in a square bracket to emphasise the differences. The
location of the TS Pinch Point is also noted.
2.3.4. Determining the requirement difference of base case and
each scenario
Next, the effect of the variations between the base case with all
plants operating normally and the various possible scenarios for
the TS centralised utility system needs to be analysed. For the cases
in which the multiple utility for the base case and the Scenario i is
located in the same Pinch region (either both above the Pinch or
both below the Pinch), the difference between the Scenario i and
the base case utility can be calculated using Eq. (2) for each utility
type j:
MUDifference,j = MUbase−case,j − MUSi,j (2)
where MUDifference,j , multiple utility variations between the base
case and the scenario for the Utility j; MUbase-case,j , multiple util-
ity of the base case for the Utility j; MUSi,j , multiple utility of the
Scenario i for the Utility j.
A positive difference indicates that there is an excess utility pro-
duced by the boiler houses or the chilling units compared with the
base case scenario, while a negative difference represents a deficit
in the external utility.
For scenarios where the MUbase-case,j and the MUSi,j are located at
different Pinch regions, the utility was changed from a hot to a cold
utility requirement or vice versa. For example, for the MUbase-case,j
located above the Pinch, this situation would require a Low Pressure
Steam (LPS) as the hot utility. With the MUSi,j located below the
Pinch, the situation does not require a LPS, but rather generates a
LPS that will require the stream to be cooled down with cooling
water. This scenario indicates that the centralised utility system
will have an excess LPS from the boiler during Scenario i, generating
an LPS from the process without a reduction in the boiler capacity.
This situation would result in a double cooling utility needed during
Scenario i. A summary of the Scenario analysis is as follows:
(i) If the MUbase-case,j is located above the Pinch and the MUSi,j is
located below the Pinch, more cold utility will be required. The
calculation of the difference is required to determine both the
heating and the cooling utility. For the heating utility,
MUDifference,j = −MUbase−case,j (3)
 ARTICLE IN PRESS
G Model
CACE-4656; No. of Pages 8

4 P.Y. Liew et al. / Computers and Chemical Engineering xxx (2013) xxx–xxx

For the cooling utility Table 1

Stream data for Processes A, B and C.

MUDifference,j = MUSi,j (4) Stream TS (◦ C) TT (◦ C) H (kW) mCp (kW/◦ C) T  S (◦ C) TT (◦ C)

Process A
(ii) If the MUbase-case,j is located below the Pinch and the MUSi,j is A1 hot 200 100 2000 20 190 90
located above the Pinch, the cold utility produced is not needed, A2 hot 150 60 3600 40 140 50
A3 cold 50 220 −2550 15 60 230
but extra steam for heating is required in this scenario. Eq. (4)
A4 hot 170 150 1000 50 160 140
can be used to determine the difference of the heating utility, Process B
while the difference for the cooling utility could be determined B1 hot 200 50 450 3 190 40
by Eq. (3). B2 hot 200 119 1863 23 190 109
B3 cold 30 200 −680 4 40 210
B4 cold 130 150 −300 15 140 160
2.3.5. Obtaining the difference after cascade Process C
The third step in the TSST is to calculate the “difference after C1 hot 240 100 210 1.5 230 90
C2 cold 50 250 −400 2 60 260
cascade”. For the hot utility, the excess is kept at the highest level C3 cold 40 190 −1500 10 50 200
with the deficit at the lowest level. For the cold utility, the deficit is C4 cold 109 140 −186 6 119 150
kept at the lowest temperature with the excess at the original utility
level. This value could be determined using the heuristic below:
3.1. Step 1: steam data extraction
• For an excess utility (positive MUDifference,j ) above the TS Pinch
The Illustrative Case Study consisted of three process units, Pro-
location in a scenario, this value is cascaded down to the lower
cesses A, B, and C, as shown in Table 1. The minimum temperature
utility level with a heating utility deficit (negative difference) to
difference between the process streams (Tmin,pp ) and the min-
form the “difference after cascade”. As the higher hot utility lev-
imum temperature difference between the process streams and
els (e.g., HPS) are typically more expensive than the lower hot
the utility (Tmin,up ) were assumed to be 20 ◦ C and 10 ◦ C, respec-
utility levels (e.g., MPS), the lower temperature hot utilities are
tively. The existing available on site utilities were High Pressure
preferred to be cascaded downward to the heating utility deficit,
Steam (HPS) at 270 ◦ C, Medium Pressure Steam (MPS) at 180 ◦ C,
ensuring that the production of the higher temperature utility
Low Pressure Steam (LPS) at 133 ◦ C and cooling water (CW) at
can be reduced to reduce operating costs. For a heating utility in
15–20 ◦ C.
excess of 200 kW of HPS and 100 kW of MPS, a value of 200 kW
The simplified utility system configuration is presented in Fig. 1.
for the LPS would be insufficient in this case. The total of 300 kW
A natural gas fuelled boiler was used to produce the HPS to satisfy
of the higher utility level could be let down to the LPS level, but
the site steam demand. The reduction of the steam production from
with only 100 kW of the MPS and the HPS available to be let down
the boiler would directly contribute to the carbon emission reduc-
to the LPS level. The net utility excess of 100 kW could remain in
tion of the site. Steam can also be produced from the removal of
the HPS to ensure that this high quality heat could be sold or the
heat from the high temperature process stream. For example, the
heating utility generation can be reduced.
oil and gas processing industries typically employ very high tem-
• For the scenarios below the TS Pinch region, the deficit of the
perature processes, generating both HPS and MPS at the processing
cooling utility (the negative difference in brackets) for the steam
site. The HPS produced from the boiler house or the processes
generation can be cascaded downwards to the cooling water level
could be stepped down to a lower energy level through a turn-
as there is no need to generate steam without a requirement. The
down valve or a steam turbine. Industries with excessive amounts
generation of the unused steam would only require an additional
of high temperature heat have frequently used steam turbines in
boiler capacity with the excess heat being removed with cooling
their steam systems to generate electricity. The condensate from
water.
the used steam can be sent back to a de-aeration tank to eliminate
dissolved oxygen to prevent corrosion.
2.3.6. Determining the storage capacity required
The capacity requirements for the centralised utility system can 3.2. Step 2: TS utility requirement determination
be summarised using the results from each scenario:
In the next step, the Multiple Utility Problem Table Algorithm
• The maximum requirement for each type of utility can be calcu- (MU-PTA) by Liew et al. (2012) was used to target the external
multiple utility requirements for each individual process. Table 2
lated by adding the absolute value of the most negative number
shows a sample MU-PTA for Process A. Process A had the poten-
in the “difference after cascade” with the utility requirement at
tial to generate 50 kW MPS and 1515 kW LPS, requiring 3085 kW
normal operating conditions. With no negative values, the maxi-
of cooling water. This information can be used to contribute to the
mum utility requirement would be equal to the normal operation
utility.
• The minimum steam utility requirement for each type of util-
Boiler
ity can be calculated by deducting the most positive value of the Houses
“difference after cascade” from the utility requirement at nor-
HPS
mal operating conditions. The utility requirement during normal
MPS
operations would be used for the cases without a positive differ-
LPS
ence. CW

Process Process Process

3. Case study De-aerator
A B C

The construction of the Total Site Sensitivity Table (TSST) is illus-

trated using a case study. Fig. 1. Total Site with a centralised utility system (after Klemeš et al., 2010).

Please cite this article in press as: Liew, P. Y., et al. Centralised utility system planning for a Total Site Heat Integration network. Computers and
Chemical Engineering (2013), http://dx.doi.org/10.1016/j.compchemeng.2013.02.007
 ARTICLE IN PRESS
G Model
CACE-4656; No. of Pages 8

P.Y. Liew et al. / Computers and Chemical Engineering xxx (2013) xxx–xxx 5

Table 2
MU-PTA for Process A.

Initial MU Utility
T’ T” ΔT mCp (kW/oC) mCp ΔH
Cascade Cascade Requirement
(oC) (oC) (oC) 20 40 15 50 (kW/oC) (kW) (kW) (kW) (kW)
270 0 0 HPS 600
0 0 0
230 230 600 0
40 -15 -600 600
190 190 0 0
PINCH
190 190 0 0
10 5 50 -50
180 50 0 MPS 50
20 5 100 -100
160 160 150 0
20 55 1100 -1100
140 140 1250 0
7 45 315 -315
133 1565 0 LPS 1515
43 45 1935 -1935
90 90 3500 0
30 25 750 -750
60 60 4250 0
10 40 400 -400
50 50 4650 0
30 0 0 0
20 20 4650 0 CW 3085

Table 3 The TS-PTA method by Liew et al. (2012) was used to target the
Summary of multiple utility targets for each process in the Total Site.
Total Site multiple utility. Table 4 shows the TS-PTA for the base case
Below Pinch (heat source) [kW] Above Pinch (heat sink) [kW] scenario. In order to illustrate the utility consumption and genera-
tion, TSP and SCC are plotted based on the TS-PTA (Table 4). Fig. 2
Process A B C A B C
shows the TSP and SCC for the Illustrative Case Study. The TS Pinch
HPS 600 180 285
point was located between the LPS and the CW. A multiple util-
MPS 50 100 595.5
LPS 1515 34 995.5 ity targeting on the TS-PTA demonstrated that the system required
CW 3085 1065 kW of HPS and 645.5 kW of MPS as the heating utility for the
entire TS. For the cooling utility, 553.5 kW of LPS and 3085 kW of CW
TS heat sources. Process A also required 600 kW of HPS for heating, was required. The SDN for the integrated TS system in the Illustra-
representing the TS heat sinks. MU-PTA was also performed for Pro- tive Case Study is shown in Table 5, the Total Site Utility Distribution
cesses B and C. Table 3 summarises the multiple utility targets for (TSUD) Table. These values were recorded in the second column of
each process. the TSST as shown in Table 6.

Table 4
Total Site Problem Table Algorithm (TS-PTA) for the Illustrative Case Study on normal operation.

1 2 3 4 5 6 7 8
Utility Heat source (kW) Heat sink (kW) Heat requirement (kW) Initial cascade (kW) Final cascade (kW) MU cascade (kW) Utility requirement (kW)

0 1710.5 0
HPS 1065 −1065 1065
−1065 645.5 0
MPS 50 695.5 −645.5 645.5
−1710.5 0 0 TS Pinch
LPS 1549 995.5 553.5 −553.5
−1157 553.5 0
CW 3085 3085 −3085
1928 3638.5 0

Table 5
Total Site Utility Distribution (TSUD) Table for the Illustrative Case Study.

Heat Source [kW] Heat Sink [kW]

Process Process Process Site Process Process Process Site
A B C Utility A B C Utility

HPS 1,065 600 180 285

MPS 50 645.5 100 595.5

LPS 1,515 34 995.5 553.5

CW 3,085 3,085

Please cite this article in press as: Liew, P. Y., et al. Centralised utility system planning for a Total Site Heat Integration network. Computers and
Chemical Engineering (2013), http://dx.doi.org/10.1016/j.compchemeng.2013.02.007
 6

CACE-4656; No. of Pages 8

G Model
Chemical Engineering (2013), http://dx.doi.org/10.1016/j.compchemeng.2013.02.007
Please cite this article in press as: Liew, P. Y., et al. Centralised utility system planning for a Total Site Heat Integration network. Computers and

Table 6
Total Site Sensitivity Table (TSST).

Scenario 1 (all processes in Scenario 2 (Processes B and C in operation) Scenario 3 (Processes A and C in operation) Scenario 4 (Processes A and B in operation)

ARTICLE IN PRESS
operation – base case)

P.Y. Liew et al. / Computers and Chemical Engineering xxx (2013) xxx–xxx
Utility Requirement Requirement Difference Difference after Requirement Difference Difference after Requirement Difference Difference after
(kW) (kW) (kW) cascade (kW) (kW) (kW) cascade (kW) (kW) (kW) cascade (kW)

(a)
HPS 1065 465 600 0 885 180 180 780 285 285

MPS 645.5 695.5 −50 0 545.5 100 100 50 595.5 595.5

Pinch Pinch Pinch
LPS 961.5 −961.5 −411.5
[553.5] [553.5] [553.5] [519.5] [34] [34] [1549] [−995.5] [0]
Pinch
CW [3085] [0] [3085] [3085] [3085] [0] [0] [3085] [0] [−995.5]

Scenario 1 (all processes in Scenario 5 (Process A in operation) Scenario 6 (Process B in operation) Scenario 7 (Process C in operation) Summary
operation – base case)

Utility Requirement Requirement Difference Difference after Requirement Difference Difference after Requirement Difference Difference after Maximum Minimum
cascade cascade cascade requirement requirement

(b)
HPS 1065 600 465 99 180 885 885 285 780 0 1065 180
Pinch
MPS 645.5 645.5 645.5 100 545.5 545.5 595.5 50 0 695.5 0
[50] [−50] [0] [0] [0]
Pinch Pinch
LPS 995.5 −995.5 −165.5 411.5 0
[553.5] [1515] [−961.5] [0] [34] [519.5] [519.5] [553.5] [553.5] [553.5] [0]
Pinch
CW [3085] [3085] [0] [−1011.5] [0] [3085] [3085] [0] [3085] [3085] [4186.5] [0]
 ARTICLE IN PRESS
G Model
CACE-4656; No. of Pages 8
P.Y. Liew et al. / Computers and Chemical Engineering xxx (2013) xxx–xxx
300
HPS
250
200
MPS
150
LPS
Site Source Profile Site Sink Profile
100
50
CW
0
-6000 -5000 -4000 -3000 -2000 -1000 0 1000 2000
Fig. 2. Total Site Profile (TSP) and Site Composite Curves (SCC) for the Illustrative
Case Study.
3.3. Step 3: Total Site Sensitivity Table (TSST) construction
For the three processes involved in this Illustrative Case Study,
the number of possible scenarios involving process shut downs is
equal to seven, as suggested by Eq. (1). The TS-PTAs for each sce-
nario were constructed. The multiple utility targets with the TS
Pinch location are listed in Table 6.
The differences between the normal operating conditions and
the various scenarios of the TS utilities requirement are determined
based on above or below TS Pinch location, by using Eqs. (2)–(4) (as
explained in Section 2.3). This step is corrected from the difference
calculation without considering the location of TS Pinch in the pre-
vious published methodology in Liew et al. (2012). For example, the
TS Pinch location changed from a point between the MPS and the
LPS for the base case to a point between the LPS and the CW for Sce-
nario 2. The difference between the base case and Scenario 2 for the
HPS, the MPS and the CW can be calculated directly using Eq. (2),
as these values are located in the same Pinch region. These values
are given in Column 4, Table 6a. The positive difference indicated
that the boiler was producing an excess of 600 kW of the HPS and
the cooling tower had an excess of 3850 kW of the CW that was not
used during Scenario 2. The negative difference indicated there was
a deficit of 50 kW of MPS that was needed to be generated. For the
LPS, the utility location changed from below the Pinch (base case)
to above the Pinch (Scenario 2). The previous methodology by Liew
et al. (2012) is expected to yield 408 kW of extra LPS requirement
in Scenario 2. However, the extended TSST in this paper indicates
that the system requires an additional boiler to generate the LPS
instead. The negative difference indicated a deficit of 961.5 kW in
the LPS. The additional CW was added at the CW difference row.
The “difference after cascade” was calculated for each of the
scenarios. The excess heating utility, which is represented by a pos-
itive difference, can be cascaded towards the lower steam utility
level characterised by a negative difference for the point above the
Pinch region. This type of heat recovery is not included in the pre-
vious published TSST (Liew et al., 2012) methodology. In Scenario
2, Process A can be shut down due to a scheduled maintenance.
This shutdown would not need the 600 kW HPS generated from
the centralised utility system for Process A during the base case,
while increased levels of the MPS and the LPS would be needed to
satisfy the requirements for Processes B and C. To minimise the
changes in the boiler utility generation, the excess unused HPS
(indicated by a positive difference) can be cascaded downwards to
satisfy the deficits in the MPS and the LPS. This cascade would allow
the HPS and the MPS boiler to run as usual, with only an additional
Please cite this article in press as: Liew, P. Y., et al. Centralised utility system planning for a Total Site Heat Integration network. Computers and
Chemical Engineering (2013), http://dx.doi.org/10.1016/j.compchemeng.2013.02.007
7
411.5 kW LPS needed to be generated. For the points below the
Pinch region, there is only one utility type (CW) in excess, indicat-
ing that the 3085 kW CW production would need to be shutdown
during Scenario 2.
After analysing the various ‘what if’ scenarios in the TSST, the
TS system in this demonstration case study required a natural gas
powered boiler with a maximum capacity of 2172 kW of super-
heated HPS production. The superheated steam generated with
this system could be stepped down to 1065 kW of HPS, 695.5 kW
of MPS and 411.5 kW of LPS, compared with the 1710 kW of total
steam required during normal operations. A backup boiler may be
a good option to cater to these production changes or the steam
requirement may also be obtained externally.
In the most extreme case, the minimum total steam require-
ment for the TS system was 180 kW during Scenario 6, as shown
in Table 6b. For instance, the boiler in the TS required a turn down
3000 4000
ratio of not less than 50%, assuming the lowest capacity of the boiler
to be 1086 kW. These conditions would suggest that the remaining
906 kW of steam needed to be cooled down using cooling water
or sold to other plants. Another alternative would be to divert
the excess steam production for the Combined Heat and Power
(CHP) system to generate more power. Another choice could also
include dividing the capacity between two boilers, allowing the
boiler capacity to be manipulated based on the equipment restric-
tions. For the cooling water, the maximum capacity required was
5092.5 kW. An additional 4186.5 kW for the maximum capacity and
906 kW of steam generated was required due to the limitations
from boiler’s turn down ratio. The minimum capacity for the cooling
tower was 0 kW.
4. Conclusion
The Total Site Sensitivity Table (TSST) has been extended for
planning the TSHI centralised utility system. This approach pro-
vides insights on the consequences of a plant shutdown or process
upsets on the entire TS system integration and the centralised util-
ity system. By identifying heating or cooling utility requirements
during a plant shut down, a suitable size utility system can be
planned to ensure a flexible and undisrupted operation for the
integrated processes. These decisions must be balanced with the
operational challenges as well as the capital and operating costs.
The proposed tool is able to determine the optimal design and
operation of the centralised utility system by assuming 100% effi-
ciency for the utility system. This is actually not representing the
actual situation. For the future work, the effect of load changes on
the TS utility system efficiencies can be further investigated. For
example, the turbine and boiler efficiencies are affected by their
loads (Möller, Genrup, & Assadi, 2007; Zhou, Liu, Li, Pistikopoulos,
& Georgiadis, 2013). Hence, the off-design operation may require
special considerations during the utility system sizing.
Acknowledgements
The authors would like to thank the Ministry of Higher Edu-
cation (MOHE) Malaysia and the UTM in providing the research
funding for this project under Vote No. Q.J130000.2544.03H44
and the EC supported project Energy – 2011-8-1 Efficient Energy
Integrated Solutions for Manufacturing Industries (EFENIS) –
ENER/FP7/296003/EFENIS.
References
Bandyopadhyay, S., Varghese, J., & Bansal, V. (2010). Targeting for cogeneration
potential through total site integration. Applied Thermal Engineering, 30, 6–14.
Chen, C., & Lin, C. (2011). A flexible structural and operational design of steam
systems. Applied Thermal Engineering, 31, 2084–2093.
 ARTICLE IN PRESS
G Model
CACE-4656; No. of Pages 8
8 P.Y. Liew et al. / Computers and Chemical Engineering xxx (2013) xxx–xxx
Dhole, V. R., & Linnhoff, B. (1993). Total site targets for fuel, co-generation, emission
and cooling. Computers and Chemical Engineering, 17, 101–109.
Halasz, L., Nagy, A. B., Ivicz, T., Friedler, F., & Fan, L. T. (2002). Optimal retrofit design
and operation of the steam-supply system of a chemical complex. Applied Ther-
mal Engineering, 22, 939–947.
Kapil, A., Bulatov, I., Smith, R., & Kim, J. (2012). Site-wide low-grade heat recovery
with a new cogeneration targeting method. Chemical Engineering Research and
Design, 90(5), 677–689.
Klemeš, J., Dhole, V. R., Raissi, K., Perry, S. J., & Puigjaner, L. (1997). Targeting and
design methodology for reduction of fuel, power and CO2 on total site. Applied
Thermal Engineering, 7, 993–1003.
Klemeš, J., Friedler, F., Bulatov, I., & Varbanov, P. (2010). Sustainability in process
industry: Integration and optimization. New York, US: McGraw Hill.
Liew, P. Y., Wan Alwi, S. R., Varbanov, P. S., Manan, Z. A., & Klemeš, J. J. (2012). A
numerical technique for Total Site sensitivity analysis. Applied Thermal Engi-
neering, 40, 397–408.
Marechal, F., & Kalitventzeff, B. (2003). Targeting the integration of multi-period
utility systems for site scale process integration. Applied Thermal Engineering,
23, 1763–1784.
Mavromatis, S. P., & Kokossis, A. C. (1998). Conceptual optimisation of utility
networks for operational variations: I. Targets and level optimisation. Chemical
Engineering Science, 53(8), 1585–1608.
Möller, B. F., Genrup, M., & Assadi, M. (2007). On the off-design of a natural gas-fired
combined cycle with CO2 capture. Energy, 32, 353–359.
Please cite this article in press as: Liew, P. Y., et al. Centralised utility system planning for a Total Site Heat Integration network. Computers and
Chemical Engineering (2013), http://dx.doi.org/10.1016/j.compchemeng.2013.02.007
Nemet, A., Klemeš, J. J., Varbanov, P. S., & Kravanja, Z. (2012). Methodology for
maximising the use of renewables with variable availability. Energy, 44(1),
29–37.
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.
Raissi K. (1994). Total site integration. PhD thesis. Manchester, UK: UMIST.
Salisbury, J. K. (1942). The steam-turbine regenerative cycle – An analytical
approach. Transactions of the ASME, 64(4), 231.
Varbanov, P. S., Fodor, Z., & Klemeš, J. J. (2012). Total site targeting with process
specific Tmin . Energy, 44(1), 20–28.
Varbanov, P., & Klemeš, J. (2011). Integration and management of renewables into
total sites with variable supply and demand. Computers and Chemical Engineer-
ing, 35(9), 1815–1826.
Varbanov, P., Perry, S., Makwana, Y., Zhu, X. X., & Smith, R. (2004). Top-level anal-
ysis of site utility systems. Chemical Engineering Research and Design, 82(A6),
784–795.
Wan Alwi, S. R., Liew, P. Y., Varbanov, P. S., Manan, Z. A., & Klemeš, J. J. (2012). A
numerical tool for integrating renewables into total sites with variable supplies
and demands. Computer Aided Chemical Engineering, 30, 1347–1351.
Zhou, Z., Liu, P., Li, Z., Pistikopoulos, E. N., & Georgiadis, M. C. (2013). Impacts of
equipment off-design characteristics on the optimal design and operation of
combined cooling, heating and power systems. Computers and Chemical Engi-
neering, 48, 40–47.