Microsoft Word - 0381_CET45_021

CCHHEEMMIICCAALL EENNGGIINNEEEERRIINNGG TTRRAANNSSAACCTTIIOONNSS

VOL. 45, 2015

A publication of

The Italian Association

of Chemical Engineering

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

ISBN 978-88-95608-36-5; ISSN 2283-9216 DOI: 10.3303/CET1545021

Please cite this article as: Nemet A., Lee K., Klemeš J.J., Varbanov P.S., Moon I., Kravanja Z., 2015, Safety analysis

embedded in total site synthesis, Chemical Engineering Transactions, 45, 121-126 DOI:10.3303/CET1545021

121

Safety Analysis Embedded in Total Site Synthesis

Andreja Nemeta, Kwanghee Leeb, Jiří J. Klemeša, Petar S. Varbanova,

Il Moon*b, Zdravko Kravanjac

aCentre for Process Integration and Intensification – CPI2, University of Pannonia, Veszprém, Hungary
bProcess System Engineering Laboratory, Dpt of Chemical and Biomolecular Engng, Yonsei University, Seoul, Korea
cFaculty of Chemistry and Chemical Engineering, University of Maribor, Maribor, Slovenia

ilmoon@yonsei.ac.kr

Decreasing utility consumptions by Total Site (TS) process-to-process Heat Integration can be performed

via several possible configurations, namely: i) Indirect, via intermediate utility and ii) Direct, by transporting

either hot or cold process streams from one process to a heat exchanger placed within another process,

undertaking heat exchange and afterwards returning them to the original process. The indirect heat

transfer requires more complex network compared to direct, whilst the direct configuration can have safety

issues. This contribution presents a five-step approach for TS heat exchanger network (HEN) synthesis.

During the first steps the risk assessment is determined for each possible heat transfer match between

processes. The matches can be classified by applying a risk-ranking matrix to matches with high, medium

and low-level risks. In the fifth step synthesis of the Total Site Heat Integration Network is performed by a

mixed-integer nonlinear programming (MINLP) model by considering the classifications of the matches

obtained during the first step. The matches can be assigned as forbidden (high risk), allowed with penalty

(medium risk) and allowed (low-level risk). The objective of the MINLP model is to maximise the Expected

Net Present Value of the Total Site. This methodology was tested on an illustrative case study for

analysing the impact of risk assessment. The obtained TS heat exchanger networks (HENs) by applying

the described methodology were inherently safer and yet economically viable.

1. Introduction

Total Sites Heat Integration can significantly contribute to all three pillars of sustainability - the

environmental burden can be decreased, the economic efficiency increased, and the social aspect

improved (Čuček et al., 2015). There are two main approaches for Total Site evaluation (Dhole and

Linnhoff, 1993).The first is the thermodynamic approach that sets thermodynamically achievable targets at

a certain minimal temperature difference (e.g. Klemeš, 2013) and a mathematical programming approach,

where different parts or the whole Total Site Heat Exchanger Network can be optimised, usually by

applying an economic objective function (e.g. Klemeš and Kravanja, 2013). Whichever approach is applied

any of them can lead to solutions that can hardly be implemented due to the high risk rate. The public’s

interest in risk assessment has been continuously increasing over the last three decades as a result of the

need for obtaining safer and more reliable processes (Marhavilas et al., 2011). Despite great interest in

risk analysis the majority of scientific publications on risk assessment cover quantitative methods (66 %)

whilst the qualitative methods are only present as a smaller proportion (28 %) – see Marhavilas et al,

2011. The majority of work has focused on the risk assessment and analysis method, assuming a retrofit

of an already existing or designed plant. It is used for identifying places with high risks regarding

operability requirements. However, the possibility of affecting the inherent safety of a process decreases

as the decision or investment on design has already been made (Heikkilä, 1999). Risk assessment at an

early stage of design can have substantial impact on the later operations of processes. Chan et al. (2014)

combined the inherent safety index with Stream Temperature vs. Enthalpy Plot (STEP) analysis developed

for HEN design. It is a graphical approach based on a heuristic that can provide a deep understanding of

the planned HEN. However, it can become too complex for solving larger problems and obtaining solutions

122

can be left to the engineer’s expertise. The framework presented in this work provides a systematic

procedure for synthesis of a Total Site regarding a wider scope of Heat Integration, including risk

assessment at an early stage during planning. The acceptable solutions are often trade-offs between low-

risk designs and economic viabilities. Obtaining designs by considering only risk analysis might not be

implemented due to poor economic performance. In order to evaluate this trade-off during Total Site

synthesis, the synthesis is performed by adding penalties to heat transfers with high risks by establishing a

trade-off between risk level and economic performance.

2. Methodology

The aim of the developed methodology is to consider safety during the process of obtaining a design for a

Total Site. For this purpose a five-step approach has been developed in order to obtain a Total Site Heat

Exchanger Network. An extension of the Grid Diagram (Linnhoff and Flower, 1978) has been used for the

illustration. The Hu represents the heaters whilst Cu stands for the coolers within the graphical

presentation (Figure 1).

Process 1 Process 2

HE 3

Hu2

Cu1

HE 1

Hu1

Pipe 1

Pump 5

Pump 6

Hot process stream

Cold process
stream

HE 2

Valve 3

Valve 4
Valve 5 Valve 6

Valve 9

Pump 3

Pump 4

Valve 8

Valve 7

Fresh Water

Hu3

IU- water IU- steam

IU- condensate

Valve 2

Pump 2

Pump 1 Valve 1

IU- cond.
pipe

Preheating
Evaporation Transport steam

Condensation

Transport condensate

Process 1 Process 2

Pipe 1

HE 1

Hu2

Hot process
stream

Cold process
stream

Pipe 2

Pump 3

Pump 4

Pump 1 Valve 1

Pump 2Valve 2

Valve 3

Valve 4Valve 5

Trans -Hot
process stream

Heat exchange

Transport from Process 1 to Process 2

Transport from Process 2 to Process 1

Process 1
Process 2

HE 1

Pipe 1

Cu1 Hot process
stream

Cold process
stream

Pipe 2

Pump 3

Pump 4

Valve 1Pump 1

Valve 2 Pump 2

Valve 3
Valve 5Valve 4

Transport from Process 2 to Process 1

Transport from
Process 1 to Process 2

Heat exchange

Hu1

Cu1

HE- Heat exchanger
IU- Intermediate utility stream
Hu-Heater
Cu-Cooler
Trans-Cold process stream- Transported cold stream in Process 1
Trans-Hot process stream- Transported hot stream in Process 2

Trans -Cold
process stream

Figure 1: TS HEN, when heat transfer configuration is a) Indirect, via intermediate utility, b) Direct,

transport of cold stream and c) Direct, transport of hot stream heat transfer between different processes

Step 0: Total Site description. The Total Site identification and breakdown has to be performed before risk

assessment can be performed. On a Total Site three different configurations of heat transfer are possible:

i) Indirect, via an intermediate utility (Figure 1a) and direct, by transporting either ii) cold (Figure 1b) or iii)

hot (Figure 1c) process streams of one process to exchange heat within another process. In the case of

Indirect heat transfer the whole heat transfer process can be composed of five basic operations:

preheating, evaporation, transport of steam, condensation and transport of condensate (Figure 1a). In the

case of direct heat transfers only three operations occur: transport of cold (hot) stream to another process,

heat exchange, and transport of cold (hot) process back to its original process, as presented in Figure 1b

(Figure 1c). Moreover, only matches with feasible heat transfer enabled by temperature difference should

be included in the further steps of evaluation.

Step 1: Deviations (hazards) identification. The HAZOP (Hazard and Operability study) method is a

systematic examination of design documents for identifying and documenting hazards through imaginative

thinking. It is the more systemised qualitative technique enabling thorough identifications of deviations

(Marhavilas et al, 2011). The specific Guide Words used for Total Site when performing HAZOP analysis

are presented in Table 1. They could be extended by more deviation parameters; however, this would also

contribute to the complexities of analysis. In this work the temperature and pressure have not been

included as parameters. They have rather been treated as a consequence of decreased or no velocity.

The planned or emergency shutdown of a process can be treated as the consequence of any failure in this

process; therefore, this case has not been included separately as a parameter.

123

Table 1: Guide words for HAZOP analysis used for Total Site

Parameter Guide word Deviation caused by

Flow or velocity More Reduced backpressure, surging, controller failure, valve stuck open

Less Leakage, partial blockade, fouling, sediment, cavitation, low suction head

No Clogging, blockage, pump failure, closed stuck valve, leak

Reverse Valve failure or wrongly inserted, poor isolation, wrong routing, control failure

Composition Other than Presence of impurities (e.g. air, water, acids, corrosion)

Part of Different phases, foaming, change in viscosity or density

Contamination Leakage of one media to other

Step 2: Determining frequency and severity of deviation from planned operating conditions. Based on

Table 1 and Figure 1 the deviation events can be created for each of the heat exchange matches. The

failure frequency is determined based on historical statistical data (e.g. Flemish government, 2009). It

should be noted that the pipe failure frequency is determined by considering its length (Flemish

government, 2009), which is an important aspect in Total Site. The severity of the leakage is determined

following the guideline set in the so-called “Purple book” (Uijt and Ale, 2005). It suggests the calculation of

a dimensionless indication number of an installation Ai for a substance i as presented in Eq(1) of an

installation for substance i:

( )1 2 3 /i i iA Q Q Q Q G= ⋅ ⋅ ⋅ (1)

where Qi is the quantity of the substance i present the installation in kg, Q1 the factor for installation type,

whether process or storage, Q2 the factor for positioning of the installation, Q3 the factor for process

conditions including the amount of substance during the vapour phase after release based on process

temperature, atmospheric boiling point, the substance phase and the ambient temperature, and Gi is the

limit value representing a measure of dangerous properties on both physical properties and the

toxic/explosive/flammable properties of the substance. Determining the risks at each installation can

indicate the place/reason for high risks. In the Total Site approach the consequences are not limited only

to the loss of some of the media but also the missing of the same media in other places can be critical, due

to uncovered heat demand or excess of heat. For this purpose an indication number Aqi of heat loss or

heat cumulating within a system due to cold stream loss is determined by applying the same equation

Eq(1), where the loss presents the amount of heat due media loss within 10 min in pipes and the amount

of heat present within the mass of media in the heat exchanger. The limiting value is the overall hot or cold

utility consumption.

Step 3: Determining the risks of each heat transfer match between processes. The overall risk of one

match is obtained by the summation of risk over each installation required for the heat transfer. The total

failure frequency for one heat transfer match can be determined as by the summations of failure

frequencies of any of the installations involved in the heat transfer. This is done based on the assumptions

that a failure of one installation would lead to a failure of the overall network. The total risk should be

obtained in order to gain information regarding the total severity of the consequences of deviations for a

certain heat transfer between processes. The total severity can then be determined by division of the total

risk by the total frequency of failure. The distribution between frequency and severity can still be estimated

by revealing the basic problems within a network leading to risk (Figure 2a).

Severity

Frequency

High risk

Acceptable risk

ALARP

Heat transfer matches

Severity

Frequency

Forbidden
matches

Allowed matches
-No penalty

Allowed
matches
-Penalty

Penalty-
to reduce
frequency

Penalty-
to reduce severity

ALARP- “as low as reasonably practicable”

a) b)

Figure 2: Presentation of possible heat exchanger match as Severity vs. Frequency a) after risk

assessment and b) as an input data for TS synthesis

Step 4: Connection between Risk Assessment and Total Site synthesis. The results obtained by risk

assessment can be presented as graphical representations Severity versus Frequency (Figure 2a). This

plot is divided in three regions: i) high risk, ii) ALARP (“as low as reasonably practicable”) and iii)

acceptable risk. The same regions are taken as a basis for selecting matches during optimisation to be: i)

forbidden matches, with unacceptable high risks, ii) allowed matches with penalties for medium risks, and

124

iii) allowed matches with acceptable risks. The penalties for medium risk matches are determined as

investments required to obtain matches with acceptable risks (Figure 2b). The observed matches can then

be transferred to the acceptable risk region. Three types of investments are possible in order to obtain this:

i) to decrease frequency, ii) to decrease severity and iii) both simultaneously. The investment to decrease

frequency can be a selection of different materials for equipment or increased thickness (diameter) of pipe

etc. Decreasing the severity consequences can be more complex as it mostly depends on the media of the

process, which would lead to process changes; however, it can be assumed as an investment for

additional safety equipment for detection, fire safety equipment etc. Only one type of investment is

assumed, representing the minimal investment for obtaining a network with acceptable risk.

Step 5: Synthesis of Total Site considering input data obtained in Step 4. For Total Site synthesis a

stochastic mixed integer nonlinear programming model has been used (Nemet, 2015a; Nemet et al,

2015b) considering the risk assessment results.

3. Illustrative case study

This illustrative case study consisted of Process 1 representing a sulphuric acid production plant and

Process 2 representing part of a refinery complex. The input data of process streams is presented in Table

2. The utility and cost data has been taken from Nemet et al. (2015a). The distance between processes is

assumed to be 2 km.

Table 2: Input data for process streams of the case study

Process Stream Type Tin/°C Tout/°C CP/ kW°C-1 h/kW m-2°C-1 Medium

Process 1 H1P1 Hot 980 430 20 0.065 Sulphuric acid

Process 2 C1P2 Cold 210 368 35 0.62 Oil

C2P2 Cold 75 95 250 0.5 water

Step 0: All three mechanisms of transfer were considered for the heat transfer between H1P1 and C1P2

as well as for H1P1 and C2P2 (Figure 1). Accounting for the required temperature difference ∆Tmin = 5 °C

the maximal possible heat transfers were determined for all possible heat transfer matches (Table 3).

Step 1: The deviations were determined based on Table 1 for each operation involved in heat transfer,

namely: heat exchangers, pipes, and pumps. In regard to heat exchangers the shell and tube types and for

pipes the centrifugal pumps with gaskets were assumed. Process streams, trans-streams at direct heat

transfer, Fresh W and IU water at indirect heat transfer were all assumed as above ground pipelines,

whilst Pipe 1 and Pipe 2 at direct heat transfer and IU steam and IU condensate at indirect heat transfer

were assumed as underground pipelines. The valve failure was unconsidered in this case study.

Table 3: Maximal possible heat transfer between all possible process-to-process heat transfers

Match type Streams involved Intermediate utility level range ∆Hmax/ kW

Direct H1P1-C1P2 - 5,530

H1P1-C2P2 - 5,000

Indirect H1P1-C1P2 LPS: 120-148 0

H1P1-C2P2 LPS: 120-148 5,000

Indirect H1P1-C1P2 MPS: 148-208 0

H1P1-C2P2 MPS: 148-208 5,000

Indirect H1P1-C1P2 HPS: 208-252 1,295

H1P1-C2P2 HPS: 208-252 5,000

*At THPS = 252°C and ∆Tmin = 5°C

Step 2: The failure frequencies were determined for the above listed equipment. The quantity of substance

present in the installation had to be determined in order to determine the severity of consequences of

deviations. As the methodology is developed for an early stage design the following simplified correlation

for quantities were considered in the heat exchanger. The area of heat exchanger AHE could be determined

from the amount of exchanged heat; overall heat transfer coefficient U and the logarithm mean

temperature ∆Tln. The correlation between the volume in the tubes and the area could be determined as

presented in Eq(4) The more common tube diameter sizes are between 19 mm and 25.24 mm (Edwards,

2008). For a preliminary calculation a 22 mm tube diameter was selected (11 mm radius) for a preliminary

calculation.

( )max ln/
HE

A H U T= ∆ ⋅ ∆ (3)

125

, 2
,

2 2 2

HE tube
HE tube HE

V r l r r
V A

A r l

⋅ ⋅
= = => = ⋅

⋅ ⋅ ⋅

(4)

The mass of content can be determined from the volume by multiplication of the VHE,tube by the density of

the content. The mass on the shell side can be determined from the mass determined by the ratio between

mass flow-rate on the tube side and the shell side of the heat exchanger Eq(5). The mass flow-rate is

determined from the heat transferred within the heat exchanger. Sensible heat is determined from Eq(6). If

the heat is stored in the form of latent heat (evaporation and condensation of intermediate utility), the mass

flow-rate is determined as a maximal enthalpy flow divided by the difference between the specific enthalpy

of steam, and water Eq(7).
shell shell shell shell

shell tube

tube tube tube tube

m m t m m
m m

m m t m m

⋅
= = => = ⋅

⋅

& & &

(5)

( )
max

/m H cp T= ∆ ⋅ ∆& (6)

( ) ( ) ( )( )max( ) /steam steam steam steam water steamm T H T h T h T= ∆ −& (7)
The substance has to be selected carefully for flowing into the tube as it can lead to substantially different

results. In regard to pipes the quantity present is determined as the amount within the pipeline, with the

length equal to the velocity multiplied by 600 s (Flemish Government, 2009). The mass flow-rate within the

pipelines is determined from the heat content in the process pipes. Regarding direct heat transfer the

mass flow-rate can be determined similarly as in Eq(6). For the indirect heat transfer the steam mass-flow

rate is determined by applying steam tables and maximal possible heat transfer, as presented in Eq(7).

Step 3 The risk is determined for each deviation separately for all the instruments involved during heat

transfer covering four major consequences of failure namely toxicity, flammability, explosiveness and

additionally for heat loss. The overall risk for a certain heat transfer is determined by summation of all the

risks regarding the equipment involved during heat transfer. The frequency of failure is similarly summed

up for equipment enabling heat transfer. The average severity is afterwards determined by dividing the

overall risk by the overall failure frequency. The obtained results are presented in Figure 3. It should be

noted that indirect heat transfer configuration for same match have similar risks, therefore, they can

overlap. As can be seen the frequency of the failure is higher in the cases when indirect heat transfer

configurations are planned. This is expected as more instruments are involved in heat transfer. However,

the severity of the failure is multiple times higher when direct heat transfer occurs. It is a consequence of

transporting either sulphur acid or crude between processes. This leads to generally higher risk during

direct heat transfer.

0.5

1.5

0 0.02 0.04 0.06

Severity

Frequency

H1P1-C1P2:C->H

H1P1-C1P2:H->C

0 0.02 0.04 0.06

Severity

Frequency

d)
H1P1-C2P2:H->C

H1P1-C1P2:H->C

H1P1-C1P2:C->H

H1P1-C2P2:C->H

100

120

0 0.02 0.04 0.06

Severity

Frequency

a) H1P1-C2P2:H->C

H1P1-C1P2:H->C

H1P1-C1P2:C->H

H1P1-C2P2:C->H
0

0.1

0.2

0.3

0.4

0.5

0 0.02 0.04 0.06

Severity

Frequency

b)
H1P1-C1P2:C->H

H1P1-C1P2:H->C
H1P1-C1P2:HPS

H->C: direct transfer; transport of hot process stream

to cold process stream

C->H: direct transfer; transport of cold process stream

to hot process stream

HPS: indirect transfer; high pressure steam

Figure 3: Overall frequency and severity of a certain heat transfer by considering a) toxicity, b)

flammability, c) explosiveness and d) heat loss risk assessment.

Step 4: As can be seen in Figure 3 matches H1P1-C2P2 and H1P1-C1P2 have significant risks for toxicity,

for direct heat transfer configuration involving heat transport by hot stream. It is a consequence of

sulphuric acid being the media in hot stream H1P1. The mentioned matches should be forbidden during

synthesis due to the high severity consequences of failure. The flammability risk and explosiveness risk

assessment support the decision of forbidding the match between H1P1-C1P2, when cold stream is

transported to hot stream, as the media is crude oil that is flammable and might become explosive. The

heat loss risk assessment highlights that match where the hot stream is transported having high risk of

heat loss.

Step 5: The results of the synthesis indicated that considering risk can provide different solutions

compared to those considering only economic criteria. In this particular case, the Expected Net Present

126

Value of network savings was decreased when considering the risk aspect, namely from 13,853 k€ to

10,831 k€, and the hot utility consumption was increased by 2,369 kW, from 748.5 kW to 3,117.3 kW.
Process 1Process 2

H1P1

C1P2

C2P2

Preheating

∆H = 5,781.2 kW∆H_loss =151.8 kW

∆H_loss = 99.4 kW

∆H =
4,251.5 kW

∆H =
748.5 kW

∆H = 890.5 kW∆H = 4,328.3 kW

Evaporation

Pinch in
Process 1

Pinch in
Process 2

Process 1Process 2

H1P1

C1P2

C2P2

Preheating

∆H = 1,160.3 kW

∆H = 1,907.6 kW

∆H =

311.9 kW

∆H =5,090.5 kW

∆H = 2,841.6 kW

∆H_loss = 90.5 kW

∆H_loss = 117.0 kW

∆H =
5,000 kW

∆H =
2,724.6 kW

∆H =
2,805.4 kW

Evaporation

Pinch in
Process 1

Pinch in
Process 2

∆H_loss = 76.8 kW
LPSLPS

LPS

MPS

Figure 4: Comparison between solutions obtained considering a) only economic value and b) economic

value and risk assessment

4. Conclusions

A framework has been presented for considering safety analysis during Total Site synthesis. The case

study indicates that solutions obtained when considering risk aspects can be significantly different

compared to those obtained without including the safety aspect-significantly safer solutions can thus be

obtained; however, with considerably higher consumption of external utilities.

Acknowledgement

The authors are grateful for the financial support from EC FP7 project FP7-PEOPLE-2011-IRSES “Energy

Systems Engineering – ESE”, project No. 294987, EC FP7 project ENER/FP7/296003/EFENIS ‘Efficient

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

Agency (Program No. P2-0032).

References

Chan I., Wan Alwi S.R., Hassim M.H., Manan Z.A., Klemeš J.J., 2014, Heat Exchanger Network Design

Considering Inherent Safety, Energy Procedia 61, 2469-2473.

Čuček L., Klemeš J.J., Varbanov P.S., Kravanja Z., 2015, Significance of environmental footprints for

evaluating sustainability and security of development, Clean Technologies and Environmental Policy, doi:

10.1007/s10098-015-0972-3

Dhole V.R., Linnhoff B., 1993, Total site targets for fuel, co-generation, emissions and cooling. Computers

and Chemical Engineering 17, 101–109.

Edwards J. E., 2008, Design and rating shell and tube heat exchangers, P&I Design Ltd, Teesside, UK

Flemish Government, LNE Department, Environment, Nature and Energy Policy Unit, Safety Reporting

Division, 2009, Handbook Failure Frequencies 2009 for drawing up a safety report.

<www.lne.be/themas/veiligheidsrapportage/rlbvr/bestanden-rlbvr/tr/vr_rlbvr_rl_hbff_EN.pdf> accessed

05.05.2015.

Heikkilä A-M., 1999, Inherent safety in process plant design an index-based approach PhD thesis,

University of Technology, Helsinki, Finland.

Klemeš J.J. (ed), 2013, Handbook of Process Integration (PI) Minimisation of Energy and Water Use,

Waste and Emissions, Woodhead Publishing Ltd/Elsevier, Cambridge, UK.

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

Programming (MP), Current Opinion of Chemical Engineering 2(4), 461-474.

Linnhoff B., Flower J.R., 1978, Synthesis of Heat Exchanger Networks: I. Systematic Generation of Energy

Optimal Networks. AIChE Journal, 24, 633–642.

Marhavilas P.K., Koulouriotis D., Gemeni V., 2011, Risk analysis and assessment methodologies in the

work sites: On a review, classification and comparative study of the scientific literature of the period

2000-2009, Journal of Loss Prevention in the Process Industries 24, 477-523.

Nemet A., 2015a, Synthesis of processes and process subsystems for entire lifetime, PhD Thesis,

University of Maribor, Slovenia.

Nemet A., Klemeš J.J., Kravanja Z., 2015b, Designing a Total Site for an entire lifetime under fluctuating

utility prices, Computers and Chemical Engineering, 72, 159-182.

Uijt de Haag P.A.M., Ale B.J.M., 2005, Guidelines for qualitative risk assessment, CPR18E, Gevaarlijke

Stoffen, Hague, The Netherland.